United States Department Evaluating the …...United States Department of Agriculture Forest Service...

United StatesDepartmentof Agriculture

Forest Service

Rocky MountainResearch Station

General TechnicalReport RMRS-GTR-63

September 2000

Evaluating the Effectivenessof Postfire RehabilitationTreatments

Peter R. RobichaudJan L. BeyersDaniel G. Neary

The Authors ________________________________________Peter R. Robichaud is a Research Engineer with the Station’s Soil and Water Engineering ResearchWork Unit, and is stationed at the Forestry Sciences Laboratory, 1221 South Main St., Moscow, ID 83843.

Jan L. Beyers is a Research Plant Ecologist with the Pacific Southwest Research Station’s PrescribedFire and Fire Effects Research Work Unit, and is stationed at the Riverside Fire Laboratory, 4955 CanyonCrest Dr., Riverside, CA 92507.

Daniel G. Neary is a Research Soil Scientist and Project Leader with the Sustainability of RiparianEcological Systems Research Work Unit and is stationed at the Southwest Forest Science Complex,2500 S. Pine Knoll, Flagstaff, AZ 86001.

Robichaud, Peter R.; Beyers, Jan L.; Neary, Daniel G. 2000. Evaluating the effectiveness of postfirerehabilitation treatments. Gen. Tech. Rep. RMRS-GTR-63. Fort Collins: U.S. Department of Agricul-ture, Forest Service, Rocky Mountain Research Station. 85 p.

Spending on postfire emergency watershed rehabilitation has increased during the past decade. Awest-wide evaluation of USDA Forest Service burned area emergency rehabilitation (BAER) treatmenteffectiveness was undertaken as a joint project by USDA Forest Service Research and National ForestSystem staffs. This evaluation covers 470 fires and 321 BAER projects, from 1973 through 1998 in USDAForest Service Regions 1 through 6. A literature review, interviews with key Regional and Forest BAERspecialists, analysis of burned area reports, and review of Forest and District monitoring reports wereused in the evaluation. The study found that spending on rehabilitation has increased to over $48 millionduring the past decade because the perceived threat of debris flows and floods has increased where firesare closer to the wildland-urban interface. Existing literature on treatment effectiveness is limited, thusmaking treatment comparisons difficult. The amount of protection provided by any treatment is small. Ofthe available treatments, contour-felled logs show promise as an effective hillslope treatment becausethey provide some immediate watershed protection, especially during the first postfire year. Seeding hasa low probability of reducing the first season erosion because most of the benefits of the seeded grassoccurs after the initial damaging runoff events. To reduce road failures, treatments such as properlyspaced rolling dips, water bars, and culvert reliefs can move water past the road prism. Channeltreatments such as straw bale check dams should be used sparingly because onsite erosion control ismore effective than offsite sediment storage in channels in reducing sedimentation from burnedwatersheds. From this review, we recommend increased treatment effectiveness monitoring at thehillslope and sub-catchment scale, streamlined postfire data collection needs, increased training onevaluation postfire watershed conditions, and development of an easily accessible knowledge base ofBAER techniques.

Keywords: burn severity, erosion control, BAER, burned area emergency rehabilitation, mitigation,seeding, monitoring.

Abstract ___________________________________________

The use of trade or firm names in this publication is for reader information and does notimply endorsement by the U.S. Department of Agriculture of any product or service.

Acknowledgments _____________Preparation of this report represents a true partnership

between Forest Service Research and National Forest Sys-tems staffs. Much of the work of collecting, collating, andtabulating the information contained in this report was con-ducted by research assistants J. Tenpas, K. Hubbert, and B.Nelson. We could not have completed it without them. Thecooperation of the six regional Burn Area Emergency Rehabili-tation (BAER) coordinators and many Forest BAER teammembers was essential to the success of this project, and wegratefully acknowledge their contributions. Funding for theproject was provided by the Forest Service, Watershed and AirStaff, with additional contributions from the Rocky Mountainand Pacific Southwest Research Stations.

Contents_______________________Page

Introduction ................................................................1USDA Forest Service Burn Area Emergency

Rehabilitation (BAER) History .......................1Problem Statement and Objectives .......................2Definitions ..............................................................2

Literature Review .......................................................5Fire’s Impact on Ecosystems .................................5

Fire Effects on Watersheds ...............................5Water Quality .....................................................8Surface Erosion .................................................8Sediment Yield and Channel Stability ...............9Mass Wasting ....................................................9Dry Ravel .........................................................11

Emergency Watershed RehabilitationTreatment Effectiveness .............................11

Hillslope Treatments ........................................11Channel Treatments ........................................20Road Treatments .............................................21

Methods ...................................................................21Burned Area Report Data ....................................21Interview Survey ..................................................22

Project Review Interview Form ........................22

Page

No Action Review Interview Form ...................22Treatment Actions Interview Form ..................22Relative Benefits Interview Form .....................22

Monitoring Reports ..............................................22Analysis Methods .................................................22

Burned Area Reports and Interview Forms .....22Monitoring Reports ..........................................23

Results .....................................................................23Overview of Data Collected .................................23Fire Severity .........................................................23Erosion Estimates ................................................26Hydrologic Estimates ...........................................27Risk Analysis .......................................................27Probability of Success .........................................29Cost of No Action/Alternatives .............................29BAER Team Members .........................................30Monitoring Reports ..............................................30Treatment Effectiveness Ratings .........................36

Hillslope Treatments ........................................36Channel Treatments ........................................37Road and Trail Treatments ..............................39

Treatment Rankings ............................................39Discussion ................................................................44

BAER Assessments .............................................44BAER Project Monitoring .....................................46Treatment Effectiveness ......................................47

Hillslope Treatments ........................................47Channel Treatments ........................................52Road Treatments .............................................53

Conclusions..............................................................53Recommendations ...................................................54References ...............................................................54Appendix A—Example Data and

Interview Forms...........................................60Appendix B—Treatment Effectiveness

Summaries ..................................................74Hillslope Treatments ............................................74Channel Treatments ............................................79Road Treatments .................................................83

You may order additional copies of this publication by sending your mailing information in labelform through one of the following media. Please specify the publication title and number.

Telephone (970) 498-1392

FAX (970) 498-1396

E-mail [email protected]

Mailing Address Publications DistributionRocky Mountain Research Station240 West Prospect RoadFort Collins, CO 80526

1USDA Forest Service Gen. Tech. Rep. RMRS-GTR-63. 2000

Introduction ____________________Recent large, high severity fires coupled with sub-

sequent major hydrological events have generatedrenewed interest in the linkage between fire andonsite and downstream effects. Fire is a natural andimportant disturbance mechanism in many ecosys-tems. However, the intentional human suppression offires in the Western United States, beginning in theearly 1900’s, has altered natural fire regimes in manyareas (Agee 1993). Fire suppression can allow fuelloading and forest floor material to increase, resultingin fires of greater intensity and extent than mighthave occurred otherwise (Norris 1990). High severityfires are of particular concern because they can affectsoil productivity, watershed response, and downstreamsedimentation, causing threats to human life andproperty. During severe fire seasons, the USDA ForestService and other land management agencies spendmillions of dollars on postfire emergency watershedrehabilitation measures intended to minimize floodrunoff, onsite erosion, and offsite sedimentation andhydrologic damage. Increased erosion and floodingare certainly the most visible and dramatic impactsof fire apart from the consumption of vegetation.

USDA Forest Service Burn AreaEmergency Rehabilitation (BAER) History

The first formal reports on emergency watershedrehabilitation after wildfires were prepared in the1960’s and early 1970’s, although postfire seedingwith grasses and other herbaceous species was con-ducted in many areas in the 1930’s, 1940’s and 1950’s(Christ 1934, Gleason 1947). Contour furrowing andtrenching were used when flood control was a majorconcern (DeByle 1970b, Noble 1965). No formal emer-gency rehabilitation program existed, and funds forwatershed rehabilitation were obtained from firesuppression accounts, emergency flood control pro-grams, or appropriated watershed restoration ac-counts. In response to a Congressional inquiry onfiscal accountability, in 1974 a formal authority forpostfire rehabilitation activities was provided in theInterior and Related Agencies appropriation. This

BAER authority integrated the evaluation of fireseverity, funding request procedures, and treatmentoptions.

The occurrence of many large fires in California andsouthern Oregon in 1987 caused expenditures forBAER treatments to exceed the annual BAER authori-zation of $2 million. On several occasions inappropriaterequests were made for nonemergency items, and clari-fications were issued that defined real emergencysituations warranting rehabilitation treatments. Poli-cies were incorporated into the Forest Service Manual(FSM 2523) and the BAER Handbook (FSH 2509.13)that required an immediate assessment of siteconditions following wildfire and, where necessary,implementation of emergency rehabilitation mea-sures to: (1) minimize the threat to life and propertyonsite or offsite; (2) reduce the loss of soil and onsiteproductivity; (3) reduce the loss of control of water; and(4) reduce deterioration of water quality. A concertedeffort was made to train BAER team leaders, andregional and national BAER training programs be-came more frequent. At the same time, debates aroseover the effectiveness of grass seeding and its negativeimpacts on natural regeneration. Seeding was still themost widely used treatment, though often applied inconjunction with other hillslope treatments, such ascontour-felled logs, and channel treatments, includingstraw bale check dams. National Forest specialistswere encouraged to do implementation monitoring oftreatment establishment, as well as some form ofeffectiveness monitoring of treatment performance,using regular watershed appropriation funds.

In the mid 1990’s, a major effort was undertaken torevise and update the BAER handbook. A steeringcommittee, consisting of regional BAER coordinatorsand other specialists, organized and developed thebulk of the handbook used today. The issue of usingnative species for emergency revegetation emerged asa major topic, and the increased use of contour-felledlogs caused rehabilitation expenditures to escalate.During the busy 1996 fire season, for example, theForest Service spent $11 million on BAER projects.

Improvements in the BAER program in the late1990’s included increased BAER training and fundingreview. Increased needs were identified for BAER

Evaluating the Effectiveness ofPostfire Rehabilitation TreatmentsPeter R. RobichaudJan L. BeyersDaniel G. Neary

2 USDA Forest Service Gen. Tech. Rep. RMRS-GTR-63. 2000

team leader training, project implementation train-ing, and on-the-ground treatment installation train-ing. Courses were developed for the first two trainingneeds but not the last. Current funding requests arescrutinized by the Regional and national BAER coor-dinators to verify that they are minimal, necessary,reasonable, practicable, cost-effective, and will pro-vide significant improvement over natural recovery.

Also in the late 1990’s, a program was initiated tointegrate national BAER policies across different Fed-eral agencies, as each agency interpreted BAER fund-ing differently. The U.S. Department of Agriculture andDepartment of the Interior approved a joint policy fora consistent approach to BAER in 1998. The new policybroadened the scope and application of BAER analysisand treatment. Major changes included: (1) monitor-ing to determine if additional treatment is needed andevaluating to improve treatment effectiveness; (2) re-pairing facilities for safety reasons; (3) stabilizing bioticcommunities; and (4) preventing unacceptable degra-dation of critical known cultural sites and naturalresources. These changes affect the Forest Service, theBureau of Land Management, the National Park Ser-vice, the Fish and Wildlife Service, and the Bureau ofIndian Affairs.

Problem Statement and Objectives

In spite of the improvements in the BAER processand the wealth of practical experience obtained overthe last several decades, the effectiveness of manyemergency rehabilitation methods has not been sys-tematically tested or validated. BAER team leadersand decisionmakers often do not have informationavailable to thoroughly evaluate the short- and long-term benefits (and costs) of various treatment options.

In 1998, at the request of and funded by the USDAForest Service Washington office Watershed and Airstaff, a joint study was initiated by the USDA ForestService Rocky Mountain Research Station and thePacific Southwest Research Station to evaluate theuse and effectiveness of postfire emergency rehabili-tation methods. The objectives of the study were to:(1) evaluate the effectiveness of rehabilitation treat-ments at reducing postfire erosion, runoff, or othereffects; (2) assess the effectiveness of rehabilitationtreatments in mitigating downstream effects of in-creased sedimentation and peakflows; (3) investigatethe impacts of rehabilitation treatments on naturalprocesses of ecosystem recovery, both in the short- andlong-term; (4) compare hillslope and channel treat-ments in terms of relative benefits, and how theycompare to a no-treatment option; (5) collect availableinformation on economic, social, and environmentalcosts and benefits of various rehabilitation treatmentoptions, including no treatment; (6) determine howknowledge of treatments gained in one location can be

transferred to another location; and (7) identify infor-mation gaps needing further research and evaluation.

The study collected and analyzed information onpast use of BAER treatments. Specifically, we sought todetermine attributes and conditions that led to treat-ment success or failure, and the effectiveness oftreatments in achieving BAER goals. Because muchof the information was unpublished and qualitative innature, resource specialists were interviewed regard-ing their BAER activity experiences.

This report is divided into six major sections: (1) areview of published literature on fire effects and BAERtreatments; (2) information acquisition and analysismethods; (3) description of results, which include hy-drologic, erosion and risk assessments, monitoring re-ports, and treatment evaluations; (4) discussion of BAERassessments and treatment effectiveness; (5) conclu-sions drawn from the analysis; and (6) recommendations.

Definitions

The literature of emergency watershed rehabilita-tion contains may terms from hydrological, ecologicaland fire science disciplines. For clarity the terms usedin this manuscript are defined below.

Aerial Seeding: See Seeding.Allelopathy: Inhibition of competing plant growth

by exudation of naturally produced, phytotoxicbiochemicals.

Annuals (Annual Plants): Plant that completes itsgrowth and life cycle in one growing season.

Ash-bed Effect: Stimulation of plant growth causedby the sudden availability of fire-mineralizedplant nutrients contained in ash residues froma fire.

Armored Ford Crossing: Road crossing of a peren-nial or ephemeral stream at or near the exist-ing cross-section gradient that is generallyconstructed of large rocks capable of bearingthe weight of the vehicles and resisting trans-port by the stream.

Armoring: Protective covering, such as rocks, vegeta-tion or engineering materials used to protectstream banks, fill or cut slopes, or drainagestructure outflows from flowing water.

BAER: Burned Area Emergency Rehabilitation.Best Management Practices: Preferred activities

which minimize impacts on soil, water, andother resources.

Broadcast seeding: See Seeding.Burn Severity: Qualitative and quantitative mea-

sure of the effects of fire onsite resources suchas soil and vegetation. Fire intensity contrib-utes to severity but does not alone define it.

Chaparral: Shrub-dominated evergreen vegetationtype abundant in low- to mid-level elevationsin California and the Southwest.


Channel Clearing: Removal of woody debris fromchannels by heavy equipment or cable yarding.

Channel Loading: Sediment inputs into ephemeralor perennial stream channels.

Check Dam: Small structure in zero or first orderchannels made of rocks, logs, plant materials,or geotextile fabric designed to stabilize thechannel gradient and store a small amount ofsediment.

Contour-Felled Logs: System for detaining runoffand sediment on slopes by felling standing tim-ber (snags) along the contour, delimbing andanchoring the logs, and backfilling to createsmall detention basins. Also known as contour-felling, contour log terraces, log erosion barriers(LEBs). In some regions, contour-felling describesonly felling the standing timber in the directionof the contour but not anchoring or backfilling.

Contour Furrowing: See Contour Trenching.Contour Trenching: Construction of trenches on

slope contours to detain water and sedimenttransported by water or gravity downslopegenerally constructed with light equipment.These are also known as contour terraces orcontour furrowing.

Cross Drain: A ditch relief culvert or other structureor shaping of a road surface designed to captureand remove surface water flow.

Culvert Overflow: Specially designed sections ofroadway that allow for overflow of relief cul-verts or cross-drain culverts without compro-mising the integrity of the road surface.

Culvert Riser: Vertical extension of culvert on theuphill side to create a small pond for detainingsediment.

Culvert Upgrading: Replacing existing culvertswith large diameter ones. May also includearmoring of inlet and outlet areas.

Debris Avalanche: Mass failure of variably sizedslope segments characterized by the rapiddownhill movement of soil and underlyinggeologic parent material.

Debris Basin: Specially engineered and constructedbasin for storing large amounts of sedimentmoving in an ephemeral stream channel.

Debris Clearing: See Channel Clearing.Design Storm: Estimate of rainfall amount and duration

over a particular drainage area. Often used inconjunction with the design storm return period,which is the average number of years withinwhich a given hydrological event is equaled orexceeded (i.e., 5-year return period).

Ditch Maintenance: Various maintenance activi-ties to maintain or restore the capacity ofditches to transport water. Activities includesediment and woody debris removal, reshap-ing, and armoring.

Dry Ravel: Downhill movement of loose soil androck material under the influence of gravityand freeze-thaw processes.

Ephemeral Stream or Channel: Drainage waywhich carries surface water flow only afterstorm events or snow melt.

Energy Dissipater: Rock, concrete, or imperviousmaterial structure which absorbs and reducesthe impact of falling water.

Erosion: Detachment and transport of mineral soilparticles by water, wind, or gravity

Fire Intensity: Rate at which fire is producingthermal energy in the fuel-climate environ-ment in terms of temperature, heat yield perunit mass of fuel, and heat load per unit area.

Fire Severity: See Burn Severity.Forb: Herbaceous plant other than grasses or grass-

like plants.Gabion: A woven galvanized wire basket sometimes

lined with geotextiles and filled with rock, stackedor placed to form an erosion resistant structure.

Geotextile (Geowebbing): Fabric, mesh, net, etc.made of woven synthetic or natural materialsused to separate soil from engineering mate-rial (rocks) and add strength to a structure.

Grade Stabilizer: Structure made of rocks, logs, orplant material installed in ephemeral channelsat the grade of the channel to preventdowncutting.

Ground Seeding: See Seeding.Hand Trenching: Contour trenching done manually

rather than mechanically.Hydrophobic Soil: See Water Repellency.In-channel Felling: Felling of snags and trees into

stream channel to provide additional woodydebris for trapping sediment.

Infiltration: Movement of rainfall into litter and thesoil mantle.

Lateral Keying: Construction or insertion of log orrock check dam 1.5 to 3 ft (0.4 to 1.0 m) intostream or ephemeral channel banks.

Log Check Dam: See Check Dam.Log Erosion Barriers (LEBS): see Contour-Felled

Logs.Log Terraces: See Contour-Felled Logs.Mass Wasting: Movement of large amounts of soil

and geologic material downslope by debrisavalanches, soil creep, or rotational slumps.

Mg ha–1: Metric ton per hectare or megagram perhectare, equivalent to 0.45 tons per acre(0.45 t ac–1).

Monitoring: The collection of information to deter-mine effects of resource management or specifictreatments, used to identify changing condi-tions or needs.

Monitoring, Compliance: Monitoring done to assurecompliance with Best Management Practices.


Monitoring, Effectiveness: Monitoring done to de-termine the effectiveness of a treatment inaccomplishing the desired effect.

Monitoring, Implementation: Monitoring done toverify installation of treatment was accom-plished as specified in installation instructiondocuments.

Mulch: Shredded woody organic material, grass, orgrain stalks applied to the soil surface to pro-tect mineral soil from raindrop impact andoverland flow.

Mychorrhizae: Fungi which symbiotically functionwith plant roots to take up water and nutrients,thereby greatly expanding plant root systems.

Outsloping: Shaping a road surface to deflect waterperpendicular to the traveled way rather thanparallel to it.

Peakflow: Maximum flow during storm or snow meltrunoff for a given channel.

Perennials (Perennial Plants): Plants that con-tinue to grow from one growing season to thenext.

Perennial Stream and Channel: Drainage ways inwhich flow persists throughout the year withno dry periods.

Plant Cover: Percentage of the ground surface areaoccupied by living plants.

Plant Species Richness: Number of plant speciesper unit area.

Ravel: See Dry Ravel.Re-bar: Steel reinforcing bar, available in various

diameters, used to strengthen concrete or an-chor straw bales and wattles.

Regreen: Commercially available sterile wheatgrasshybrid used to stabilize slopes immediatelyafter a fire but not interfere with subsequentnative plant recovery.

Relief Culvert: Conduit buried beneath road sur-face to relieve drainage in longitudinal ditchat the toe of a cut slope.

Return Interval: Probabilistic interval for recurrence(1, 2, 5, 10, 20, 50, 100 years etc.) of stormflow,rainfall amount or rainfall intensity.

Rill: Concentrated water flow path, generally formedon the surface of bare soil.

Riparian Area: Area alongside perennial or ephem-eral stream that is influenced by the presenceof shallow groundwater.

Ripping: See Tilling.Risk: The chance of failure.Rock Cage Dam: See Gabion or Check Dam.Rolling Dip: Grade reversal designed into a road to

move water off of short slope section ratherthan down long segment.

Rotational Slump: Slope failure characterized rota-tion of the soil mass to a lower angle of repose.

Runoff: Movement of water across surface areas of awatershed during rainfall or snowmelt events.

Sediment: Deposition of soil eroded and transportedfrom locations higher in the watershed.

Sedimentation: Deposition of water, wind, or gravityentrained soil and sediment in surface depres-sions, side slopes, channel bottoms, channelbanks, alluvial flats, terraces, fans, lake bot-toms, etc.

Sediment Trap Efficiency: Percent of contour-felledlog length showing accumulated sedimentrelative to available length of log. Or percent ofsediment accumulated behind logs relative toavailable storage capacity of the logs. Or per-cent of sediment stored behind logs relative tosediment that was not trapped and moved tothe base of a hillslope.

Sediment Yield (Production): Amount of sedimentloss off of unit area over unit time period usu-ally expressed as tons ac–1 yr–1 or Mg ha–1 yr–1.

Seeding: Application of plant seed to slopes by air-craft (Aerial Seeding or Broadcast Seeding), orby ground equipment or manually (GroundSeeding).

Silt Fence: Finely woven fabric material used todetain water and sediments.

Slash Spreading: Dispersal of accumulations ofbranches and foliage over wider areas.

Slope Creep: Slow, downhill movement of soil mate-rial under the influence of gravity.

Soil/Site Productivity: Capability of a soil type orsite to produce plant and animal biomass in agiven amount of time.

Soil Wettability: See Water Repellency.Storm Duration: Length of time that a precipitation

event lasts.Storm Magnitude: Relative size of precipitation event.Storm Patrol: Checking and cleaning culvert inlets

to prevent blockage during storm runoff.Straw Bale Check dam: Check dam made of straw

or hay bales often stacked to provide additionalstorage capacity. Designed to store sedimentand/or prevent downcutting.

Straw Wattle: Woven mesh netting (1 ft diameter by6 to 20 ft in length, 0.3 m diameter by 1.8 m to6.1 m in length) filled with straw or hay andsometimes seed mixes, used to trap sedimentand promote infiltration.

Stream Bank Armoring: Reinforcing of streambankwith rock, concrete, or other material to reducebank cutting and erosion.

Streamflow: Movement of water in a drainagechannel.

Temporary Fencing: Fencing installed on a grazingallotment or other unit to keep cattle or nativeungulates out of burned area.


Terracette: See Contour-Felled Logs.Tilling: Mechanical turning of the soil with a plow or

ripping device. Often used to promote soil infil-tration by breaking up water repellent soillayers.

Trash Rack: Barrier placed upstream of a culvert toprevent woody debris from becoming jammedinto the inlet.

Ungulate: Herbivorous animals with hooves, e.g., cow,elk, deer, horses, etc.

Water Bar: Combination of ditch and berm installedperpendicular or skew to road or trail centerlineto facilitate drainage of surface water; some-times nondriveable and used to close a road.

Water Repellency: Tendency of soil to form a hydro-phobic (water resistant) layer during fire thatsubsequently prevents infiltration and perco-lation of water into the soil mantle.

Watershed: An area or region bounded peripher-ally by ridges or divides such that all precipi-tation falling in the area contributes to itswatercourse.

Water Yield: Total runoff from a drainage basin.

Literature Review _______________Our evaluation of BAER treatment effectiveness

began with the published scientific literature. Thegeneral effects of fire on Western forested landscapesare well documented (Agee 1993, DeBano and others1998, Kozlowski and Ahlgren 1974). Conversely, manyof the processes addressed by BAER treatments havenot been extensively studied, and relatively little in-formation has been published about most emergencyrehabilitation treatments with the exception of grassseeding. To put BAER treatment effectiveness intoecosystem context, we summarize the scientific litera-ture on postfire conditions that are relevant to BAERevaluations. Then we examine published studies onspecific BAER treatments.

Fire’s Impact on Ecosystems

All disturbances produce impacts on ecosystems.The level and direction of impact (negative or positive)depends on ecosystem resistance and resilience, aswell as on the severity of the disturbance. The vari-ability in resource damage and response from site tosite and ecosystem to ecosystem is highly dependenton burn or fire severity.

Burn severity (fire severity) is a qualitative measureof the effects of fire onsite resources (Hartford andFrandsen 1992, Ryan and Noste 1983). As a physical-chemical process, fire produces a spectrum of effectsthat depend on interactions of energy release (inten-sity), duration, fuel loading and combustion, vegeta-tion type, climate, topography, soil, and area burned.

Fire intensity is an integral part of burn severity,and the terms are often incorrectly used synony-mously. Intensity refers to the rate at which a fire isproducing thermal energy in the fuel-climate environ-ment (DeBano and others 1998). Intensity is mea-sured in terms of temperature and heat yield. Surfacetemperatures can range from 120 to greater than2,730 °F (50 to greater than 1,500 °C). Heat yields perunit area can be as little as 59 BTU ft–2 (260 kg-cal m–2)in short, dead grass to as high as 3700 BTU ft–2

(10,000 kg-cal m–2) in heavy logging slash (Pyne andothers 1996). Rate of spread is an index of fire dura-tion and can vary from 1.6 ft week–1 (0.5 m week–1)in smoldering peat fires to as much as 15 mi hr–1

(25 km hr–1) in catastrophic wildfires.The component of burn severity that results in the

most damage to soils and watersheds, and henceecosystem stability, is duration. Fast moving fires infine fuels, such as grass, may be intense in terms ofenergy release per unit area, but do not transfer thesame amounts of heat to the forest floor, mineral soil,or soil organisms as do slow moving fires in moderateto heavy fuels. The impacts of slow moving, low orhigh intensity fires on soils are much more severe andcomplex. The temperature gradients that develop canbe described with a linked-heat transfer model(Campbell and others 1995) and are a function ofmoisture and fuel loadings.

Some aspects of burn severity can be quantified,but burn severity cannot be expressed as a singlequantitative measure that relates to resource impact.Therefore, relative magnitudes of burn or fire sever-ity, expressed in terms of the postfire appearance oflitter and soil (Ryan and Noste 1983), are better criteriafor placing burn or fire severity into broadly defined,discrete classes, ranging from low to high. A generalburn severity classification developed by Hungerford(1996) relates burn severity to the soil resource re-sponse (table 1).

Fire Effects on Watersheds—Soils, vegetation,and litter are critical to the functioning of hydrologicprocesses. Watersheds with good hydrologic condi-tions and adequate rainfall sustain stream baseflowconditions for much or all of the year and produce littlesediment. With good hydrologic condition (greaterthan 75 percent of the ground covered with vegetationand litter), only about 2 percent or less of rainfallbecomes surface runoff, and erosion is low (Bailey andCopeland 1961). When site disturbances, such as se-vere fire, produce hydrologic conditions that are poor(less than 10 percent of the ground surface coveredwith plants and litter), surface runoff can increaseover 70 percent and erosion can increase by threeorders of magnitude.

Within a watershed, sediment and water responsesto wildfire are often a function of burn severity and the


Table 1—Burn severity classification based on postfire appearances of litter and soil and soiltemperature profiles (Hungerford 1996, DeBano et al. 1998).

Burn SeveritySoil and Litter Parameter Low Moderate High

Litter Scorched, Charred, Consumed ConsumedConsumed

Duff Intact, Surface Deep Char, ConsumedChar Consumed

Woody Debris - Small Partly Consumed, Consumed ConsumedCharred

Woody Debris - Logs Charred Charred Consumed,Deeply Charred

Ash Color Black Light Colored Reddish, Orange

Mineral Soil Not Changed Not Changed Altered Structure,Porosity, etc

Soil Temp. at 0.4 in (10 mm) <120 °F 210-390 °F >480 °F(<50 °C) (100-200 °C) (>250 °C)

Soil Organism Lethal Temp. To 0.4 in (10 mm) To 2 in (50 mm) To 6 in (160 mm)

occurrence of hydrologic events. For a wide range ofburn severities, the impacts on hydrology and sedi-ment loss can be minimal in the absence of precipita-tion. However, when a precipitation event follows alarge, moderate- to high-burn severity fire, impactscan be far reaching. Increased runoff, peakflows, andsediment delivery to streams can affect fish popula-tions and their habitat (Rinne 1996).

Fire can destroy accumulated forest floor materialand vegetation, altering infiltration by exposing soilsto raindrop impact or creating water repellent condi-tions (DeBano and others 1998). Loss of soil fromhillslopes produces several significant ecosystem im-pacts. Soil movement into streams, lakes, and ripar-ian zones may degrade water quality and change thegeomorphic and hydrologic characteristics of thesesystems. Soil loss from hillslopes may reduce siteproductivity.

Total water yields across the Western United Statesvary considerably depending on precipitation, evapo-transpiration (ET), soil, and vegetation. The magni-tude of measured increases in water yield the firstyear after fire can vary greatly within a location orbetween locations depending on fire severity, cli-mate, precipitation, geology, soils, topography, veg-etation type, and proportion of the vegetation burned.Because increases in water yield are primarily due toelimination of plant cover, with subsequent reduc-tions in the transpiration component of ET, flowincreases are greater in humid ecosystems withhigh prefire ET (Anderson and others 1976). El-evated streamflow declines through time as woody

and herbaceous vegetation regrow, with this recov-ery period ranging from a few years to decades.

Increases in annual water yield after wildfires andprescribed fires are highly variable (table 2). Hibbertand others (1982) reported a 12 percent increase inwater yield after prescribed fire in an Arizona pinyon-juniper forest. A wildfire in the mostly ponderosa pineEntiat watershed in Washington produced a 42 per-cent increase in water yield the first postfire year(Helvey 1980). The first-year increase in water yieldafter a prescribed burn in a Texas grassland was 1,150percent of the unburned control watershed, but theincrease over the control was only 400 percent wherea rehabilitation treatment (seeding) was done afterthe fire (Wright and others 1982). Seeding also short-ened the recovery period from 5 to 2 years. In Arizonachaparral burned by wildfire, the first-year wateryield increase exceeded 1,400 percent (Hibbert 1971).Where soil wettability becomes a problem, water yieldincreases can be very high due to greater stormflows.

The effects of fire disturbance on storm peakflowsare highly variable and complex. They can producesome of the most profound watershed and riparianimpacts that forest managers have to consider. In-tense short duration storms that are characterized byhigh rainfall intensity and low volume have beenassociated with high stream peakflows and significanterosion events after fires (Neary and others 1999). Inthe Intermountain West, high intensity, short dura-tion rainfall is relatively common (Farmer and Fletcher1972). Five minute rainfall rates of 8.4 to 9.2 in hr–1

(213 and 235 mm hr–1) have been associated with


Table 2—Effects of prescribed fires and wildfires on water yield based in different vegetation types.

Flow RecoveryLocation Precipitation Flow Added Period Reference

(in) (mm) (in) (mm) (%) (years)Douglas-fir, OR 98 2480 Bosch and Hewlett 1982

Control 74 1890 — —Cut 82%, Burned 88 2230 20 >5

Douglas-fir, OR 94 2390 Bosch and Hewlett 1982Control 54 1380 — —Cut 100%, Burned 72 1840 34 >5

Ponderosa Pine/Douglas-fir, WA 23 580 Helvey 1980Control (Preburn) 9 220 — —Wildfire (Postburn) 12 315 42 ?

Chaparral, AZ 29 740 Davis 1984Control 3 75 — —Prescribed Fire 6 155 144 >11

Chaparral, AZ 23 580 Hibbert and others 1982Control 3 75 — —Wildfire 5 130 59 ?

Chaparral, AZ 26 655 Hibbert 1971Control 0 0 — —Wildfire 5 125 >99 >9Control 0.7 20 — —Wildfire 11 290 1421 >9

Pinyon-Juniper, AZ 19 480 Hibbert and others 1982Control 1 25 — —Prescribed Fire 1.5 40 12 5

Juniper-Grass, TX 26 660 Wright and others 1982Control 0.1 2 — —Prescribed Fire 1 25 1150 5Rx Fire, Seeded 0.4 10 400 2

Aspen-Mixed Conifer Bosch and Hewlett 1982Control 6 155 — —Wildfire 8 190 22 5

peakflows from recently burned areas that were in-creased 556 percent above that for adjacent areas(Croft and Marston 1950). Anderson and others (1976)produced a good review of peakflow response to dis-turbance (table 3). Wildfires generally increasepeakflows. Peakflow increases of 500 to 9,600 percentare common in the Southwest, while those measuredin the Cascade region are much lower (Anderson andothers 1976). For example, the Tillamook burn in 1933in Oregon increased the total annual flow of twowatersheds by 9 percent and increased the annualpeakflow by 45 percent (Anderson and others 1976). A310 ac (127 ha) wildfire in Arizona increased summerpeakflows by 500 to 1,500 percent, but had no effect onwinter peakflows. Another wildfire in Arizona pro-duced a peakflow 58 times greater than an unburnedwatershed during record autumn rainfalls. Peakflowincreases following wildfires in Arizona chaparral ofup to 45,000 percent have been reported (Glendeningand others 1961). Watersheds in the Southwest are

prone to these enormous peakflow responses becauseof climatic, topographic, and soil conditions. Theseinclude intense monsoon rainfalls common in thatregion at the end of the spring fire season; steepterrain; shallow, skeletal soils; and water repellency,which often develops in soils under chaparral vege-tation. Recovery times can range from years to manydecades. Studies have shown both increases (+35 per-cent) and decreases (–50 percent) in snowmeltpeakflows following fires (Anderson and others 1976).

Burned watersheds generally respond to rainfall fasterthan unburned watersheds, producing more “flashfloods” (Anderson and others 1976). Water repellentsoils and cover loss will cause flood peaks to arrivefaster, rise to higher levels, and entrain significantlygreater amounts of bedload and suspended sediments.Flood warning times are reduced by “flashy” flow, andthe high flood levels can be devastating to property andhuman life. Although these concepts of stormflow tim-ing are well-understood within the context of wildland


hydrology, some studies have confounded results be-cause of the combined changes in volume, peak andtiming at different locations in the watershed, and theseverity and size of the disturbance in relation to thesize of watershed (Brooks and others 1997).

Water Quality—Increases in streamflow after firecan result in substantial to little effect on the physicaland chemical quality of streams and lakes, depending onthe size and severity of the fire (DeBano and others1998). Higher streamflows and velocities result in addi-tional transport of solid and dissolved materials that canadversely affect water quality for human use and dam-age aquatic habitat. The most obvious effects are pro-duced by suspended and bedload sediments, but sub-stantial changes in anion/cation chemistry can occur.

Undisturbed forest, shrub, and range ecosystemsusually have tight cycles for major cations and anions,resulting in low concentrations in streams. Distur-bances such as cutting, fires, and insect outbreaks inter-rupt or temporarily terminate uptake by vegetationand may affect mineralization, microbial activity, nitri-fication, and decomposition. These processes result inthe increased concentration of inorganic ions in soilwhich can be leached to streams via subsurface flow(DeBano and others 1998). Nutrients carried to streamscan increase growth of aquatic plants, reduce thepotability of water supplies, and produce toxic effects.

Most attention relative to water quality after firefocuses on nitrate nitrogen (NO3-N) because it ishighly mobile. High NO3-N levels, in conjunction withphosphorus, can cause eutrophication of lakes and

streams. Most studies of forest disturbances showincreases in NO3-N, with herbicides causing the larg-est increases (Neary and Hornbeck 1994, Tiedemannand others 1978).

Surface Erosion—Surface erosion is the move-ment of individual soil particles by a force and isusually described by three components: (1) detach-ment, (2) transport, and (3) deposition. Inherenterosion hazards are defined as site properties thatinfluence the ease which individual soil particles aredetached (soil erodibility), slope gradient and slopelength. Forces than can initiate and sustain themovement of soil particles include raindrop impact(Farmer and Van Haveren 1971), overland flow(Meeuwig 1971), gravity, wind, and animal activity.Protection is provided by vegetation, surface litter,duff, and rocks that reduce the impact of the appliedforces and aid in deposition (Megahan 1986, McNabband Swanson 1990).

Erosion is a natural process occurring on land-scapes at different rates and scales, depending ongeology, topography, vegetation, and climate. Natu-ral erosion rates increase as annual precipitationincreases (table 4). Landscape disturbing activitiessuch as mechanical site preparation, agriculture,and road construction lead to the greatest erosion,which generally exceeds the upper limit of naturalgeologic erosion (Neary and Hornbeck 1994). Firesand fire management activities (fireline construc-tion, temporary roads, heli-pad construction, andpostfire rehabilitation) can also affect erosion.

Table 3—Effects of harvesting and fire on peakflows in different habitat types (from Anderson andothers 1976).

Location Treatment Other Information Peakflow Change

(%)Douglas-fir, OR Clearcut Fall Storms +90

Clearcut Winter Storms +28

Douglas-fir, OR Clearcut, 100% Burn +30Clearcut, 50% Burn +11

Douglas-fir, OR Wildfire +45

Chaparral, CA Wildfire +2282

Chaparral, AZ Wildfire Summer Flows +500Wildfire Summer Flows +1500Wildfire Winter Flows 0

Chaparral, AZ Wildfire Fall Flows +5800

Ponderosa Pine, AZ Wildfire Summer Flows +9605

Mixed Conifer, AZ Wildfire (Rich 1962) Low Summer Flow +1521Wildfire (Rich 1962) Inter. Summer Flow +526Wildfire (Rich 1962) High Summer Flow +960

Aspen-Conifer, CO Clearcut, 100%


Table 4—Natural watershed sediment losses in the USA based on published literature.

Location Watershed Conditions Sediment Loss Reference

(t ac–1) (Mg ha–1)USA Geologic Erosion Schumm and Harvey 1982

Natural Rate, Lower Limit 0.3 0.6Natural Rate, Upper Limit 7 15

Eastern USA Forests, Lower Baseline 0.05 0.1 Patric 1976Forests, Upper Baseline 0.1 0.2

Western USA Forests, Lower Baseline 0.0004 0.001 Biswell and Schultz 1965Forests, Upper Baseline 2 6 DeByle and Packer 1972

Sediment yields 1 year after prescribed burns andwildfires range from very low, in flat terrain and in theabsence of major rainfall events, to extreme, in steepterrain affected by high intensity thunderstorms(table 5). Erosion on burned areas typically declines insubsequent years as the site stabilizes, but the ratevaries depending on burn or fire severity and vegeta-tion recovery. Soil erosion after fires can vary fromunder 0.4 to 2.6 t ac–1 yr–1 (0.1 to 6 Mg ha–1 yr–1) inprescribed burns and from 0.2 to over 49 t ac–1 yr–1

(0.01 to over 110 Mg ha–1 yr–1) in wildfires (Megahanand Molitor 1975, Noble and Lundeen 1971, Robichaudand Brown 1999) (table 5). For example, Radek (1996)observed erosion of 0.1 to 0.8 t ac–1 (0.3 to 1.7 Mg ha–1)from several large wildfires that covered areas rang-ing from 375 to 4,370 ac (200 to 1,770 ha) in thenorthern Cascades mountains. Three years after thesefire, large erosional events occurred from spring rain-storms, not from snowmelt. Most of the sedimentproduced did not leave the burned area. Sartz (1953)reported an average soil loss of 1.5 in (37 mm) after awildfire on a north-facing slope in the Oregon Cascades.Raindrop splash and sheet erosion accounted for themeasured soil loss. Annual precipitation was 42 in(1070 mm), with a maximum intensity of 3.5 in hr–1

(90 mm hr–1). Vegetation covered the site within 1 yearafter the burn. Robichaud and Brown (1999) reportedfirst-year erosion rates after a wildfire from 0.5 to1.1 t ac–1 (1.1 to 2.5 Mg ha–1) decreasing by an order ofmagnitude by the second year, and to no sediment bythe fourth, in an unmanaged forest stand in easternOregon. DeBano and others (1996) found that follow-ing a wildfire in ponderosa pine, sediment yieldsfrom a low severity fire recovered to normal levelsafter 3 years, but moderate and severely burnedwatersheds took 7 and 14 years, respectively. Nearlyall fires increase sediment yield, but wildfires in steepterrain produce the greatest amounts (12 to 165 t ac–1,28 to 370 Mg ha–1) (table 5). Noble and Lundeen (1971)reported an average annual sediment production rateof 2.5 t ac–1 (5.7 Mg ha–1) from a 900 ac (365 ha) burnon steep river breaklands in the South Fork of the

Salmon River, Idaho. This rate was approximatelyseven times greater than hillslope sediment yieldsfrom similar, unburned lands in the vicinity.

Sediment Yield and Channel Stability—Fire-related sediment yields vary, depending on fire fre-quency, climate, vegetation, and geomorphic factorssuch as topography, geology, and soils (Swanson 1981).In some regions, over 60 percent of the total landscapesediment production over the long-term is fire-related.Much of that sediment loss can occur the first yearafter a wildfire (Agee 1993, DeBano and others 1996,DeBano and others 1998, Rice 1974, Robichaud andBrown 1999, Wohlgemuth and others 1998). Conse-quently, BAER treatments that have an impact thefirst year can be important in minimizing damage toboth soil and watershed resources.

After fires, suspended sediment concentrations instreamflow can increase due to the addition of ash andsilt-to-clay sized soil particles in streamflow. Highturbidity reduces municipal water quality and canadversely affect fish and other aquatic organisms. It isoften the most easily visible water quality effect offires (DeBano and others 1998). Less is known aboutturbidity than sedimentation in general because it isdifficult to measure, highly transient, and extremelyvariable.

A stable stream channel reflects a dynamic equilib-rium between incoming and outgoing sediment andstreamflow (Rosgen 1996). Increased erosion afterfires can alter this equilibrium by transporting addi-tional sediment into channels (aggradation). How-ever, increased peakflows that result from fires canalso produce channel erosion (degradation). Sedimenttransported from burned areas as a result of increasedpeakflows can adversely affect aquatic habitat, recre-ation areas, roads, buildings, bridges, and culverts.Deposition of sediments alters habitat and can fill inlakes and reservoirs (Rinne 1996, Reid 1993).

Mass Wasting—Mass wasting includes slope creep,rotational slumps, debris flows and debris avalanches.Slope creep is usually not a major postfire source of


Table 5—Published first-year sediment losses after prescribed fires and wildfires.

Location Treatment Sediment Loss Reference

(t ac–1) (Mg ha–1)

Mixed Conifer, WA Wildfire 130 300 Sartz 1953Mixed Conifer, WA Control 0.01 0.03 Helvey 1980

Wildfire 1 2

Mixed Conifer, WA McCay Wildfire 0.8 2 Radek 1996Bannon Wildfire 0.6 1Thunder Mtn. Wildfire 0.2 0.5Whiteface Wildfire 0.2 0.3

Ponderosa Pine, CA Control <0.0005 <0.001 Biswell and Schultz 1965Prescribed Fire <0.0005 <0.001

Chaparral, CA Control 0.02 0.04 Wells 1981Wildfire 13 30

Chaparral, CA Control 2 6 Krammes 1960Wildfire 25 60

Chaparral, CA Control, Steep Slope 0.0009 0.002 DeBano and Conrad 1976Rx Fire, Steep Slope 3 7Control, Gentle Slope 0 0Rx Fire, Gentle Slope 1 3

Chaparral, AZ Control 0 0 Pase and Lindenmuth 1971Prescribed Fire 2 4

Chaparral, AZ Control 0.04 0.1 Pase and Ingebo 1965Wildfire 13 29

Chaparral, AZ Control 0.07 0.2 Glendening and others 1961Wildfire 91 204

Ponderosa Pine, AZ Control 0.001 0.003 Campbell and others 1977Wildfire 0.6 1

Ponderosa Pine, AZ Wildfire, Low 0.001 0.003 DeBano and others 1996Wildfire, Moderate 0.009 0.02Wildfire, Severe 0.7 1.6

Mixed Conifer, AZ Control <0.0004 <0.001 Hendricks and Johnson 1944Wildfire, 43% Slope 32 72Wildfire, 66% Slope 90 200Wildfire, 78% Slope 165 370

Juniper-Grass, TX Control 0.03 0.06 Wright and others 1982Prescribed Fire 7 15Prescribed Fire, Seed 1 3

Juniper-Grass, TX Control 0.006 0.01 Wright and others 1976Burn, Level Slope 0.01 0.03Burn, 20% Slope 0.8 2Burn, 54% Slope 4 8

Larch/Douglas-fir, MT Control <0.0004 <0.001 Debyle and Packer 1972Slash Burned 0.07 0.2

Ponderosa-pine/Douglas-fir, ID Wildfire 4 6 Noble and Lundeen 1971

Ponderosa-pine/Douglas-fir, ID Clearcut and Wildfire 92 120 Megahan and Molitor 1975

Ponderosa-pine/Douglas-fir, OR Wildfire, 20% Slope 0.5 1.1 Robichaud and Brown 1999Wildfire, 30% Slope 1.0 2.2Wildfire, 60% Slope 1.1 2.5


sediment. Rotational slumps normally do not moveany significant distance. Slumps are only major prob-lems when they occur close to stream channels, butthey do expose extensive areas of bare soil on slopesurfaces. Debris flows and avalanches are the largest,most dramatic, and main form of mass wasting thatdelivers sediment to streams (Benda and Cundy 1990).They can range from slow moving earth flows to rapidavalanches of soil, rock, and woody debris. Debrisavalanches occur when the mass of soil material andsoil water exceed the sheer strength needed to main-tain the mass in place. Steep slopes, logging, roadconstruction, heavy rainfall, and fires aggravate de-bris avalanching potential.

Many fire-associated mass failures are correlatedwith development of water repellency in soils (DeBanoand others 1998). Chaparral vegetation in the South-western United States is a high hazard zone becauseof the tendency to develop water repellent soils. Waterrepellency also occurs commonly elsewhere in theWest after wildfires. Sediment delivery to channels bymass failure can be as much as 50 percent of the totalpostfire sediment yield. Wildfire in chaparral vegeta-tion in coastal southern California increased debrisavalanche sediment delivery from 18 to 4,845 yd3

mi–2 yr–1 (7 to 1,910 m3 km–2 yr–1) (Wells 1981).Cannon (1999) describes two types of debris flow

initiation mechanisms, infiltration soil slip and sur-face runoff after wildfires in the Southwestern UnitedStates. Of these, surface runoff which increases sedi-ment entrainment was the dominate triggeringmechanism.

Dry Ravel—Dry ravel is the gravity-induced down-slope surface movement of soil grains, aggregates, androck material, and is a ubiquitous process in semiaridsteepland ecosystems (Anderson and others 1959).Triggered by animal activity, earthquakes, wind, andperhaps thermal grain expansion, dry ravel may bestbe described as a type of dry grain flow (Wells 1981).Fires greatly alter the physical characteristics ofhillside slopes, stripping them of their protective coverof vegetation and organic litter and removing barriersthat were trapping sediment. Consequently, duringand immediately following fires, large quantities ofsurface material are liberated and move downslopeas dry ravel (Krammes 1960, Rice 1974). Dry ravel canequal or exceed rainfall-induced hillslope erosion afterfire in chaparral ecosystems (Krammes 1960,Wohlgemuth and others 1998).

Emergency Watershed RehabilitationTreatment Effectiveness

Early burned area emergency rehabilitation effortswere principally aimed at controlling erosion. Work byBailey and Copeland (1961), Christ (1934), Copeland

(1961, 1968), Ferrell (1959), Heede (1960, 1970), and Noble(1965) demonstrated that various watershed manage-ment techniques could be used on forest, shrub, andgrass watersheds to control both storm runoff anderosion. Many of these techniques have been refined,improved, and augmented from other disciplines (ag-riculture, construction) to form the set of BAER treat-ments in use today.

With the exception of grass seeding, relatively littlehas been published specifically on the effectivenessand ecosystem impacts of most postfire rehabilitationtreatments. We discuss the BAER literature by treat-ment categories: hillslope, channel, and road treat-ments. BAER treatments will be categorized in thismanner throughout this report.

Hillslope Treatments—Hillslope treatments in-clude grass seeding, contour-felled logs, mulch, andother methods intended to reduce surface runoff andkeep postfire soil in place on the hillslope. These treat-ments are regarded as a first line of defense againstpostfire sediment movement, preventing subsequentdeposition in unwanted areas. Consequently, moreresearch has been published on hillslope treatmentsthan on other methods.

Broadcast Seeding—The most common BAER prac-tice is broadcast seeding of grasses, usually from air-craft. Grass seeding after fire for range improvementhas been practiced for decades, with the intent to gainuseful products from land that will not return totimber production for many years (Christ 1934,McClure 1956). As an emergency treatment, rapidvegetation establishment has been regarded as themost cost-effective method to promote rapid infiltra-tion of water, keep soil on hillslopes and out of chan-nels and downstream areas (Miles and others 1989,Noble 1965, Rice and others 1965). Grasses are par-ticularly desirable for this purpose because their ex-tensive, fibrous root systems increase water infiltra-tion and hold soil in place. Fast-growing non-nativespecies have typically been used. They are inexpensiveand readily available in large quantities when anemergency arises (Agee 1993, Barro and Conard 1987,Miles and others 1989).

Legumes are often added to seeding mixes for theirability to increase available nitrogen in the soil afterthe postfire nutrient flush has been exhausted, aidingthe growth of seeded grasses and native vegetation(Ratliff and McDonald 1987). Seed mixes were refinedfor particular areas as germination and establishmentsuccess were evaluated. Most mixes contained annualgrasses to provide quick cover and perennials to estab-lish longer term protection (Klock and others 1975,Ratliff and McDonald 1987). However, non-native spe-cies that persist can delay recovery of native flora andpotentially alter local plant diversity. More recentlyBAER teams have recommended nonreproducing


annuals, such as cereal grains or sterile hybrids, thatprovide quick cover and then die out to let nativevegetation reoccupy the site.

Chaparral: Chaparral is the shrub-dominated veg-etation type abundant in the low to middle elevationfoothills in California and the Southwestern States(Cooper 1922, Keeley and Keeley 1988). Chaparralstands are often located on steep slopes, burn withgenerally high intensity, and typically develop water-repellent soils. They become candidates for postfireseeding due to the threat of increased runoff andsediment movement (Ruby 1989).

Concern over impacts of postfire seeding has focusedon chaparral ecosystems because a specialized annualflora takes advantage of the light, space, and soilnutrients available after fire (Keeley and others 1981,Sweeney 1956). Some of the dominant shrub speciesregenerate after fire only from seed (Keeley 1991,Sampson 1944). Most published research on chaparralcomes from California (tables 6 and 7).

Brushfields prone to fire and erosion occur at theurban/wildland interface, where growing populationcenters in lowland valleys have encroached on foot-hills and steep mountain fronts. The societal impactsof wildfire and subsequent accelerated erosion inCalifornia chaparral are enormous, as are the pres-sures to treat burned hillsides with grass seed toprotect life and property (Arndt 1979, Gibbons 1995).

Foresters in southern California began seedingburned-over slopes with native shrubs in the 1920’s.After finding that seeded shrubs emerged no earlierthan natural regeneration (Department of Foresterand Fire Warden 1985), they experimented with intro-duced herbaceous species such as Mediterraneanmustards in the 1930’s and 1940’s (Gleason 1947).Mustards proved to be unpopular weeds with downslopeorchardists and suburbanites, so other species weretested, including native and non-native subshrubsand non-native grasses (Department of Forester andFire Warden 1985). By the late 1940’s annual ryegrass(Lolium multiflorum, also called Italian ryegrass), anative of temperate Europe and Asia, had became theprimary species used for postfire seeding. Like mus-tard, it was inexpensive, could be broadcast easilyfrom aircraft, was available in large quantities, and itsfibrous root system appeared effective at stabilizingsurface soil (Barro and Conard 1987).

The effectiveness of broadcast grass seeding forerosion control on steep chaparral slopes has beenquestioned (Conrad 1979), but relatively few data onerosion response exist. The first watershed-scale reha-bilitation experiment was set up at the San DimasExperimental Forest after a wildfire in 1960, includ-ing annual and perennial grass seeding. The firstwinter after the fire was one of the driest on recordwith negligible grass establishment (Corbett and Green

1965). The treatments were reseeded, and the nextyear seeded grasses did not affect peak streamflowduring four recorded storm events. The high-rateannual grass treatment produced 8 percent grasscover by the time of the last large storm event andresulted in a 16 percent reduction in sediment produc-tion over the season (Krammes and Hill 1963). Con-tour planting of barley, which included hand-hoedrows and fertilization, had the greatest impact onsediment production (Rice and others 1965). All seededtreatments had lower cover of native plants thanunseeded controls (Corbett and Green 1965).

Data collected by the California Department ofForestry showed that ryegrass establishment wastypically poor in interior southern California and moresuccessful in cooler, northern or coastal locations(Blanford and Gunter 1972). An inverse relationshipbetween ryegrass cover and native herbaceous plantcover was observed, and Blanford and Gunter (1972)felt that more data were needed to properly evaluatethe competitive effects of seeded ryegrass on nativeherbs. Range improvement studies found that highseeded grass cover could reduce shrub seedling den-sity (Schulz and others 1955). Blanford and Gunter(1972) did not observe major failure of shrub regenera-tion, though no quantitative measurements were made.A general negative relationship between ryegrass coverand erosion was observed using erosion pins. Blanfordand Gunter (1972), like Krammes and Hill (1963) andRice and others (1965), concluded that postfire annualgrass seeding was an appropriate rehabilitationmethod because its low cost made occasional seedingfailure an acceptable risk.

Cover or biomass of native chaparral vegetation,especially herbaceous species, tended to be lower onplots with high ryegrass cover, both in operationallyseeded areas (Keeley and others 1981, Nadkarni andOdion 1986) and on hand-seeded experimental plots(Gautier 1983, Taskey and others 1989). Native plantspecies richness was lower on plots containing ryegrass(Nadkarni and Odion 1986, Taskey and others 1989).Gautier (1983) and Taskey and others (1989) foundlower density of shrub seedlings, especially specieskilled by fire, on seeded plots, and warned that long-term chaparral species composition could potentiallybe affected by grass seeding. Taskey and others (1989)also noted bare areas appearing in seeded plots whereryegrass died out after 3 years, resulting in lower coverthan on unseeded plots. These studies suggested thatryegrass grows at the expense of native vegetation.

During a year in which total rainfall was exception-ally high compared to average, Gautier (1983) mea-sured less erosion from plots in which ryegrass seed-ing increased total plant cover. On the other hand,Taskey and others (1989) found no effect of ryegrasson first-year postfire erosion with average rainfall and


Tab

le 6

—P

ublis

hed

first

-yea

r se

eded

gra

ss s

peci

es c

over

and

tota

l pla

nt c

over

in v

ario

us e

cosy

stem

s. E

rosi

on r

educ

tion

effe

ctiv

enes

s is

giv

en if

it w

as te

sted

in th

e st

udy.

Co

ver

Lo

cati

on

Veg

etat

ion

See

ded

sp

ecie

sS

lop

e-

- -

- -

- -

- -

Rat

e -

- -

- -

- -

- -

See

ded

To

tal

Ero

sio

n2

So

urc

e

(%)

(lb a

c–1)

(kg

ha–1

)(s

ft–2

)(s

m–2

) -

- -

- (

%)

- -

- -

Ota

y M

ts.,

So.

CA

mix

ed c

hapa

rral

annu

al r

yegr

ass

448.

09.

035

-80

40-8

5*31

% le

ssG

autie

r 19

83co

ntro

l44

00

—20

-35

San

ta L

ucia

Mts

., C

Ach

amis

e ch

apar

ral

annu

al r

yegr

ass

40-5

513

.117

3740

037

71.4

*ns

Tas

key

and

othe

rs 1

989

cont

rol

40-5

50

00

0—

57.7

San

ta A

na M

ts.,

CA

mix

ed c

hapa

rral

Bla

ndo

brom

end

37.

58.

43

15-2

5—

Con

ard

and

othe

rs 1

995

Zor

ro fe

scue

nd7.

58.

42-

520

-25

—an

nual

rye

gras

snd

7.5

8.4

3-5

18-2

5—

nativ

e fo

rb m

ixnd

7.5

8.4

3-8

25—

cont

rol

nd0

00

0—

12-3

0—

San

ta M

onic

a M

ts.,

CA

mix

ed c

hapa

rral

annu

al r

yegr

ass

548

912

30ns

Bey

ers

and

othe

rs 1

998a

;co

ntro

l54

00

00

—35

Woh

lgem

uth

and

othe

rs 1

998

San

ta A

na M

ts.,

CA

mix

ed c

hapa

rral

annu

al r

yegr

ass

658

95

20ns

sam

eco

ntro

l65

00

00

—20

San

ta L

ucia

Mts

., C

Ach

amis

e ch

apar

ral

annu

al r

yegr

ass

528

92

5*ns

sam

eco

ntro

l52

00

00

—3

San

ta M

onic

a M

ts.,

CA

mix

ed c

hapa

rral

annu

al r

yegr

ass

498

95

30ns

sam

eco

ntro

l49

00

00

—35

San

ta Y

nez

Mts

., C

Am

ixed

cha

parr

alan

nual

rye

gras

s63

89

1275

nssa

me

cont

rol

630

00

0—

75

Sis

kiyo

u M

ts.,

OR

Dou

glas

-fir

annu

al r

yegr

ass

40-5

024

2749

50*

nsA

mar

anth

us 1

989

—

early

win

ter

cont

rol

40-5

00

00

04

9

Sis

kiyo

u M

ts.,

OR

Dou

glas

-fir

annu

al r

yegr

ass

40-5

024

2785

.287

.1*

ns

—la

te s

prin

gco

ntro

l40

-50

00

00

8.2

23.6

near

Lom

an, I

DD

ougl

as-f

irse

ed m

ix5

nd8.

219.

248

520

114

—G

eier

-Hay

es 1

997

cont

rol

nd0

00

0—

15

near

Lom

an, I

Dpo

nder

osa

pine

seed

mix

5nd

4.39

4.8

2729

23

10—

sam

eco

ntro

lnd

—11

(con

.)


near

Lom

an, I

Dsu

balp

ine

firse

ed m

ix5

nd5.

366

3133

4tr

7—

sam

eco

ntro

lnd

—9

near

Gre

envi

lle, C

Am

ixed

con

ifer

seed

mix

6ge

ntle

tr6

—R

oby

1989

cont

rol

7

Ent

iat E

xp. F

ores

t, W

Apo

nder

osa

pine

- D

ougl

as-f

irse

ed m

ix7

nd5.

896.

63.

310

.3—

Tie

dem

ann

and

Klo

ck 1

973

seed

mix

+ fe

rt.

nd5.

896.

61.

4, 2

7.5,

10.

8—

cont

rol

nd0

00

0tr

5.6

Sno

w B

asin

, OR

pine

-mix

ed fi

rse

ed m

ix8

gent

le10

11.2

1244

*—

And

erso

n an

d B

rook

s 19

75co

ntro

lnd

00

00

—12

San

ta L

ucia

Mts

., C

Aco

ulte

r pi

ne-

annu

al r

yegr

ass

25-6

58

8.96

3740

05-

7010

-75

—G

riffin

198

2su

gar

pine

unse

eded

4nd

00

00

—5

1 Ast

eris

k (*

) ne

xt to

tota

l cov

er v

alue

indi

cate

s th

at s

eede

d pl

ots

had

sign

ifica

ntly

gre

ater

tota

l cov

er th

an u

nsee

ded

plot

s. S

tatis

tical

sig

nific

ance

was

not

test

ed in

all

stud

ies.

2 For

stu

dies

in w

hich

ero

sion

was

mea

sure

d, “n

s” m

eans

ther

e w

as n

o si

gnifi

cant

diff

eren

ce b

etw

een

seed

ed a

nd u

nsee

ded

plot

s; if

ero

sion

was

sig

nific

antly

less

, am

ount

of r

educ

tion

is g

iven

.3 “n

d” in

dica

tes

no d

ata

prov

ided

for

a gi

ven

cate

gory

in th

e pu

blic

atio

n.4 U

nsee

ded

plot

was

in s

light

ly d

iffer

ent a

rea,

avo

ided

dur

ing

seed

ing

beca

use

of r

are

plan

t con

cern

s.S

eed

mix

es in

clud

ed:

5 Inte

rmed

iate

whe

atgr

ass,

tim

othy

, orc

hard

gras

s, a

nd s

moo

th b

rom

e.6 O

rcha

rdgr

ass,

tall

fesc

ue, t

imot

hy, s

lend

er w

heat

gras

s +

fert

ilize

r.7

Har

d fe

scue

, orc

hard

gras

s, p

eren

nial

rye

gras

s, ti

mot

hy, y

ello

w s

wee

tclo

ver.

8 In

term

edia

te w

heat

gras

s, ti

mot

hy, h

ard

fesc

ue, l

egum

es.

Tab

le 6

(C

on.)

Co

ver

Lo

cati

on

Veg

etat

ion

See

ded

sp

ecie

sS

lop

e-

- -

- -

- -

- -

Rat

e -

- -

- -

- -

- -

See

ded

To

tal

Ero

sio

n2

So

urc

e

(%)

(lb a

c–1)

(kg

ha–1

)(s

ft–2

)(s

m–2

) -

- -

- (

%)

- -

- -


Tab

le 7

—P

ublis

hed

seco

nd-y

ear s

eede

d gr

ass

spec

ies

cove

r and

tota

l pla

nt c

over

in v

ario

us e

cosy

stem

s. E

rosi

on re

duct

ion

effe

ctiv

enes

s is

giv

en if

it w

as te

sted

in th

e st

udy.

Co

ver

Lo

cati

on

Veg

etat

ion

See

ded

sp

ecie

sS

lop

e-

- -

- -

- -

- -

Rat

e -

- -

- -

- -

- -

See

ded

To

tal

Ero

sio

n2

So

urc

e

(%)

(lb a

c–1)

(kg

ha–1

)(s

ft–2

)(s

m–2

) -

- -

- (

%)

- -

- -

San

Gab

riel M

ts.,

CA

mix

cha

parr

alan

nual

rye

gras

snd

320

22.4

9.9

10.3

16%

less

Ric

e an

d ot

hers

196

5B

land

o br

ome

cont

rol

00

—7.

7

San

ta A

na M

ts.,

CA

mix

cha

parr

alB

land

o br

ome

nd3

7.5

8.4

10-1

540

-60

—C

onar

d an

d ot

hers

199

5Z

orro

fesc

uend

7.5

8.4

3-8

45-5

5—

annu

al r

yegr

ass

nd7.

58.

45-

1545

-60

—na

tive

forb

mix

nd7.

58.

410

-20

45-6

0—

cont

rol

nd0

0—

35-5

0—

San

ta M

onic

a M

ts.,

CA

mix

cha

parr

alan

nual

rye

gras

s54

89

450

nsB

eyer

s an

d ot

hers

199

8a;

cont

rol

540

0—

48W

ohlg

emut

h an

d ot

hers

199

8

San

ta A

na M

ts.,

CA

mix

cha

parr

alan

nual

rye

gras

s65

89

3580

nssa

me

cont

rol

650

0—

78

San

ta L

ucia

Mts

., C

Ach

amis

e ch

apar

ral

annu

al r

yegr

ass

528

920

55*

nssa

me

cont

rol

520

0—

45

San

ta M

onic

a M

ts.,C

Am

ix c

hapa

rral

annu

al r

yegr

ass

498

925

78ns

sam

eco

ntro

l49

00

—80

San

ta Y

nez

Mts

., C

Am

ix c

hapa

rral

annu

al r

yegr

ass

638

915

65ns

sam

eco

ntro

l63

00

—70

Sis

kiyo

u M

ts.,

OR

Dou

glas

-fir

annu

al r

yegr

ass

40-5

024

270

live

91*

—A

mar

anth

us 1

989

(84

mul

ch)

cont

rol

40-5

00

0—

46

Sal

mon

Riv

er, C

Am

ixed

con

ifer

seed

mix

564

ndnd

85*

nsV

an d

e W

ater

199

8co

ntro

l64

00

—20

Sal

mon

Riv

er, C

Am

ixed

con

ifer

seed

mix

582

ndnd

60*

80%

less

sam

eco

ntro

l82

00

—40

near

Lom

an, I

DD

ougl

as-f

irse

ed m

ix6

nd8.

219.

248

520

847

—G

eier

-Hay

es 1

997

cont

rol

nd0

0—

47

near

Lom

an, I

Dpo

nder

osa

pine

seed

mix

6nd

4.3

4.8

3133

410

31—

sam

eco

ntro

lnd

—48

near

Lom

an, I

Dsu

balp

ine

firse

ed m

ix6

nd5.

36

231

—sa

me

cont

rol

nd—

34(c

on.)


Tab

le 7

(Con

.)

Co

ver

Lo

cati

on

Veg

etat

ion

See

ded

sp

ecie

sS

lop

e-

- -

- -

- -

- -

Rat

e -

- -

- -

- -

- -

See

ded

To

tal

Ero

sio

n2

So

urc

e

(%)

(lb a

c–1)

(kg

ha–1

)(s

ft–2

)(s

m–2

) -

- -

- (

%)

- -

- -

near

Gre

envi

lle, C

Am

ixed

con

ifer

See

d m

ix7

gent

lend

1024

nsR

oby

1989

cont

rol

—27

Ent

iat E

xp. F

or, W

Apo

nder

osa

pine

-D

ougl

as-f

irS

eed

mix

8nd

6.5

6.6

nd20

—T

iede

man

n an

d K

lock

197

3ab

ove

mix

+ fe

rt.

nd6.

56.

6nd

17-2

3—

cont

rol

nd0

0nd

16

Sno

w B

asin

, OR

pine

-mix

ed fi

rS

eed

mix

9nd

low

1011

.230

57—

And

erso

n an

d B

rook

s 19

75co

ntro

lnd

00

—49

San

ta L

ucia

Mts

., C

Aco

ulte

r pi

ne-

annu

al r

yegr

ass

25-6

58

937

400

30-9

075

-95

—G

riffin

198

2su

gar

pine

unse

eded

4nd

00

—25

1 Ast

eris

k (*

) ne

xt to

tota

l cov

er v

alue

indi

cate

s th

at s

eede

d pl

ots

had

sign

ifica

ntly

gre

ater

tota

l cov

er th

an u

nsee

ded

plot

s. S

tatis

tical

sig

nific

ance

was

not

test

ed in

all

stud

ies.

2 For

stu

dies

in w

hich

ero

sion

was

mea

sure

d, “n

s” m

eans

ther

e w

as n

o si

gnifi

cant

diff

eren

ce b

etw

een

seed

ed a

nd u

nsee

ded

plot

s; if

ero

sion

was

sig

nific

antly

less

, am

ount

of r

educ

tion

is g

iven

.3 “n

d” in

dica

tes

no d

ata

prov

ided

for

a gi

ven

cate

gory

in th

e pu

blic

atio

n.4 U

nsee

ded

plot

was

in s

light

ly d

iffer

ent a

rea,

avo

ided

dur

ing

seed

ing

beca

use

of r

are

plan

t con

cern

s.S

eed

mix

es in

clud

ed:

5 Ken

tuck

y bl

uegr

ass,

red

clo

ver,

cer

eal g

rain

.6 In

term

edia

te w

heat

gras

s, ti

mot

hy, o

rcha

rdgr

ass,

and

sm

ooth

bro

me.

7 Orc

hard

gras

s, ta

ll fe

scue

, tim

othy

, sle

nder

whe

atgr

ass

+ fe

rtili

zer.

8 Har

d fe

scue

, orc

hard

gras

s, p

eren

nial

rye

gras

s, ti

mot

hy, y

ello

w s

wee

tclo

ver

9 Inte

rmed

iate

whe

atgr

ass,

tim

othy

, har

d fe

scue

, leg

umes

.


no intense storms, despite higher average cover onseeded plots. Higher dry season erosion was measuredon seeded plots the following year, which was attrib-uted to pocket gophers attracted to the site by theabundant ryegrass. Similar densities of pocket gophermounds were found in operationally seeded areas(Taskey and others 1989).

The most extensive study of annual ryegrass effectson erosion and vegetation response was conducted onfive sites burned in hot prescribed fires and a wind-driven wildfire in coastal southern California (Beyersand others 1998a, 1998b; Wohlgemuth and others1998). Data on prefire vegetation and hillslope sedi-ment movement were gathered, and greater replica-tion was used than in most previous studies. Fireseverity varied among sites from moderate to veryhigh, and postfire precipitation varied from half ofnormal to very high. Only plots that showed severityeffects great enough to trigger operational seedingwere retained in the study. At all five sites, postfireerosion was greatest during the first year after fire andwas not significantly affected by ryegrass seeding(Wohlgemuth and others 1998). Seeding increasedtotal plant cover the first year at only one site, by about1.5 percent, probably accounting for the lack of differ-ence in erosion rates (Beyers and others 1998a). Aver-age ryegrass cover reached 15 to 30 percent on somesites during the second year after fire. Native herba-ceous plant cover and species richness were lower onseeded plots when ryegrass cover was high (Beyersand others 1994, 1998b). Unlike some earlier studies,Beyers and others (1998a) did not find significantlylower shrub seedling density on seeded plots. In laterpostfire years, some sites had significantly less ero-sion on seeded than on unseeded plots, but this hap-pened only after erosion rates had dropped to prefirelevels, which occurred in as little as 2 years on somesites (Wohlgemuth and others 1998). Dry season ero-sion (ravel) accounted for a high proportion of first-year sediment movement on sites that burned duringearly or mid summer. Grass seeding does not affect thechannel loading that occurs by this process (Beyersand others 1998b, Wohlgemuth and others 1998).These studies concluded that postfire annual ryegrassseeding is unlikely to reduce postfire hillslope sedi-ment movement the first year after fire in southernCalifornia chaparral and has minimal impact on totalerosion from a burn site.

Grass species other than annual ryegrass havebeen used for postfire rehabilitation. Blando brome(Bromus hordaceous cv “Blando”), promoted for usein drought-prone areas, did not produce cover as wellas annual ryegrass (Blanford and Gunter 1972).Conard and others (1995) tested several non-nativegrasses and a native forb mix; only the native forbmix significantly increased total plant cover, and

then only on a north-facing slope. After the 1993firestorms in southern California, Keeley and others(1995) found complete failure where native perennialneedlegrass (Nasella) species were used, and rela-tively low levels of grass cover (1 to 23 percent)produced by non-native annuals such as Zorro fescue(Vulpia myuros cv “Zorro”) and Blando brome, used toavoid the competitive problems associated with an-nual ryegrass. The highest seeded cover, 40 percent,occurred on a site seeded with a mix of native speciesand annual ryegrass. However, natural regenerationof native and naturalized plants provided much morecover than the seeded species. Although no directerosion measurements were made, Keeley and others(1995) concluded that seeding was ineffective as asediment control measure in the cases examinedbecause it contributed very little to total plant cover.

No quantitative studies on the impact of grass seed-ing on postfire erosion in chaparral have been pub-lished from northern California or Arizona. Becauseannual ryegrass and other grasses typically establishcover more successfully in northern California (Barroand Conard 1987, Blanford and Gunter 1972), theywould be more likely to reduce erosion there. Theimpact of grass seeding on native chaparral vegetationin other areas, aside from suppression of shrub seed-lings at very high grass densities (Schultz and others1955), is largely unknown.

Conifer Forest: High intensity fire may be outsidethe range of natural variability for many conifer plantcommunities that have been subject to fire suppres-sion for the last century (Agee 1993). The loss of formerunderstory seed banks due to overgrazing and canopydensification may also reduce the likelihood of rapidregeneration of ground cover after fire. Seeding mixesused in conifer stands often include legumes such aswhite clover (Trifolium repens) or yellow sweet clover(Melilotus officinalis) to enhance nitrogen status ofthe soil. Both annual and perennial grasses may beused in mixes with non-native forage species origi-nally tested for range improvement purposes (Christ1934, Forsling 1931, McClure 1956).

Orr (1970) examined plant cover and erosion for3 years after fire in the Black Hills of South Dakotain an area operationally seeded with a mixture ofgrasses and legumes. Most of the sediment productionoccurred in two summer storms shortly after erosion-measuring apparatus was set up. Sediment produc-tion was inversely related to plant cover. Summerstorm runoff was 50 percent less on plots with highplant and litter cover than on those with sparse cover.Regression analysis showed that the decrease in run-off and sediment production with increasing groundcover leveled off at 60 percent cover, similar to resultspresented by Noble (1965). Orr (1970) concluded thatseeded species were essential for quickly stabilizing


the sites. However, unseeded plots were not includedin the study.

Seeded grasses provided greater cover than naturalregeneration in a burned area in Oregon (Andersonand Brooks 1975). Litter and mulch also developedmore rapidly on the seeded sites. After 4 years, how-ever, all sites had more than 70 percent ground cover.Legume species included in the seeding mix forwildlife forage generally did not survive. Seeded grassesappeared to suppress growth of native shrubs andannual forbs, particularly in the second and thirdyear after fire. Erosion amounted to only 5 t ac–1

(5.5 Mg ha–1) during the first 2 years after fire onseeded sites. The unseeded site was not measured butalso appeared to experience little erosion (Andersonand Brooks 1975).

In contrast, Dyrness (1976) measured negligiblecover produced by seeded species on severely burnedplots in Oregon. Total vegetation cover was only 40percent after 2 years even on lightly burned sites. Hesuggested that nitrogen fertilization might have im-proved vegetation growth. Earlier work by Dyrness(1974) found that grass vigor decreased 4 years afterseeding along forest roads for erosion control, andrefertilization in year 7 reinvigorated perennial grassesin the plots. On disturbed firelines, Klock and others(1975) seeded various grasses and legumes andfound that fertilization greatly increased initial coverof most species tested. Fertilization with 45 lb ac–1

(50 kg ha–1) drilled urea significantly increased nativeplant regrowth, but not production of seeded species,on granitic soil in Idaho (Cline and Brooks 1979).

Seeding and fertilizer treatments were compared onseparate watersheds in the Washington Cascadesafter a fire swept through the Entiat ExperimentalForest (Tiedemann and Klock 1973). Seeding increasedplant cover at the end of the first growing season byabout one third, from 5.6 percent on the unseededwatershed to 7. 5 to 10.8 percent on the seeded water-sheds. Seeded grasses made up 18 to 32 percent of totalcover on seeded sites. Nitrate concentration in streamsincreased immediately after fertilizer application, butsubsequently fertilized and unfertilized watershedshad similar stream nitrogen dynamics (Tiedemannand others 1978). Later that summer, record rainfallevents caused massive flooding and debris torrentsfrom treated and untreated watersheds alike (Helvey1975). In the second year after fire, average total plantcover increased to 16.2 percent on the unseeded water-shed and 16.4 to 23 percent on the seeded watersheds.Seeded grasses comprised about 7 percent cover onseeded watersheds (Tiedemann and Klock 1976). Onsouth-facing slopes, the unseeded watershed had asmuch or more cover than the seeded ones. Althoughfertilization did not affect plant cover either year,Tiedemann and Klock (1976) felt that it increasedseeded grass vigor and height.

From an erosion standpoint during the first winterafter fire, the amount of seeded grass present at thetime major storms occur is more important than theamount present at the end of the growing season,when it is usually assessed in studies. In southernOregon, annual ryegrass seeding and fertilizationdid not significantly increase plant cover or reduceerosion by early December, when that winter’s majorstorms occurred (Amaranthus 1989). The seeded andfertilized plots had significantly less bare groundthan the unseeded plots. Erosion was low and notsignificantly different between treatments, though ittrended lower on the seeded plots. Amaranthus (1989)pointed out that timing of rainfall is critical to bothgrass establishment and erosion, and that differentrainfall patterns could have produced different re-sults from the study.

In contrast, grass seeding plus fertilizer did notsignificantly increase total plant cover during the first5 years after a northern Sierra Nevada fire (Roby1989). Seeded grass cover did not exceed 10 percentuntil 3 years after the fire, when total cover on unseededplots was greater than 50 percent. There was nodifference in erosion between the seeded and unseededwatersheds during the first 2 years after fire. Roby(1989) concluded that grass seeding was ineffective asa ground cover protection measure in that location.Geier-Hayes (1997) also found that total plant coverdid not differ between seeded and unseeded plots for5 years after an Idaho fire. Seeded plots had lowercover of native species. Erosion was not measured.

Several species commonly used for postfire seeding,because of their rapid growth and wide adaptability(Klock and others 1975), have been found to be stronglycompetitive with conifer seedlings in experimentalplots. Orchardgrass (Dactylis glomerata), perennialryegrass (Lolium perenne), and timothy (Phleumpratense) reduced growth of ponderosa pine seedlingsin tests conducted in California (Baron 1962).Orchardgrass and crested wheatgrass (Agropyron de-sertorum) reduced ponderosa pine growth in Arizona(Elliot and White 1987). Field studies on aerialseeded sites in California found low pine seedlingdensities on most plots with annual ryegrass coverhigher than 40 percent (Conard and others 1991,Griffin 1982).

Amaranthus and others (1993) reported significantlylower survival of planted sugar pine (Pinus lam-bertiana) seedlings in plots heavily seeded with an-nual ryegrass than in unseeded controls during thefirst postfire year in southern Oregon. Soil moisturewas significantly lower and pine seedlings showedsignificantly greater water stress in the seeded plots.Ryegrass cover was 49 percent when tree seedlingswere planted and 85 percent by mid-summer, whiletotal plant cover was only 24 percent at mid-summeron the control plots. The next summer, a second group


of planted pine seedlings had significantly greatersurvival and lower water stress on seeded plots thanon controls. By then, dead ryegrass formed a densemulch on the seeded plots, but no live grass was found.Native shrub cover was significantly greater on theunseeded plots the second year and soil moisture waslower (Amaranthus and others 1993). Ryegrass thusacted as a detrimental competitor to tree seedlings thefirst year after fire, but provided a beneficial mulchand reduced competition from woody plants the sec-ond year. Conard and others (1991) also suggestedthat seeded ryegrass could benefit planted coniferseedlings if it suppressed woody competitors andcould itself later be controlled. In their study, how-ever, live ryegrass cover was exceptionally high inmany plots during the second year after fire (Conardand others 1991).

The studies examined suggest that grass seedingdoes not assure increased plant cover during the firstcritical year after fire (table 6). A wide variety ofgrass species or mixes and application rates wereused in the reported studies, making generalizationdifficult. Over 50 years ago, southern California for-esters were urged to caution the public not to expectsignificant first-year sediment control from postfireseeding (Gleason 1947). Better cover and, conse-quently, erosion control can be expected in the second(table 7) and subsequent years.

Measuring erosion and runoff is expensive, complex,and labor-intensive, and few researchers have done it.Such research is necessary to determine if seededgrasses control erosion better than natural regenera-tion. Another goal of postfire grass seeding on timbersites, soil fertility retention, does not appear to havebeen investigated. Grass establishment can clearlyinterfere with native plant growth, and grass varietiesthat will suppress native shrubs but not conifer seed-lings have not yet been developed (Ratliff and McDonald

1987). The impacts of recent choices for rehabilitationseeding, including native grasses and cereal grains, onnatural and planted regeneration in forest lands havenot been studied extensively.

Mulch—Mulch is material spread over the soil sur-face to protect it from raindrop impact. Straw mulchapplied at a rate of 0.9 t ac–1 (2 Mg ha–1) significantlyreduced sediment yield on burned pine-shrub forest inSpain over an 18-month period with 46 rainfall events(Bautista and others 1996). Sediment production was0.08 to 1.3 t ac–1 (0.18 to 2.92 Mg ha–1) on unmulchedplots but only 0.04 to 0.08 t ac–1 (0.09 to 0.18 Mg ha–1)on mulched plots. Kay (1983) tested straw mulch laiddown at four rates—0.5, 1, 1.5, and 4 t ac–1 (1.1, 2.2,3.4, and 9.0 Mg ha–1)—against jute excelsior, andpaper for erosion control. Straw was the most cost-effective mulch, superior in protection to hydraulicmulches and comparable to expensive fabrics. Excel-sior was less effective but better than paper stripsynthetic yarn. The best erosion control came fromjute applied over 1.5 t ac–1 (3.4 Mg ha–1) straw. Milesand others (1989) studied the use of wheat strawmulch on the 1987 South Fork of the Trinity River fire,Shasta-Trinity National Forest in California. Wheatstraw mulch was applied to fill slopes adjacent to peren-nial streams, firelines, and areas of extreme erosionhazard. Mulch applied at rates of 2 t ac–1 (4.5 Mg ha–1),or 1 t ac–1 (2.2 Mg ha–1) on larger areas, reduced erosion6 to 10 yd3 ac–1 (11 to 19 m3 ha–1). They consideredmulching to be highly effective in controlling erosion(table 8). Edwards and others (1995) examined theeffects of straw mulching at rates of 0.9,1,8, 2.7, and3.6 t ac–1 (2,4,6, and 8 Mg ha–1) on 5 to 9 percent slopes.Soil loss at 0.9 t ac–1 (2 Mg ha–1) mulch was signifi-cantly greater (1.4 t ac–1, 3.16 Mg ha–1 of soil) than at1.8 t ac–1 (4 Mg ha–1) mulch (0.9 t ac–1, 1.81 Mg ha–1 ofsoil loss). Above 1.8 t ac–1 (4 Mg ha–1) mulch there wasno further reduction in soil loss.

Table 8—Comparison of slope and channel BAER treatments, South Fork Trinity River fires, Shasta-Trinity National Forest, CA,1987 (modified from Miles et al. 1989). Costs are shown in 1999 dollars.

Cost Efficacy InstallTreatment Type ($ yd–3) ($ m–3) ($ ac–1) ($ ha–1) Category Rate Risk of Failure

- - - - - - - - - - - - - - $$1999 - - - - - - - - - - - - -Slope Treatment Summary

Aerial Seeding $23 $23 $79 $196 Moderate1 Rapid ModerateMulching $50 $52 $504 $1245 High2 Slow LowContour Felling $180 $183 $720 $1778 Low2 Slow High

Channel Treatment SummaryStraw Bale Check Dams $105 $107 $158 $392 High2 High LowLog and Rock Check Dams $33 $33 $1346 $3325 High2 Slow Moderate

1Soil loss estimated using Universal Soil Loss Equation (USLE).2Soil loss estimated using on-site measurements.


Contour-Felled Logs—This treatment involves fell-ing logs on burned-over hillsides and laying them onthe ground along the slope contour, providing me-chanical barriers to water flow, promoting infiltrationand reducing sediment movement; the barriers canalso trap sediment. The terms “log erosion barriers” or“log terracettes” are often used when the logs arestaked in place and filled behind. Logs were contour-felled on 22 ac (9 ha) of the 1979 Bridge Creek Fire,Deschutes National Forest in Oregon (McCammonand Hughes 1980). Trees 6 to 12 in (150-300 mm) d.b.h.were placed and secured on slopes up to 50 percentat intervals of 10 to 20 ft (3 to 6 m). Logs were stakedand holes underneath were filled. After the first stormevent, about 63 percent of the contour-felled logs werejudged effective in trapping sediment. The remainderwere either partially effective or did not receive flow.Nearly 60 percent of the storage space behind con-tour-felled logs was full to capacity, 30 percent washalf-full, and 10 percent had insignificant deposition.Common failures were flow under the log and notplacing the logs on contour (more than 25° off contourcaused trap efficiency to decrease to 20 percent). Over1,600 yd3 (1,225 m3) of material was estimatedtrapped behind contour-felled logs on the 22 ac, orabout 73 yd3 ac–1 (135 m3 ha–1). Only 1 yd3 (0.7 m3) ofsediment was deposited in the intake pond for amunicipal water supply below. Miles and others (1989)monitored contour-felling on the 1987 South ForkTrinity River fires, Shasta-Trinity National Forest inCalifornia. The treatment was applied to 200 ac (80 ha)within a 50,000 ac (20,240 ha) burned area. Trees<10 in (250 mm) d.b.h. spaced 15 to 20 ft (4.5 to 6 m)apart were felled at rate of 80-100 trees ac–1 (200-250 trees ha–1). The contour-felled logs trapped 0 to0.07 yd3 (0 to 0.05 m3) of soil per log, retaining 1.6 to6.7 yd3 ac–1 (3 to 13 m3 ha–1) of soil onsite. Miles andothers (1989) considered sediment trapping efficiencylow and the cost high for this treatment (table 8).Sediment deposition below treated areas was notmeasured, however.

Contour Trenching—Contour trenches have beenused as a BAER treatment to reduce erosion andpermit revegetation of fire-damaged watersheds. Al-though they do increase infiltration rates, the amountsare dependent on soils and geology (DeByle 1970b).Contour trenches can significantly improve revegeta-tion by trapping more snow, but they don not affectwater yield to any appreciable extent (Doty 1970,1972). This BAER treatment can be effective in alter-ing the hydrologic response from short duration, highintensity storms typical of summer thunderstorms,but does not significantly change the peakflows of lowintensity, long duration rainfall events (DeByle1970a). Doty (1971) noted that contour trenching inthe sagebrush (Artemisia spp.) portion (upper 15

percent with the harshest sites) of a watershed incentral Utah did not significantly change streamflowand stormflow patterns. The report by Doty (1971) didnot discuss sediment. Costales and Costales (1984)reported on the use of contour trenching on recentlyburned steep slopes (40 to 50 percent) with clay loamsoils in pine stands of the Philippines. Contour trench-ing reduced sediment yield by over 80 percent, from28 to 5 t ac–1 (63 to 12 Mg ha–1).

Other Hillslope Treatments—Treatments such astilling, temporary fencing, installation of erosioncontrol fabric, use of straw wattles, lopping andscattering of slash, and silt fence construction areused to control sediment on the hillslopes. No pub-lished quantitative information is available aboutthe efficiency and sediment trapping ability of thesetreatments after wildfires.

Channel Treatments—Channel treatments areimplemented to modify sediment and water move-ment in ephemeral or small-order channels, to preventflooding and debris torrents that may affect down-stream values at risk. Some in-channel structuresslow water flow and allow sediment to settle out;sediment will later be released gradually as the struc-ture decays. Channel clearing is done to remove largeobjects that could become mobilized in a flood. Muchless information has been published on channel treat-ments than on hillslope methods.

Straw Bale Check Dams—Miles and others (1989)reported on the results of installing 1300 straw balecheck dams after the 1987 South Fork Trinity Riverfires, Shasta-Trinity National Forest, California. Mostdams were constructed with five bales. About 13 per-cent of the straw bale check dams failed due to pipingunder or between bales or undercutting of thecentral bale. Each dam stored an average 1.1 yd3

(0.8 m3) of sediment. They felt that filter fabric on theupside of each dam and a spillway apron would haveincreased effectiveness. They considered straw balecheck dams easy to install and highly effective whenthey did not fail (table 8). Collins and Johnston (1995)evaluated the effectiveness of straw bales on sedi-ment retention after the Oakland Hills fire. About5000 bales were installed in 440 straw bale checkdams and 100 hillslope barriers. Three months afterinstallation, 43 to 46 percent of the check dams werefunctioning. This decreased to 37 to 43 percent by4.5 months, at which time 9 percent were side cut, 22percent were undercut, 30 percent had moved, 24 per-cent were filled, 12 percent were unfilled, and 3 percentwere filled but cut. Sediment storage amounted to55 yd3 (42 m3) behind straw bale check dams andanother 122 yd3 (93 m3) on an alluvial fan. Goldmanand others (1986) recommended that the drainagearea for straw bale check dams be kept to less than


20 ac (8 ha). Bales usually last less than 3 months, flowshould not be greater than 11 cfs (0.3 m3 s–1), and balesshould be removed when sediment depth upstream isone-half of bale height. More damage can result fromfailed barriers than if no barrier were installed(Goldman and others 1986).

Log Check Dams—Logs 12 to 18 in (300 to 450 mm)diameter were used to build 14 log check dams thatretained from 1.5 to 93 yd3 (mean 29 yd3) (1.1 to 71 m3,mean 22 m3) of sediment after the 1987 South ForkTrinity River fires on the Shasta-Trinity NationalForest, California (Miles and others 1989). While logcheck dams have a high effectiveness rating and 15 to30 year life expectancy (Miles and others 1989), theyare costly to install (table 8).

Rock Dams and Rock Cage Dams (Gabions)—Prop-erly designed and installed rock check dams and rockcage (gabion) dams are capable of halting gully devel-opment on fire-disturbed watersheds, and reducingsediment yields by 60 percent or more (Heede 1970,1976). Although these structures are relatively ex-pensive, they can be used in conjunction with vegeta-tion treatments to reduce erosion by 80 percent andsuspended sediment concentrations by 95 percent(Heede 1981). While vegetation treatments such asgrassed waterways augment rock check dams and areless expensive, their maintenance costs are consider-ably greater. Check dams constructed in Taiwan water-sheds with annual sediment yields of 10 to 30 yd3 ac–1

(19 to 57 m3 ha–1) filled within 2 to 3 years. Sedimentyield rates decreased upstream of the check dams,but were offset by increased scouring downstream(Chiun-Ming 1985).

Other Channel Treatments—No published infor-mation was found on the effectiveness of strawwattle dams, log grade stabilizers, rock grade stabi-lizers, in-channel debris basins, in-channel debrisclearing, stream bank armoring or other BAER chan-nel treatments.

Road Treatments—BAER road treatments con-sist of a variety of practices aimed at increasing thewater and sediment processing capabilities of roadsand road structures, such as culverts and bridges, inorder to prevent large cut-and-fill failures and themovement of sediment downstream. The functionalityof the road drainage system is not affected by fire, butthe burned-over watershed can affect the functional-ity of that system. Road treatments include outsloping,gravel on the running surface, rocks in ditch, culvertremoval, culvert upgrading, overflows, armored streamcrossings, rolling dips, and water bars. The treat-ments are not meant to retain water and sediment, butrather to manage water’s erosive force. Trash racksand storm patrols are aimed at preventing culvert

blockages due to organic debris, which could result inroad failure that would increase downstream flood orsediment damage.

Furniss and others (1998) developed an excellentanalysis of factors contributing to the failure ofculverted stream crossings. Stream crossings are veryimportant, as 80 to 90 percent of fluvial hillslopeerosion in wildlands can be traced to road fill failuresand diversions of road-stream crossings (Best andothers 1995). Since it is impossible to design and buildall stream crossings to withstand extreme stormflows,they recommended increasing crossing capacity anddesigning to minimize the consequences of culvertexceedence as the best approaches for forest roadstream crossings.

Comprehensive discussions of road-related treat-ments and their effectiveness can be found in Packerand Christensen (1977), Goldman and others (1986)and Burroughs and King (1989). Recently the USDAForest Service, San Dimas Technology and Develop-ment Program has developed a Water/Road Interac-tion Technologies Series (Copstead 1997), which cov-ers design standards, improvement techniques, andevaluates some surface drainage treatments for re-ducing sedimentation.

Methods _______________________This study was restricted to USDA Forest Service

BAER projects in the Western continental UnitedStates (Regions 1 through 6). We began by requestingBurned Area Report (FS-2500-8) forms and monitor-ing reports from the Regional headquarters and For-est Supervisors’ offices. Our initial efforts revealedthat information collected on the Burned Area Reportforms and in the relatively few existing postfiremonitoring reports was not sufficient to assess treat-ment effectiveness, nor did it capture the informationknowledge of BAER specialists. Therefore, we designedinterview questions to enable us to rank treatmenteffectiveness, determine aspects of the treatmentsthat lead to success or failure, and allow for commentson various BAER related topics.

Burned Area Report Data

The Forest Service Burned Area Report form con-tains the fire name, watershed location, size, suppres-sion cost, vegetation, soils, geology, and lengths ofstream channels, roads, and trails affected by the fire.The watershed description includes areas in low,moderate, and high severity burn categories and areasthat have water repellent soils. Erosion hazard ratingand estimates of erosion potential and sedimentdelivery potential are included, based on specifieddesign storms. The probability of success for hillslope,


channel, and road treatments are provided. Cost esti-mates of no action (loss) versus cost of selected alter-natives are identified, as well as BAER funds re-quested and other matching funds. This informationwas entered directly into the database.

Interview Survey

Interview forms were developed after consultationwith several BAER specialists. The forms were used torecord information when we interviewed BAER teammembers, regional and national leaders. Questionswere designed to address specific BAER projects (i.e.,individual fires), as well as to elicit opinions regardingthe interviewees’ experience with treatments used ontheir forests and other fires they had worked on. Priorto conducting interviews, information such as BurnedArea Report forms and postfire monitoring reportswas requested to familiarize the interviewer with thevarious fires and treatment used. Onsite interviewswere conducted because much of the supporting datawere located in the Supervisor’s and District’s officesand could be retrieved during the interviews. At-tempts were made to ask questions that would allowfor grouping and ranking results, because much of theinformation was qualitative. Example interviewforms are included in appendix A.

Project Review Interview Form—Questions weredesigned to identify the fire size, area treated, andtreatment. The values at risk (i.e., downstream oronsite) were identified, and questions were askedwhether the site was tested by a significant stormevent and what damages resulted. We also askedinterviewees to list up to three treatments they feltwere overused, and up to three that in hindsightshould have been used more, on specific BAER projects.Cumulative ratings were determined by totaling thenumber of times each treatment was mentioned.

No Action Review Interview Form—For fireswhere no BAER action was recommended, intervieweeswere asked to identify the rationale used. They werealso asked if the site was tested by a significant stormand their opinion about what treatments might havebeen beneficial in hindsight.

Treatment Actions Interview Form—Thesequestions identified treatments used on specific firesand what environmental factors affected success andfailure. Interviewees were also asked questions re-garding implementation of treatments and whetherany monitoring was completed. For cases where moni-toring was conducted (either formal or informal),interviewees were asked to describe the type andquality of the data collected (if applicable) and to givean overall effectiveness rating of “excellent”, “good”,“fair”, or “poor” for each treatment. Because many of

the answers were qualitative, we synthesized theresponses, highlighting the major points made foreach treatment. We summarized this information intoparagraphs on effectiveness factors, implementationand environmental factors, and other factors whenthey occurred (appendix B).

Interview forms were developed for individual hill-slope treatments such as aerial seeding, ground seed-ing, fertilizer, mulch, contour felling, straw wattles,lop and scatter, silt fences, contour trenching, ripping,tilling, temporary fencing and erosion control fabric.Channel treatment forms included straw bale checkdams, log grade stabilizers, rock grade stabilizers, logdams, in-channel debris basins, in-channel debris clear-ing, stream bank armoring, rock cage (gabion) dams,and straw wattle dams. Road treatment forms in-cluded road regrading (such as out-sloping), rock inditches, culvert removal, culvert upgrades, overflows,trash racks, armored stream crossing, storm patrol,and rolling dips and water bars. For each treatment,specific question were asked regarding the factorsthat caused the treatment to succeed or fail, such asslope classes, soil type, and type of areas treated, aswell as appropriate implementation method questionsfor each treatment.

Relative Benefits Interview Form—Intervieweeswere asked to rank hillslope, channel, and road andtrail treatments for the three most effective treat-ments in each category. Then they were asked forthree overall treatments that provide the greatestbenefits. To obtain cumulative rankings, we totaledthe number of first, second and third place “votes” foreach treatment, multiplied by 3 for first, 2 for second,and 1 for third, then added the adjusted totals to yielda cumulative preference rating. Final questions wereopen-ended to provide an opportunity for programrecommendations or other topics not addressed.

Monitoring Reports

Monitoring reports were requested from Region,Forest, and District offices. We included administra-tive trip reports, data collection efforts, and regionalburn area rehabilitation activity reviews in our re-quest. We also examined BAER accomplishment re-ports, when provided, for initial post-treatment moni-toring results.

Analysis Methods

Burned Area Reports and Interview Forms—Burned Area Report data and interview informationwere entered into the commercial Microsoft Accessdatabase management system. Categorical informa-tion (such as treatments that were over-used or under-used) was left unchanged. Ranked information results


were given a 1 to 3 value with the first rankingreceiving three points, second ranking receiving twopoints and the third receiving one point. Several ques-tions had positive or negative effects response options.Qualitative answers were grouped into categories toreduce the data to a manageable amount. Correlationanalysis and categorical t-tests were performed onselected information in the data.

BAER spending and treatment costs were trans-formed into similar units (i.e., hectares or acres) andadjusted for inflation based on consumer price index to1999 dollars (Federal Reserve Bank 1999). This mademeaningful comparisons possible for analyzingspending trends. Treatment costs were obtained fromthe final Burned Area Report forms and were assignedto the year of the fire.

Monitoring Reports—Because most of the infor-mation in monitoring reports was qualitative in na-ture, excerpts from reports were entered into thedatabase referenced to specific fires. Other excerptswere included in the general comment fields. Quan-titative information was tabulated by hand sepa-rately from the main database because of its diversenature.

Results ________________________

Overview of Data Collected

Data were collected from 470 Burned Area Reportsand 98 interviews. The results represent our bestestimate of the types and amounts of BAER treat-ments used and their attributes for the past 3 decadesin the Forest Service. However, we were not able tocollect all possible Burned Area Reports. Regions 1and 3 are nearly complete data sets, whereas Re-gions 2, 4, 5, and 6 have missing results, especiallyfrom the 1970’s and 1980’s, because materials hadbeen archived and could not easily be accessed. There-fore, all dollar and area totals reported are at bestminimum estimates.

While our goal was to collect information on BAERtreatment effectiveness, we also acquired a vast data-base of information on BAER project and no-actionfires from the Burned Area Reports. These report dataallowed us to tabulate and examine the various piecesof information that make up the BAER evaluation.

Over the past 3 decades, more than $110 million wasspent in total on emergency rehabilitation that in-volved the Forest Service. Of that, about $83 millioncame from National Forest Systems (NFS) to treat4.6 million ac (1.9 million ha) of a total of 5.4 million ac(2.2 million ha) from BAER project fires. About 72percent of the total area treated was National ForestSystem lands. The remainder was on other Federalagency, State, and private lands.

Of the 470 fires for which Burned Area Reports wereprepared, 321 had BAER treatments recommended.The rest (148) were fires for which no emergency wasidentified and no BAER treatment requested. Seventytwo of the fires were less than 1,000 ac (400 ha), 153fires were between 1,000 to 10,000 ac (400 to 4,050 ha),and 96 were greater than 10,000 ac (4,050 ha).

Expenditures for BAER treatments have increasedsubstantially, especially during the 1990’s (fig. 1).There were several large fires that represent a major-ity of the spending in the 1990’s ($48 million), includ-ing the Rabbit Creek, Foothills, and Eighth Streetfires on the Boise National Forest in Idaho and theTyee Creek Complex on the Wenatchee National For-est in Washington (table 9). Regions 4, 5, and 6 ac-counted for 86 percent of the BAER spending from1973 to 1998 (fig. 2). Total acres burned by year (fig. 3)shows a trend similar to that for spending especially inthe 1990’s. In terms of cost per acre burned, the big fireyears do not always coincide with the greatest amountper acre (hectare) spent on BAER treatments. In 1989for example, an average of $67 ac–1 ($165 ha–1) wasspent on 55,000 National Forest System ac (22,300 ha)burned. When 616,000 National Forest System ac(249,00 ha) burned in 1996, only $16 ac–1 ($40 ha–1)was spent (fig. 4).

Fire Severity

Part of the Burned Area Report form contains infor-mation on percent of the total burned area in low,medium, and high fire “intensity.” However, BAER

Figure 1—BAER spending by National Forests and other stateand private entities that include National Forests by year in1999 dollars. The insert shows spending by decade as apercent of the total spending. Spending authority changes areshown.

0

5

10

15

20

25

1973

1974

1975

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

Mill

ion

s

YEAR

BA

ER

SP

EN

DIN

G (

19

99

$)

NFS OTHER

NFS BAER SPENDING

8%

31% 61%

1970's 1980's 1990's

Not authorized toexceed $2 million

per year

Authorized toexceed $2 million

per year


Figure 2—National Forest BAER spending by Region in 1999 dollars, 1973-1998 from Burned Area Reports.The insert shows the Western Forest Service Regions used in this study.

teams actually evaluate burn severity, not intensity(DeBano and others 1998), and hereafter we use theterm “severity” instead of intensity. The Burned AreaReport form burn severity information was used tocalculate the total acreage in the Western UnitedStates of wildfire-burned lands, by National ForestSystem Region, in high, moderate, and low burnseverity classes over the last three decades. Totalreported burn area (National Forest System plus otherownerships) was greatest in Region 5 (1,800,000 ac;730,000 ha), followed by Regions 6, 4, 2, 3, and 1 (fig. 5).The total burned and treated areas of high severity(National Forest System plus other ownerships) inRegion 5 (702,000 ac, 284,000 ha) exceeded that of all

other Regions combined (670,000 ac, 271,300 ha) (fig. 6).For Region 5, the high severity areas (39 percent of thetotal reported wildfire-burned area) exceeded themoderate (29 percent) and low severity categories(33 percent); this is due to the large amount chapar-ral vegetation in Region 5 which generally burn athigh severity conditions. In all the other Regions, theacreage of burned land in the low severity class ex-ceeded the high severity class.

In terms of expenditures for BAER treatments onhigh fire severity areas, the Regions segregated intotwo groups (fig. 7). Both Regions 4 and 5 incurred BAERtreatment expenses of over $27 million, and Region 6exceeded $17 million. However, the expenditures for

Table 9—The 10 costliest fires for BAER treatment spending. All amounts are in 1999 dollars.

National NFS Total NFS TotalFire Name Forest Year (ac) (ha) (ac) (ha) ($) ($)

Rabbit Creek Boise 1994 94880 38425 94880 38425 8,420,000 8,420,000Foothills Boise 1993 139955 56680 257600 104330 8,251,500 8,346,000Tyee Creek Complex Wenatchee 1994 105600 42770 140195 56780 6,156,100 8,978,000Lowman Complex Boise 1989 95000 38475 95000 38475 3,215,500 3,215,500Stanislaus Complex Stanislaus 1987 117980 47780 139980 56690 2,109,450 2,609,450Fork Mendocino 1997 61930 25080 82993 33610 1,839,100 1,888,000Buffalo Creek Pike-San Isabel 1998 11320 4585 11900 4820 1,800,200 2,146,400Clover Mist Shoshone 1988 194000 78570 387000 156735 1,393,500 1,393,500Eighth Street Boise 1997 3160 1280 15193 66155 1,207,000 8,562,400Clarks Incident Plumas 1988 30000 12,150 40000 16,200 1,024,000 1,289,000

Pacific NorthwestRegion 6

21%

SouthwesternRegion 3

4%

Pacific SouthwestRegion 5

32%

Rocky MountainRegion 2

6% NorthernRegion 1

4%

IntermountainRegion 4

33%

6

5

4

1

2

3


0.0

0.2

0.4

0.6

0.8

1.0

1.2

1973

1974

1975

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

Mill

ions

YEAR

AR

EA

BU

RN

ED

(a

c)

0.0

0.1

0.2

0.3

0.4

Mill

ions

AR

EA

BU

RN

ED

(h

a)

NFS OTHER

0

10

20

30

40

50

60

70

80

1973

1974

1975

1976

1977

1978

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

YEAR

BA

ER

SP

EN

DIN

G B

Y N

FS

(19

99

$ a

c-1)

0

20

40

60

80

100

120

140

160

180

BA

ER

SP

EN

DIN

G B

Y N

FS

(19

99

$ h

a-1

)

Figure 3—Total area burned in National Forests and otherlands that had some portion of National Forest lands by yearfrom the Burned Area Reports.

Figure 4—BAER spending by National Forests per unit areaburned by year from Burned Area Reports.

Figure 5—Total areas burned by severity class listed byRegion, 1973-1998 from Burned Area Reports. Treated anduntreated burned areas by severity are totaled separately.

Figure 6—High severity areas burned and treated with BAERfunding by region, 1973-1998 from Burned Area Reports.

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

1 2 3 4 5 6 1 2 3 4 5 6

Mill

ions

REGIONS

AR

EA

BU

RN

ED

(ac)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Mill

ions

AR

EA

BU

RN

ED

(ha)

HIGH SEVERITYMODERATE SEVERITYLOW SEVERITY

NOT TREATED TREATED

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

1 2 3 4 5 6

Millions

REGION

HIG

H S

EV

ER

ITY

AR

EA

TR

EA

TE

D(a

c)

0.00

0.05

0.10

0.15

0.20

0.25

0.30

Millions

HIG

H S

EV

ER

ITY

AR

EA

TR

EA

TE

D

(ha)

Regions 1, 2, and 3 did not exceed $5 million. Regions4, 5, and 6 had 15 of the top 20 most-expensive BAERefforts. Region 4 had four of the top five most costlyBAER efforts. The Rabbit Creek, Foothills, andLowman Fires on the Boise National Forest in Idahoinvolved National Forest System BAER spending of$8.4, $8.2 and $3.2 million, respectively (table 9). Themost expensive BAER project in Region 5 was only$2.1 million (Stanislaus Complex), but the Region hadeight of the top 20 most expensive BAER projects.

BAER project costs per unit area treated followedthe trend of total treatment costs (fig. 2, 8). Costs peracre (ha) were higher for Regions 4, 6, and 5 thanRegions 1, 2, and 3. The most expensive cost per acre,$39 ($96 ha–1), was for Region 4 and the least expen-sive, $9 ($22 ha–1), was from Region 3. The higher costsper acre in Region 4, 6, and 5 fires reflected invest-ments in BAER projects to protect life and property.This was particularly true for high severity fires inRegion 4 (fig. 9).


Figure 8—Cost per area for BAER treatment for Forest ServiceSystems lands by Region in 1999 dollars, 1973-1998 fromBurned Area Reports.

Figure 9—Cost per area for BAER treatment on high severityburned sites in 1999 dollars, 1973-1998 from Burned AreaReports.

Erosion Estimates

The Burned Area Report form asks for an erosionhazard rating for each fire. The rating is divided intolow, moderate, and high erosion hazards categories.For the 321 project fires, the ratio of high erosionareas to high burn severity areas was greater thanone (fig. 10). More areas were rated high erosionhazard than just those with high burn severity. Thiswas probably due to natural erosion hazards associ-ated with local geology, geomorphology, and precipita-tion patterns. Regions 2 and 4 both have high ratios

Figure 7—BAER spending by Region for high severity burnareas in 1999 dollars, 1973-1998 from Burned Area Reports.

0

5

10

15

20

25

30

1 2 3 4 5 6

Mill

ions

REGION

BA

ER

SP

EN

DIN

G (

19

99

$)

0

5

10

15

20

25

30

35

40

45

1 2 3 4 5 6

REGION

0

20

40

60

80

100

BA

ER

TR

EA

TM

EN

T C

OS

T F

OR

AL

L F

IRE

S (

19

99

$ a

c-1)

BA

ER

TR

EA

TM

EN

T C

OS

T F

OR

AL

L F

IRE

S (

19

99

$ h

a-1)

0

50

100

150

200

250

1 2 3 4 5 6

REGION

0

100

200

300

400

500

600

BA

ER

TR

EA

TM

EN

T C

OS

T-H

IGH

BU

RN

SE

VE

RIT

Y (

1999 $

ac-1

)

BA

ER

TR

EA

TM

EN

T C

OS

T-H

IGH

BU

RN

SE

VE

RIT

Y (

1999 $

ha

-1)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

1 2 3 4 5 6

REGIONS

ER

OS

ION

HA

ZA

RD

: FIR

E S

EV

ER

ITY

RA

TIO

HIGH

MODERATE

LOW

due to conditions such as granitic soils and steepslopes, which create naturally high erosion hazards.On the other hand, in all regions the ratio of lowerosion hazard areas to low burn severity areas waslow, indicating that erosion potential was small.

A wide range of erosion potential estimates andwatershed sediment yield (delivered to the channel)potential estimates was found in the Burned AreaReport forms, some with very high values that could beconsidered unrealistic (fig. 11). Erosion potential variedfrom 1 to 7,000 ton ac–1 (2 to 15,500 Mg ha–1), andsediment yield varied over six orders of magnitude.Erosion potential and sediment yield potential did notcorrelate well (r = 0.18, n = 117). Different methods

Figure 10—Average ratio of areas described as low, moderate,and high erosion hazard to areas of low, moderate, and highburn severity by Region from Burned Area Reports.


were used to calculate these estimates on differentfires, making comparisons difficult. Methods includedempirical base models such as Universal Soil LossEquation (USLE), values based on past estimates ofknown erosional events, and professional judgment.

Hydrologic Estimates

Part of the BAER process evaluates the potentialeffects of wildfire on hydrologic responses. One facet ofthis involves determining storm magnitude, duration,and return interval for which treatments are to bedesigned. On the Burned Area Report forms, the mostcommon design storms were 10-year return events(fig. 12a and b). Storm durations were usually lessthan 24 hr with the common design storm magnitudesfrom 1 to 6 in (25 to 150 mm). Five design storms weregreater than 12 in (305 mm) with design return inter-vals of 25 years or less. The variation in estimatesreflects some of the climatic differences throughoutthe Western United States.

The Burned Area Report form also contains anestimate of the percentage of burned watersheds thatis water repellent. Water repellent soils are oftenreported after wildfires, and we expected to find themmore common on coarse-grained soils, such as thosederived from granite. However, there was no statisti-cal difference among geologic parent materials in thepercent of burned area that was water repellent (t-test; fig. 13). Water repellent conditions appeared to bedistributed evenly among soil parent materials. BAERteams also estimate a percentage reduction in infiltra-tion capacity as part of the Burned Area Report.Comparison of reduction in infiltration rate to per-centage of area that was water repellent showed nostatistically significant relationship (fig. 14). Factorsother than water repellent soil conditions, such as loss

of the protective forest floor layers, obviously affectinfiltration capacity.

Estimation methods for expected changes in chan-nel flow due to wildfire were variable but primarilybased on predicted change in infiltration rates. Thus a20 percent reduction in infiltration resulted in a esti-mated 20 percent increase in channel flows. Variousmethods were used such as empirical-based models,past U.S. Geological Survey records from nearby wa-tersheds that had a flood response, and professionaljudgment. Some reports show a very large percentincrease in design flows (fig. 15).

Risk Analysis

The kinds of resources or human values judged bythe BAER evaluation team to be at risk from postfiresedimentation and flooding are listed on the BurnedArea Report form. These consisted of life, water qual-ity, threatened and endangered (T & E) species, soil

Figure 11—Estimated hillslope erosion potential and water-shed sediment yield potential (log scale) for all fires request-ing BAER funding.

1

10

100

1000

10000

100000

1000000

1(2.24)

10(22.4)

100(222.4)

1000(2224)

10000(22240)

2.95

0.295

29.5

295

2950

29500

295000

Erosion Potential t ac-1 (Mg ha-1)

Wa

ters

he

d S

ed

ime

nt Y

ield

(yd

s3 m

i-2)

Wa

ters

he

d S

ed

ime

nt Y

ield

(m3 k

m-2

)

0

5

10

15

20

25

30

0 20 40 60 80 100 120

Design Storm Return Period (yr)

Des

ign

Sto

rm D

urat

ion

(hr)

0

5

10

15

20

25

0 20 40 60 80 100 120

Design Storm Return Period (yr)

Des

ign

Sto

rm M

agni

tude

(in)

0

100

200

300

400

500

600

Des

ign

Sto

rm M

agni

tude

(mm

)

Figure 12—(a) Design storm duration and (b) design stormduration by return period for all fires requesting BAER funding.

(a)

(b)


Figure 14—Fire-induced water repellent soil areas comparedto the estimated reduction in infiltration for all fires requestingBAER funding. Regression line shows a poor correlationbetween increased water repellent soil areas and the reductionin infiltration (R2 = 0.31).

Figure 15—Estimated design peakflow change (log scale) dueto burned areas related to the estimated reduction in infiltrationfor all fires requesting BAER funding.

Figure 13—Fire-induced water repellent soil areas and their geologic parent material for allfires requesting BAER funding. Fire-induced water repellency was not significantly differentby parent material (t-test, α = 0.05).

productivity, and property. The latter category in-cludes homes, roads, cultural features, water suppliesand reservoirs, and agriculture.

Property, water quality, and soil productivity werecited as reasons for conducting BAER projects in abouta third of all projects (table 10). Region 5 (California),with its high population, had the highest response (51percent) for property, while sparsely populated Region1 had the lowest. In terms of property protection, roadsand homes were mentioned most frequently as rea-sons for treatments in Region 5 (34 and 28 percent of

the BAER project responses, respectively) (table 11).In the other regions, homes constituted a reason forimplementing BAER treatments in less than 11 per-cent of the projects. Protection of homes was citedmore frequently in the 1990’s as a major fire suppres-sion activity objective than it was in previous decades.It is very likely that the same will occur for futureBAER projects. Cultural features, water supplies, andagriculture were listed as factors in BAER projects inless than 10 percent of the responses, except for agri-culture in Region 2 (20 percent). Considering the

0

10

20

30

40

50

60

70

80

90

100

Geologic Parent Material

Wat

er R

epel

lent

Are

a (%

)

Metamorphic

mean=34n=41

mean=29n=51

mean=40n=47

mean=25n=43

Sedimentary Granitic Volcanic

R2 = 0.31n=125

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50 60 70 80 90 100

Water Repellent Area (%)

Infil

trat

ion

Red

uctio

n (%

)

1:1

1

10

100

1000

10000

100000

0 20 40 60 80 100

Infiltration Reduction (%)

Desi

gn P

eak

Flo

w C

hange (

%)


Table 10—Described values at risk for spending on BAER projects by Region.

Number of T and E Water SoilRegion Projects Life Property Species Quality Productivity

- - - - - - - - - - - - - - - - - - - Percent of Projects - - - - - - - - - - - - - - - - - - -1 56 2 29 14 14 142 20 5 35 0 70 503 69 4 29 7 26 584 45 18 47 33 60 445 201 10 51 8 41 246 79 8 33 11 58 52All Regions 470 9 41 11 41 36

Table 11—Property subcategory breakdown of values at risk for spending on BAER projects by Region.

Number of Cultural WaterRegion Projects Homes Roads Feature Supplies Agriculture

- - - - - - - - - - - - - - - - - - Percent of Projects - - - - - - - - - - - - - - - - - -1 56 11 9 2 4 92 20 5 0 5 0 203 69 9 20 1 0 44 45 11 20 0 7 75 201 28 34 3 1 56 79 6 24 0 0 9All Regions 470 17 25 2 2 7

rapidly growing wildland-urban interface fire prob-lem in the West, property protection is likely tokeep growing as a reason for implementing BAERtreatments.

Protection of life was listed as a reason for conduct-ing BAER projects in Region 4 (18 percent) more oftenthan in the other Regions. Water quality was citedover 50 percent of the time in Regions 2, 4, and 6, butonly 14 percent of the time in Region 1. Soil productiv-ity was mentioned as a major purpose for BAER inRegions 2, 3, 4, and 6, with the most concern (58percent) expressed in the Region 3 (Arizona and NewMexico). Region 1 had a relatively low response for soilproductivity. Protection of threatened and endangered(T & E) species values was mentioned most frequentlyin Region 4 and not even listed as a reason for BAERprojects in Region 2.

Probability of Success

The Burned Area Report form contains a section forestimating the probability of success for land, chan-nel, and road treatments 1, 3, and 5 years afterimplementation. This is required by FSH 2509.13—Burned Area Emergency Rehabilitation Handbook,

WO Amendment 2509.13-95-9, effective 1/12/95,Chapter 30—Cost Risk Analysis and Evaluation ofAlternatives for Emergency Rehabilitation, Part 31.4Probability of Success and Potential Resource ValueLoss. The handbook states that the BAER team“…should provide an interdisciplinary decision onthe estimated probability of each alternative’s abilityto successfully minimize or eliminate emergencywatershed conditions….” Probabilities of success wereprovided for 321 of the 470 fires for which BAERreports were completed. The data did not contain anyparticular Region-to-Region trends. The combinedtreatment probability of success data (averages andranges) showed a consistently higher predictedprobability of success for road treatments than forhillslope and channel treatments (table 12). Theseestimations are the product of an interdisciplinaryteam decision and represent the combined experi-ence of the individual BAER team members.

Cost of No Action/Alternatives

Another section of the Burned Area Report formrequires estimates of the costs of no action and pos-sible treatment alternatives, as well as determination


of the cost plus estimated loss for the proposed treat-ments. For the 321 project fires, estimates of the costsof no action and treatment alternatives ranged from$9,000 to $100 million. BAER teams calculated theseestimates based on downstream property value atrisk, soil productivity value, water quality value, T &E species value, and other resource values estimatedto be affected by the fire and possible floods or debrisflows. Potential soil productivity losses may be basedon: estimated site index changes due to fire and pos-sible loss in harvestable timber during the next regen-eration cycle; the cost of top soil if purchased commer-cially to replace that anticipated to be lost; or estimatesby professional judgment. Water quality values arebased on the cost of cleaning reservoirs, increasedcosts of treating drinking water, and estimates ofaquatic habitat degradation.

Costs of BAER treatments were compared to esti-mated losses (without treatment) from the BurnedArea Report forms (fig. 16). BAER treatments appearto be very cost effective, generally costing one-tenth asmuch as the expected losses if no treatment were to beimplemented. Expected losses are just estimates; we

do not have data on actual losses that may haveoccurred.

BAER Team Members

The composition of BAER teams by discipline andRegion was determined from the Burned Area Reportforms to determine appropriate disciplines to targetfor additional training (table 13). Just under 43 per-cent of all the BAER teams included in this data set(470) came from Region 5. The smallest number wasfrom Region 2 (4 percent). Regions 4, 1, 3, and 6 had 10,12, 15, and 17 percent, respectively.

The predominant disciplines on the BAER teamswere hydrology and soil science (table 13). Except forRegion 3, the percentages of BAER teams containinghydrologists and soil scientists were fairly consistent(78 to 87 percent) across Regions. Only two-thirds ofRegion 3 BAER teams had members from these disci-plines. The next most common BAER team disci-plines, wildlife biology (34 to 71 percent), timber man-agement (30 to 65 percent), and engineering (22 to 56percent), exhibited a two-fold range between Regions.Region 1 had the lowest representation on its BAERteams for engineering, range management, geology,archeology, fire management, contracting, and re-search disciplines. Region 4 had the highest represen-tation of wildlife biology, fire management, ecology,fisheries, contracting, and research disciplines.

Monitoring Reports

A wide variety of monitoring reports was collectedfrom the six Regions. Most were internal administra-tive reports dealing with one fire or several fires inproximity. Several were regional burn area rehabilita-tion activity reviews, resulting from interdisciplinaryteam review of multiple fires over several forests toevaluate current policies and techniques.

We obtained 157 documents that contained postfiremonitoring information. Of those, 55 (35 percent)contained quantitative data of some kind. The rest(65 percent) contained qualitative evaluations oftreatment success, such as trip report narratives or

Figure 16—BAER spending compared to projected value lossif no action was taken (log scale). BAER spending did notexceed estimated values.

10,000

100,000

1,000,000

10,000,000

100,000,000

1,000,000,000

10,000 100,000 1,000,000 10,000,000

BAER Spending (1999 $)

Exp

ect

ed V

alu

e L

oss

If N

oA

ctio

n (

1999 $

)

1:1

Table 12—Probability of treatment success by Regions. Treatments are grouped into threecategories: hillslope, channel, and road.

BAER Year 1 Year 3 Year 5Treatment Mean Range Mean Range Mean Range

- - - - - - - - - - - - - - - - - - - - - - - - - Percent - - - - - - - - - - - - - - - - - - - - - - -Hillslope 69 60-80 82 74-87 90 85-94Channel 74 57-80 83 74-88 89 84-93Road 86 79-90 89 78-94 94 87-100


photos. We also received 17 published reports, some ofwhich did not evaluate BAER treatments specificallybut included incidental information as part of anotherstudy. Most of the published reports were discussed inthe Literature Review section.

The type of information contained in the monitoringreports varied widely. Quantitative reports on a singletreatment (e.g., seeding) tended to use different mea-surements (cover, density, biomass, sediment pro-duced), making tabulation and comparison of theresults from different projects difficult. Treatmentswere monitored at varying times after the fires, from3 months to 12 years. Where “cover” was measured,the category sometimes included only plants, some-times litter, and sometimes also rock or wood. “Groundcover density” was sometimes used to refer to plantcover only where it was rooted in the ground. In othercases, “ground cover” included the aerial portions ofplants. Many reports did not specify what was in-cluded in the category “cover.” Often reports containeddata on plant cover or sediment movement, but notother site variables that could have put the results ina wider context. In particular, vegetation type, water-shed size, slope angle, and aspect of monitored siteswere frequently missing from data presentations andnarrative accounts. Most reports were prepared forinternal use, where these variables would be betterknown to likely readers. However, the lack of descrip-tive site information made the results of monitoringmore difficult to interpret for this analysis.

A wealth of information was recorded in the moni-toring reports. To capture the considerable but ex-tremely varied experience represented, qualitativeinformation from the reports was entered into thedatabase in various “comments” fields, along with

interview remarks. Comments were aggregated andused to compose effectiveness and implementationfactor summaries for each treatment (appendix B).

The quantitative reports covered 46 fires, with somefires covered by multiple reports and some reportscovering several fires in one document. Report datesranged from 1967 to 1998. Most of the data collectedconcerned ground cover production or erosion reduc-tion by seeded species (32 reports), effectiveness ofcontour-felled logs (5 reports) or straw bale checkdams (3 reports), and water quality parameters suchas turbidity (5 reports). Reports sometimes coveredmore than one treatment. Only a few of the monitoringefforts compared treated areas to untreated areas.The others based effectiveness conclusions on amountof plant cover present, whether structures trappedsediment, and so forth. Many reports simply docu-mented some facet of hillslope or stream recovery afterfire, sometimes in areas that did not receive BAERtreatments.

Nonquantitative reports documented treatment ef-fectiveness qualitatively or made rough visual esti-mates of success parameters, such as amount of grasscover or storage effectiveness of log erosion barriers.They covered approximately 85 different fires. Manywere trip reports that simply pronounced a treatmentsuccessful or not. Most, however, also analyzed rea-sons for success or failure and made recommendationsfor improving future projects. Those comments wereused extensively to develop treatment effectivenessand implementation factor summaries (appendix B).The bulk of the nonquantitative reports dealt withseeding (54 reports), straw bale check dams (18),contour-felled logs (15), or channel treatments (16).Most reports covered more than one treatment. A

Table 13—Percentage of BAER teams by Region having personnel from various disciplines.

RegionDiscipline Overall 1 2 3 4 5 6

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - Percent - - - - - - - - - - - - - - - - - - - - - - - - - - - - -Hydrology 81.1 81.8 85.0 65.2 84.4 83.2 86.1Soil Science 78.7 87.3 85.0 65.2 82.2 78.2 82.3Wildlife Biology 61.9 34.5 65.0 59.4 71.1 66.3 65.8Timber Management 46.4 49.1 60.0 30.4 44.4 43.1 64.6Engineering 43.4 21.8 45.0 30.4 37.8 55.9 40.5Range Management 40.6 18.2 70.0 60.9 68.9 26.2 51.9Geology 31.5 9.1 20.0 15.9 26.7 41.6 40.5Archeology 23.4 5.5 20.0 17.4 26.7 28.2 27.8Fire Management 21.7 16.4 35.0 23.2 37.8 19.3 17.7Ecology 20.9 23.6 25.0 24.6 51.1 12.4 19.0Fisheries 18.5 21.8 5.0 5.8 40.0 14.4 29.1Contracting 7.2 3.6 15.0 4.3 22.2 4.5 8.9Research 5.1 0.0 5.0 2.9 20.0 3.0 7.6

Total No. of BAER Teams 470 55 20 69 45 202 79


number of reports focused on salvage logging “bestmanagement practices” evaluation but also mentionedBAER treatments. Reports dated from 1962 through1998.

Grass seeding (aerial or ground) was usually per-ceived as “effective” if: (1) it produced at least 30percent cover by the end of the first growing season;(2) seeded species comprised a significant amount ofthe total plant cover at the end of the first growingseason; or (3) less sediment movement was measuredcompared to an unseeded plot or watershed. Statisti-cal significance of observed differences between seededand unseeded sites was seldom tested. In the first yearafter fire, seeding was considered generally effectivein 9 of 16 quantitative monitoring reports (56 percent).Second year effectiveness was similar (10 of 16 reportsor 62.5 percent). One 1978 rehabilitation review fromRegion 3 collected data from 12 fires, ranging from 1 to12 years after treatment. They found that seeding wasgenerally successful (produced cover) on forested sitesbut not on chaparral sites (Taylor and others 1979).

The amount of cover produced by seeded grassesduring the first and second years after fire variedwidely (tables 14 and 15). Many of the monitoringreports contained data on annual ryegrass and cerealgrains (rye, barley or oats), the species most in use inrecent years. More information was available fromCalifornia (Region 5) than any other area, most of itfrom chaparral sites. Annual ryegrass and cereal grainsproduced considerable cover in some cases. In others,they did not appreciably increase plant cover or reduceerosion, especially the first year after fire.

In the nonquantitative reports, seeding was judgedto be “effective” or successful the first year after fire in22 of 28 cases (79 percent), generally based on thepresence or absence of grass, evidence of rilling, oramount of cover compared to unseeded areas. Second-year results were similar, with 11 of 14 cases (79percent) considered successful. In some cases seedingwas considered “effective” in producing cover but prob-ably not necessary, as natural vegetation regrowthwas abundant as well (Bitterroot National Forest1997). In others it was given a mixed rating becausethe seeded species persisted for many years or ap-peared to crowd out native vegetation (Isle 1988,Loftin and others 1998). Sometimes seeding was judgedeffective in one part of a project but not another(Herman 1971, Liewer 1990, Ruby 1995, Story andKracht 1989). Loftin and others (1998; Region 3)suggested that protection from grazing could be thesingle most effective method for enhancing cover pro-duction by both seeded grasses and recovering nativevegetation.

Contour-felled logs were judged to be effective in all5 documents in which some kind of data were reported.Accumulation of sediment uphill of the barriers (Green

1990), lack of rilling in the treated area, or reductionin sediment collected downhill compared to an un-treated plot were considered “effective” outcomes. Forexample, DeGraff (1982) measured “sediment trapefficiency” (STE) at 0.7 on slopes of less than 35percent on the Sierra National Forest, meaning that70 percent of the length of a log, on average, hadaccumulated sediment. Logs on steeper slopes exhib-ited an average STE of 0.57. Griffith (1989a) observed1.5 t ac–1 (3.4 Mg ha–1) of sediment behind a silt fencebelow a watershed treated with contour-felled logs,compared to 10.7 t ac–1 (24.2 Mg ha–1) from an un-treated watershed, during the first postfire year on theStanislaus National Forest. Both watersheds weresalvage-logged the following year, and sediment out-put increased to 10 t ac–1 (23 Mg ha–1) on the treatedand over 34 t ac–1 (77 Mg ha–1) on the untreated wa-tershed. Several reports from the first few years afterthe Foothills Fire (Boise National Forest) stated thatno significant amounts of sediment were producedfrom any of several experimental watersheds treatedwith contour-felled logs, whether or not they weresalvage-logged (e.g., Maloney and Thornton 1995).The reports noted that the area experienced no majorthunderstorms until late summer 2 years after thefire.

In nonquantitative reports, contour-felled logs wereconsidered effective in 11 of 13 cases (85 percent) inwhich they were actually tested by storms. Severalreports pointed out that contour-felled logs are de-signed to reduce water flow energy and promote infil-tration, not trap sediment, but they showed somebenefit as direct sediment traps and also enhancedestablishment of seeded grasses. Different terminol-ogy was sometimes used by different Regions in thereports. The term “contour-felled log” was synony-mous with “log erosion barrier” in most areas, but inRegion 3 contour-felling referred only to felling ofmaterial, not anchoring and sealing it. There theterms “log erosion barrier” or “log terracette” wereused for anchored logs. The nonanchored logs were notconsidered a successful treatment.

Mulch was evaluated in two quantitative monitor-ing reports and found to be very effective. For example,Faust (1998) collected only 0.8 t ac–1 (1.8 Mg ha–1) ofsediment below a slope mulched and seeded with oats,compared to 5.8 t ac–1 (12.9 Mg ha–1) below a slopeseeded with oats alone.

Kidd and Rittenhouse (1997) rated mechanicallydug contour trenches as the “best” treatment in trap-ping sediment after the Eighth Street Fire, BoiseNational Forest, Idaho. However they rated hand-dugcontour trenches as the “worst” treatment due to poorconstruction (shallow depth) and layout (off contour).These trenches often contributed to rilling. Mechani-cally constructed contour trenches worked better


Tab

le 1

4—S

elec

ted

first

-yea

r se

eded

spe

cies

cov

er,

tota

l pla

nt c

over

, an

d so

il lo

ss m

easu

red

in m

onito

ring

repo

rts.

Loc

atio

n an

d ve

geta

tion

type

as

wel

l as

slop

e ar

ein

clud

ed. S

eede

d co

ver

perc

enta

ges

incl

ude

only

thos

e sp

ecie

s th

at w

ere

seed

ed. T

otal

cov

er p

erce

ntag

es in

clud

e al

l veg

etat

ion.

Co

ver

Lo

cati

on

Veg

etat

ion

See

ded

Sp

ecie

sS

lop

e-

- -

- -

Rat

e -

- -

- -

See

ded

To

tal

- -

-So

il L

oss

1 - -

-S

ou

rce

(%)

(lb a

c–1)

(kg

ha–1

)-

- -

(%)

- -

-(T

ac–1

)(M

g ha

–1)

Cle

vela

nd N

F, C

Ach

apar

ral

annu

al r

yegr

ass

5310

111

196.

514

.6B

lank

en-

Agu

anga

fire

, 198

4an

nual

rye

gras

s56

2022

224

12.5

28ba

ker

and

annu

al r

yegr

ass

5840

451

1814

.031

.4ot

hers

198

5un

seed

ed56

00

020

6.8

15.2

Los

Pad

res

NF,

CA

chap

arra

lan

nual

ryeg

rass

, lan

a ve

tch

nd7

849

63nd

ndE

splin

& S

hack

elfo

rd 1

980

Cac

hum

a fir

e, 1

977

chap

arra

lm

isse

d st

rips

ndtr

ace

526

ndnd

San

Ber

nard

ino

Mts

., C

Ach

apar

ral

annu

al r

yegr

ass

nd8

9<

115

6013

4T

yrre

l 198

1P

anor

ama

fire,

198

0ch

apar

ral

annu

al r

yegr

ass

+ fe

rt.

nd16

18<

130

8819

7B

land

o br

ome

nd12

13<

545

7216

1Z

orro

fesc

uend

1011

<15

5572

161

unse

eded

nd0

070

2147

Los

Pad

res

NF

, CA

chap

arra

lB

land

o br

ome

nd3

352

82nd

ndD

ucum

mon

198

7La

s P

ilitis

fire

, 198

5un

seed

ednd

00

56nd

ndT

uolu

mne

Can

yon,

CA

chap

arra

lB

land

o br

ome,

Zor

ro>

505

616

± 7

41 ±

9nd

ndJa

nick

i 198

9S

tani

slau

s co

mpl

ex, 1

987

fesc

ueM

endo

cino

NF

, CA

chap

arra

lce

real

oat

s35

ndnd

~ 4

242.

65.

8F

aust

199

8F

orks

Fire

, 199

6un

seed

ed35

00

025

3.7

8.3

nativ

e m

ix +

oat

s38

ndnd

~ 6

382.

86.

3na

tive

mix

alo

ne40

ndnd

nd40

1.1

2.5

seed

+ m

ulch

35nd

ndnd

950.

81.

8T

onto

NF

, AZ

coni

fer/

shru

bw

eepi

ng lo

vegr

ass

ndnd

nd1

9nd

ndA

mbo

s 19

92D

ude

fire,

199

0in

tens

e bu

rn+

oth

ers

coni

fer/

shru

bw

eepi

ng lo

vegr

ass

ndnd

nd8

33nd

ndm

oder

ate

burn

+ o

ther

sC

ibol

a N

F, N

Mpo

nder

osa

pine

/m

ix2

nd45

s ft

–248

4 s

m–2

1556

ndnd

Tay

lor

and

othe

rs 1

979

Gal

inas

fire

, 197

6pi

nyon

-juni

per

San

ta F

e N

F, N

Mpo

nder

osa

pine

orch

ardg

rass

, har

dnd

41 s

ft–2

441

s m

–20.

89.

4nd

ndT

aylo

r an

d ot

hers

197

9P

orte

r fir

e, 1

976

fesc

ue, t

imot

hy, a

lfalfa

sam

e, r

esee

ded

in J

uly

310

.4nd

ndS

tani

slau

s N

F, C

Api

ne/o

ak;

annu

al r

yegr

ass

10-6

510

1120

± 8

29 ±

7nd

ndJa

nick

i 198

9S

tani

slau

s co

mpl

ex, 1

987

mix

ed c

onife

rsa

me

unse

eded

10-6

50

00

19 ±

5nd

ndm

ixed

con

ifer

mix

325

-65

7.5

85

± 6

21 ±

5nd

nd(h

ighe

r el

ev.)

(con

.)


Tab

le 1

4—(C

on.)

Co

ver

Lo

cati

on

Veg

etat

ion

See

ded

Sp

ecie

sS

lop

e-

- -

- -

Rat

e -

- -

- -

See

ded

To

tal

- -

-So

il L

oss

1 - -

-S

ou

rce

(%)

(lb a

c–1)

(kg

ha–1

)-

- -

(%)

- -

-(T

ac–1

)(M

g ha

–1)

Sta

nisl

aus

NF

, CA

coni

fer/

met

a-an

nual

rye

gras

snd

1011

ndnd

3.6

8.1

Grif

fith

1989

bS

tani

slau

s co

mpl

ex, 1

987

sedi

men

tary

soi

lsa

me

unse

eded

nd0

0nd

nd9.

821

.9co

nife

r/an

nual

rye

gras

s “p

oor

nd10

11nd

nd>

8.3

18.6

gran

tic s

oil

catc

h”sa

me

annu

al r

yegr

ass

“goo

dnd

1011

ndnd

>2.

14.

7ca

tch”

sam

ean

nual

rye

gras

s +

nd10

11nd

nd1.

02.

2ne

edle

cas

tsa

me

unse

eded

nd0

0nd

nd>

8.4

18.8

Nez

Per

ce N

F I

DD

ougl

as-f

irce

real

rye

>60

ndnd

8092

0.04

in0.

9 m

mG

reen

199

0G

reen

Cr.

Pt.

fire,

198

8N

ezP

erce

NF

, ID

pond

eros

a pi

nece

real

rye

+ y

ello

w59

1011

5.4

27.6

ndnd

Clin

e an

d B

rook

s 19

79C

otte

r B

ar II

fire

, 197

7sw

eetc

love

rce

real

rye

+cl

over

+59

10 +

45

11 +

50

2.6

27.4

ndnd

drill

ed u

rea

drill

ed u

rea

5945

50(2

.5)

48.5

ndnd

untr

eate

d59

00

(3.2

)27

.6nd

ndou

tsid

e of

cere

al r

ye+

clov

er +

nd10

+ 4

511

+ 5

013

.662

.9nd

ndex

pmt.

plot

sdr

illed

ure

aS

tani

slau

s N

F, C

Am

ixed

con

ifer

cere

al b

arle

ynd

7382

2.8

t ac–1

3.3

t ac–1

ndnd

Grif

fith

1991

; 199

3A

rch

Roc

k fir

e, 1

990

cere

al b

arle

ynd

7382

ndnd

>1.

3>

2.9

unse

eded

nd0

0nd

nd4.

39.

6E

ldor

ado

NF

, CA

coni

fer/

cere

al b

arle

y52

ndnd

nd60

.05.

4 lb

2.4

kgP

arso

ns a

ndgr

aniti

c so

iltr

ough

–1tr

ough

–1M

cDon

ald

1993

unse

eded

500

0nd

32.5

9.2

lb4.

2 kg

trou

gh–1

trou

gh–1

coni

fer/

met

a-ce

real

bar

ley

38nd

ndnd

28.8

0.8

lb0.

4 kg

mor

phic

soi

ltr

ough

–1tr

ough

–1

unse

eded

400

0nd

3.6

lb1.

6 kg

trou

gh–1

trou

gh–1

Gila

NF

, NM

mix

con

ifer

Reg

reen

+ o

ther

s40

ndnd

1-2

140

0S

oude

rs 1

997

HB

fire

, 199

5po

nder

osa

pine

Reg

reen

+ o

ther

s40

ndnd

<1

20

0

1 Val

ues

in ta

ble

wer

e av

erag

ed fr

om in

divi

dual

plo

ts in

repo

rt in

som

e ca

ses.

Sta

tistic

al s

igni

fican

ce w

as n

ot te

sted

in th

ese

stud

ies.

Som

e va

lues

are

min

imum

s be

caus

e si

lt fe

nces

ove

r-to

pped

or fa

iled

durin

g w

inte

r. O

ne s

tudy

just

mea

sure

d se

dim

ent l

oss

with

ero

sion

pin

s, a

nd a

noth

er m

easu

red

poun

ds o

f sed

imen

t in

trou

ghs

with

a 2

5 to

30

ft (7

.7 to

9.1

m)

cont

ribut

ing

area

.2 H

ard

fesc

ue, s

and

drop

seed

, sm

ooth

bro

me,

Rus

sian

wild

rye,

yel

low

sw

eetc

love

r, b

lack

med

ic, w

eepi

ng lo

vegr

ass.

3 Tim

othy

, har

ding

gra

ss, s

moo

th b

rom

e, B

land

o br

ome.


Tab

le 1

5—S

elec

ted

seco

nd-y

ear

seed

ed s

peci

es c

over

, tot

al p

lant

cov

er, a

nd s

oil l

oss

mea

sure

d in

mon

itorin

g re

port

s. L

ocat

ion

and

vege

tatio

n ty

pe a

s w

ell a

s sl

ope

are

incl

uded

. See

ded

cove

r pe

rcen

tage

s in

clud

e on

ly th

ose

spec

ies

that

wer

e se

eded

. Tot

al c

over

per

cent

ages

incl

ude

all v

eget

atio

n.

Co

ver

Lo

cati

on

Veg

etat

ion

See

ded

Sp

ecie

sS

lop

e-

- -

- -

Rat

e -

- -

- -

See

ded

To

tal

- -

-So

il L

oss

1 - -

-S

ou

rce

(%)

(lb a

c–1)

(kg

ha–1

)-

- -

(%)

- -

-(T

ac–1

)(M

g ha

–1)

Ton

to N

F, A

Zch

apar

ral

wee

ping

love

gras

s,nd

39 s

ft–2

433

s m

–20

41nd

ndT

aylo

r an

d ot

hers

1979

Bob

fire

, 197

5an

nual

rye

gras

sP

resc

ott N

F, A

Zch

apar

ral

unse

eded

nd0

038

Tay

lor

and

othe

rs19

79M

ingu

s R

x fir

e, 1

975

Los

Pad

res

NF

, CA

chap

arra

lan

nual

rye

gras

snd

78

878

ndnd

Esp

lin a

ndC

achu

ma

fire,

197

7m

isse

d st

rips

ndtr

ace

trac

e4

78nd

ndS

hack

lefo

rd19

80M

endo

cino

NF

, CA

chap

arra

lce

real

oat

s35

ndnd

ndnd

0.7

1.6

Fau

st 1

998

For

ks fi

re, 1

996

unse

eded

350

0nd

nd0.

40.

9T

ahoe

NF

, CA

brus

hce

real

bar

ley

20-2

5nd

ndnd

228.

418

.8Ja

nick

i 199

2C

otto

n fir

e, 1

990

unse

eded

20-2

50

0nd

287.

115

.9S

tani

slau

s N

F, C

Aco

nife

r/m

eta-

annu

al r

yegr

ass

nd10

11nd

nd>

1.9

>4.

2G

riffit

h 19

89a

Sta

nisl

aus

com

plex

, 198

7se

dim

enta

ryun

seed

ednd

00

nd4.

29.

4co

nife

r/an

nual

rye

gras

snd

1011

ndnd

>10

.6>

23.7

gran

itic

soil

“poo

r ca

tch”

annu

al r

yegr

ass

nd10

11nd

nd>

5.0

>11

.2“g

ood

catc

h”un

seed

ednd

00

nd>

1.3

>2.

9S

tani

slau

s N

F, C

Aco

nife

rce

real

bar

ley

nd73

82tr

ace

7.1

T a

c–1nd

ndG

riffit

h 19

91A

rch

Roc

k fir

e, 1

990

coni

fer

–ce

real

bar

ley

nd73

826.

6 T

ac–1

13 T

ac–1

ndnd

salv

age

logg

edG

ila N

F, N

Mm

ixed

con

ifer

Reg

reen

+ o

ther

s40

ndnd

517

gain

Sou

ders

199

7H

B fi

re, 1

995

pond

eros

a pi

neR

egre

en +

oth

ers

40nd

nd2-

37

0.23

in5.

9 m

m1 V

alue

s in

tabl

e w

ere

aver

aged

from

indi

vidu

al p

lots

in re

port

in s

ome

case

s. S

tatis

tical

sig

nific

ance

was

not

test

ed in

thes

e st

udie

s. S

ome

valu

es a

re m

inim

ums

beca

use

silt

fenc

es o

ver-

topp

edor

faile

d du

ring

win

ter.

One

stu

dy ju

st m

easu

red

sedi

men

t los

s w

ith e

rosi

on p

ins,

and

ano

ther

mea

sure

d po

unds

of s

edim

ent i

n tr

ough

s w

ith a

25

to 3

0 ft

(7.6

to 9

.1 m

) co

ntrib

utin

g ar

ea.


because of their greater depth, better layout on thecontour, and improved infiltration from deep ripping.

Straw bale check dams were judged to be effective in11 of 16 qualitative reports (69 percent), based onaccumulation of sediment behind the structures andstructural integrity after first year storms. Failuresresulted from poor implementation or placement, orfrom exceptionally large storms that exceeded damdesign.

Fites-Kaufman (1993) reported on the failure ofstraw bale, log, and sandbag check dams after theCleveland Fire on the Eldorado National Forest, Cali-fornia. Thirty percent of straw bale check dams failedfrom undercutting and blow outs compared to only 3percent of log and sand bag check dams. Failuresoccurred in narrow, steep drainages where only twobales comprised the check dam. Downstream supportfrom rocks or logs reduced the failure rate. No esti-mates of the sediment trapping efficiency were made.

Niehoff (1995) noted that straw bale check dams hadmixed success after the Mary-Mix Fire, ClearwaterNational Forest, Idaho in 1986. Straw bales placed inlow-to-moderately incised first and second orderchannels were in place and functioning to stabilizestream grade 1 and 9 years postfire. Straw bale checkdams placed in deeply incised drainages were com-pletely blown out at the end of the first year.

Kidd and Rittenhouse (1997) reported that 800 strawbale check dams installed in channels after the EighthStreet Fire on the Boise National Forest, Idaho had a99 percent structural integrity rate. Although thesestructures were still being monitored, no estimates ofsediment trapping efficiency were available. On ascale of “1” to “10”, straw bale check dams were rated“9” in terms of their effectiveness. Observations of logand rock check dams installed after the ClevelandFire on the Eldorado National Forest, California indi-cated that they were effective in trapping sedimentand held up well over time (Parsons 1994). No esti-mates of sediment storage were made. Other channeltreatments of various kinds were also regarded aseffective most of the time (13 of 17, or 76 percent ofevaluations). These included channel clearing, log silldams, and similar measures.

Road treatments (outsloping, trashracks at culverts,armored crossings, etc.) were specifically evaluatedonly in a few narrative reports. Herman (1971) notedthat immediately after the Entiat Fire, WenatcheeNational Forest, Washington trash racks were inplace and still functioning, but had collected onlysmall amounts of debris due to postfire removal ofwoody material from channels. He believed that long-term maintenance of trash racks was necessary sincefire-killed trees would at some point begin contribut-ing large amounts of woody debris into channels.

Boyd and others (1995) reported on the hydrologicfunctioning of roads and their structures within theCleveland Fire, Cleveland National Forest, Californiaafter a winter storm of 4+ in (100 mm) in 48 hours. Anoversized culvert put in place after the fire success-fully processed large chunks of wood and rocks. Anearby normal-sized culvert was repeated pluggedduring the storm, resulting in numerous overflowsonto the road. Flanagan and Furniss (1997) describedthe reduction in flow capacity by partial blockage.During the same storm in which they examined cul-vert functioning, Boyd and others (1995) observed thatsome correctly constructed postfire water bars did nothave sufficient rocks or slash to dissipate the energy ofhigher surface runoff. The resulting concentrationand channelization of runoff produced additional smallgullies and one large, entrenched gully.

Road treatments were generally judged to be effec-tive. Trip reports sometimes mentioned road treat-ment effectiveness incidental to evaluating other typesof treatments.

Treatment Effectiveness Ratings

Interviewees rated the effectiveness of treatmentsused on specific fires with which they were familiar(table 16). In-channel felling, slash spreading,streambank armoring, trail work, rock gabion dams,culvert inlet/outlet armoring, culvert overflow bypasses,debris basins, culvert risers, outsloping roads, waterbars, storm patrol, and armored fords received two orfewer evaluations per treatment and are not tabulated.Treatments were rated across the spectrum from “ex-cellent” to “poor,” but just over 76 percent of the effec-tiveness ratings were either “good” or “excellent.”

Hillslope Treatments—Hillslope treatments areimplemented to keep soil in place and comprise thegreatest effort in most BAER projects. Aerial seeding,the most frequently used BAER treatment, was ratedabout equally across the spectrum from “excellent” to“poor.” The rating for contour-felled logs was “excel-lent” or “good” in 66 percent of the evaluations. Mulch-ing was rated “excellent” about the same amount(67 percent), and nobody considered it a “poor” treat-ment. Nearly 82 percent of the evaluations placedground seeding effectiveness in the “good” category.There was a 100 percent concurrence that silt fenceswere “excellent” or “good” as a BAER treatment. Evalu-ations of seeding plus fertilizer covered the spectrumfrom “excellent” to “poor,” although most responseswere “fair” or “poor.” The remainder of the hillslopetreatments, received only three evaluations each, so itis difficult to come up with conclusions beyond the factthat they were generally rated “excellent,” “good,” or“fair” and none were evaluated as being “poor.”


BAER spending on hillslope treatments was com-pared. From 1973 through 1998, over $20 million (in1999 dollars) was spent on contour-felled logs and onaerial seeding (fig. 17). Less than $1.5 millions wasspent on other treatments during the same time pe-riod. Clearly these two treatments were the mostpopular. In the 1970’s, there was little spending oncontour-felled logs, and in the 1980’s over $4 millionwas spent. Spending increased dramatically in the1990’s as this treatment gained popularity (fig. 17).Among Regions, Region 4 (mostly Boise National For-est) spent the most ($18.7 million) on contour-fellingtreatments, while Region 5 spent the most on aerialseeding ($8.5 million) (fig. 18). Region 6 spent the moston seeding plus fertilizer and ground seeding.

There were enough evaluations of aerial seeding andcontour-felled logs to assess effectiveness by Regionfor these treatments (table 17). A majority of inter-viewees from Regions 1, 4, and 6 rated aerial seedingas “excellent” or “good.” However, in Regions 3 and 5the majority rated aerial seeding as “fair” or “poor.”For contour-felled logs, a majority of interviewees inRegions 1, 3, 4, and 5 believed that its effectivenesswas “excellent” or “good.” Region 2 evaluations wereevenly split between “good” and “poor.” Region 6 evalu-ations of contour-felled logs were evenly balanced.

Table 16—BAER treatment effectiveness ratings from individual fires as provided by interviewees. Totalresponses are listed as percentages in four classes. Only treatments which received three or moreevaluations are included.

Hillslope Treatment Number Excellent Good Fair Poor

- - - - - - - - - - - - - - - - - - Percent - - - - - - - - - - - - - - - - -Aerial Seeding 83 24.1 27.7 27.7 20.5Contour Felling 35 28.6 37.1 14.3 20.0Mulching 12 66.8 16.6 16.6 0.0Ground Seeding 11 9.1 81.8 9.1 0.0Silt Fence 8 37.5 62.5 0.0 0.0Seeding and Fertilizer 4 25.0 0.0 50.0 25.0Rock Grade Stabilizers 3 0.0 33.3 67.7 0.0Contour Trenching 3 67.7 33.3 0.0 0.0Temporary Fencing 3 0.0 67.7 33.3 0.0Straw Wattles 3 33.3 33.3 33.3 0.0Tilling/Ripping 3 33.3 33.3 33.3 0.0

Channel TreatmentsStraw Bale Check Dams 10 30.0 30.0 30.0 10.0Log Grade Stabilizers 10 30.0 30.0 10.0 30.0Channel Debris Clearing 7 0.0 71.4 0.0 28.6Log Dams 5 40.0 60.0 0.0 0.0Rock Grade Stabilizers 3 0.0 33.3 67.7 0.0Straw Wattle Dams 3 33.3 67.7 0.0 0.0

Road TreatmentsCulvert Upgrading 6 6.7 66.6 0.0 16.7Trash Racks 4 50.0 0.0 25.0 25.0

Comparison of unit costs for contour-felled logs(fig. 19) and aerial seeding (fig. 20) shows that aerialseeding was considerably less expensive per unit area.Contour-felling had a wide range of costs due to ter-rain, access, and whether contract or FS labor wasused. Region 5 had an average cost of about $450 ac–1

($1,100 ha–1) (adjusted to 1999 dollars). Regions 4 and6 costs averaged $260 ac–1 ($640 ha–1). Region 1 costsaveraged $165 ac–1 ($410 ha–1), Region 3 costs were$78 ac–1 ($193 ha–1), Region 2 only used contour-felledlogs four times. Some low unit costs for contour-felledlogs were probably due to low density or linear feet perarea of logs. The high unit costs were often due todifficult terrain and expensive crew costs.

Aerial seeding costs ranged from $4 to $115 ac–1

($10 to 284 ha–1) (adjusted to 1999 dollars). Averagecost by Region varied from $25 ac–1 ($62 ha–1) forRegion 3 to $47 ac–1 ($116 ha–1) for Region 2. Region 5used aerial seeding for 65 fires, whereas Region 2 usedaerial seeding for 16 fires.

Channel Treatments—Effectiveness ratings forstraw bale check dams and log grade stabilizers rangedrelatively evenly from “excellent” to “poor” (table 16).While most interviewees (71 percent) thought thatchannel debris clearing effectiveness fell into the “good”category, 29 percent rated it “poor.” Log dams and


Figure 18—BAER spending on the five most expensive hillslopetreatments by Region in 1999 dollars.

straw wattle dams were rated “excellent” or “good” ineffectiveness, and better than rock grade stabilizers.No one considered the effectiveness of these BAERtreatments to be “poor.”

BAER spending on debris basins, straw bale checkdams, and channel debris clearing was about threetimes greater than spending on the other channeltreatments (fig. 21). When comparing the change inuse over the past three decades, straw bale check damswere extensively used only in the 1990’s. BAER spend-ing on debris basins was non-existent in the 1970’s,and doubled each decade from the 1980’s to the 1990’s(debris basins were in use in the 1970’s but fundingcame from sources other than the Forest Service orpostfire emergency treatments). These treatmentswere generally installed in channels to protect down-stream urban areas in California. Interestingly, spend-ing on channel debris clearing decreased five-foldduring the last 30 years, as the value of instreamdebris was realized.

Figure 17—BAER spending on hillslope treatments by decade in 1999 dollars. Treatments are orderedby decreased spending.

contour fellin

g

seeding - aeria

l

seeding and fertili

zer

mulching

seeding - ground

geotextile fa

brics/ geowebbing

contour trenching

temporary fencing/ cattle

exclusion

straw wattle

s

tilling/rip

ping

slash spreading

sand, soil, or g

ravel bags

silt fence

70's

80's

90's

21.7

0

1

2

3

4

5

6

7

8

9

10

BA

ER

SP

EN

DIN

G(1

999

$)

Millions

HILLSLOPE TREATMENT

18.7

0

1

2

3

4

5

6

7

8

9

10

contourfelling

seeding -aerial

seedingand

fertilizer

mulching

Millions

HILLSLOPE TREATMENTS BY REGION

BA

ER

SP

EN

DIN

G (

1999

$)

1 2 4 5 63 1 2 4 5 63 1 2 4 5 63 1 2 4 5 63 1 2 4 5 63seeding -ground


Table 17—Effectiveness ratings for aerial seeding and contour-felling effectiveness as provided by interviewees,sorted by Region. Percentages of total replies in each rating class are shown.

No. ofBAER Hillslope Treatment Region Replies Excellent Good Fair Poor

- - - - - - - - - - - - - - Percent - - - - - - - - - - - - - - -Aerial Seeding 1 8 62.5 12.5 12.5 12.5

2 6 33.3 33.3 0.0 33.33 16 6.3 18.7 37.5 37.54 11 63.6 18.2 0.0 18.25 32 3.0 34.4 43.8 18.86 10 40.0 40.0 20.0 0.0

Contour Felling 1 9 44.4 44.4 11.2 0.02 2 0.0 50.0 0.0 50.03 6 50.0 16.7 16.7 16.64 4 25.0 50.0 0.0 25.05 6 16.7 50.0 0.0 33.36 8 12.5 25.0 37.5 25.0

n=9

n=5

n=14

n=9

n=9 n=21

0

200

400

600

800

1000

1200

1 2 3 4 5 6

REGION

0

500

1000

1500

2000

2500

3000

CO

ST

(1

99

9 $

ac-

1)

CO

ST

(1

99

9 $

ha-1

)

n=29

n=29

n=17

n=66

n=16

n=28

0

20

40

60

80

100

120

140

1 2 3 4 5 6

REGION

0

50

100

150

200

250

300

350

CO

ST

(1

99

9 $

ac-

1)

CO

ST

(1

99

9 $

ha-1

)

Figure 19—Cost per area for BAER spending on contour-felledlogs by Region in 1999 dollars. Mean unit cost with range foreach region is shown.

Figure 20—Cost per area for BAER spending on aerial seedingby Region in 1999 dollars. Mean unit cost with range for eachregion is shown.

Region 5 spent the most on debris basins (1990’s)and channel debris clearing (1970’s) (fig. 22). Strawbale check dams were used mostly by Regions 4, 5 and6. There were far fewer interviewee comments re-corded in our database for channel treatments (38)than for hillslope treatments (168), indicating a lowerfrequency of use in BAER projects.

Road and Trail Treatments—Only two road treat-ments, culvert upgrading and trash racks, received morethan three effectiveness evaluations. The responses cov-ered the range from “excellent” to “poor,” although three-quarters of the interviewees rated culvert upgrading“excellent” or “good” in effectiveness (table 16).Interviewees were evenly split on their assessment oftrash racks as “excellent,” “fair,” or “poor.”

BAER spending on armored ford crossings wasthree times greater than for any other road treat-ment. This was due to the extensive use of armoredcrossing on the 1994 Tyee Fire, Wenachee NationalForest in Washington (fig. 23). Culvert upgrades,ditch maintenance/cleaning and armoring, road rip-ping, drainage improvement and stabilization, andtrail work accounted for the majority of funds spent.In the 1990’s, more funds were spent on ditch main-tenance than during the other two decades combined(adjusted to 1999 dollars). Spending on most otherroad treatments increased during the 1980’s andagain in the 1990’s. Region 5 spent more on roadtreatments, other than armored ford crossing, thanother regions (fig. 24). Region 4 invested the most onditch cleaning and armoring.

Treatment Rankings

The composition of interviewees was examined tosee if different disciplines would rank treatment


Figure 21—BAER spending on channel treatment by decade in 1999 dollars. Treatments are ordered bydecreased spending.

Figure 22—BAER spending on the five most expensive channel treatmentsby Region in 1999 dollars.

straw bale ch

eck dams

debris basin

s

channel d

ebris cl

earing

stream bank/c

hannel arm

oring

log dams

rock ca

ge (gabion) d

ams

log grade stabiliz

ers

in-channel fe

lling

rock grade st

abilizers

stack

ed straw w

attle dams

70's80's

90's

0.00.20.40.60.81.01.21.41.61.8

BA

ER

SP

EN

DIN

G(1

999

$)

Millions

CHANNEL TREATMENTS

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

debris basins straw balecheck dams

channeldebris

clearing

streambank/channel

armoring

log dams

Millions

CHANNEL TREATMENTS BY REGION

BA

ER

SP

EN

DIN

G (

1999

$)

2.2 1.8

1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 1 2 3 4 5 6


armor

ed fo

rd cr

ossin

g

culve

rt up

grad

ing

ditch

- cle

aning

, arm

oring

road

rip, d

rain,

and s

tabiliz

e

rollin

g dips

/wate

r bar

s/cro

ss dr

ain

trash

rack

s

outsl

oping

road

storm

patro

l

culve

rt inl

et/ou

tlet a

rmor

ing

culve

rt re

moval

trail w

ork

other

70's

80's

90's

3.7

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

BA

ER

SP

EN

DIN

G(1

999

$)

Millions

ROAD TREATMENTS

Figure 23—BAER spending on road and trail treatments by decade in 1999 dollars. Treatments areordered by decreased spending except for trail work and other treatment that were not categorized.

Figure 24—BAER spending on the five most expensive road and trailtreatments by Region in 1999 dollars.

3.7

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

armored fordcrossing

culvertupgrading

ditch - cleaning,armoring

road rip, drain,and stabilize

rolling dips/waterbars/cross drain

Millions

ROAD TREATMENTS BY REGION

BA

ER

SP

EN

DIN

G (

19

99

$)

1 2 4 5 63 1 2 4 5 63 1 2 4 5 63 1 2 4 5 63 1 2 4 5 63


preferences differently. There was no difference inrankings from all interviewees (n = 105) compared tothose of soil scientists (n = 29) or hydrologists (n =21), who accounted for the majority of interviewees;therefore, rankings were not stratified by discipline.Interviewees did not name over- or underused treat-ments on every fire.

The overall rankings show that hillslope treat-ments are preferred methods for controlling erosionand runoff after fire, comprising five of the top 10ranked treatments (fig. 25). Contour-felled logs andseeding had scores twice or more as high than anyother treatment. These rankings are reflected in spend-ing on these methods (fig. 17). Road treatments werenext in overall preference, and only one channeltreatment was highly ranked.

Aerial seeding had the highest ranking amonghillslope treatments, followed by contour-felled logs,slash spreading, mulch, and temporary fencing. Othertreatments received relatively low scores. The highrank for seeding is not surprising considering its highlevel of use (fig. 26). On the other hand, aerial seedingwas listed as the most overused treatment by far, withground seeding second (table 18). Seeding also gar-nered a few votes as underused, and it was most oftenmentioned (three times) as a treatment that shouldhave been used on no-action fires. These seeminglycontradictory results reflect the wide differences inopinion about seeding’s effectiveness (table 17) andthe on-going controversies surrounding the use ofgrass seeding as a rehabilitation treatment.

Figure 25—Cumulative ranking of treatment effectiveness forall treatments combined. Cumulative rankings are taken frominterviewees ranking of their top three treatment preferences.The top 14 treatment preferences are shown out of a total of 26treatments.

Figure 26—Cumulative ranking of treatment effectiveness forhillslope treatments. Cumulative rankings are taken frominterviewees ranking of their top three treatment preferences.The top 10 preferences are shown out of a total of 16treatments.

0

5

10

15

20

25

30

35

cont

our f

elling

seed

ing

culve

rt up

grad

ing

slash

spre

ading

mulc

hing

tem

pora

ry fe

ncing

rollin

g dip

s/wat

er b

ars/c

ross

dra

in

storm

pat

rol

straw

bale

chec

k dam

sot

her

culve

rt ris

ers/a

rmor

ing

culve

rt re

mov

al

straw

wat

tles

outsl

oping

road

ALL BAER TREATMENTS

CU

MU

LA

TIV

E R

AN

KIN

G(P

oin

ts)

0

10

20

30

40

50

60

70

80

seed

ing

cont

our f

elling

slash

spre

ading

mulc

hing

tem

pora

ry fe

ncing

straw

wat

tles

cont

our t

renc

hing

tilling

/ripp

ing othe

r

geot

extile

s

HILLSLOPE TREATMENTS

CU

MU

LAT

IVE

RA

NK

ING

(Poi

nts)

Contour-felled logs, the second highest rankedhillslope treatment, was also rated the most oftenunderused treatment on project fires and second moston no-action fires. This treatment received the highestoverall ranking (fig. 26), barely beating seeding, atrend reflected in its increasing popularity in recentyears (fig. 17). However, it was listed as overused twiceand, like seeding, received mixed ratings on effective-ness (table 17).

Among channel treatments, straw bale check damsreceived the highest ranking, followed by log gradestabilizers, rock grade stabilizers, channel clearing,bank and channel armoring, and in-channel felling(fig. 27). Straw bale check dams ranked ninth inoverall preference, the only channel method fallingwithin the top 10 (fig. 25). On the other hand, strawbale check dams were listed as overused twice, morethan any other channel treatment, and were not listedas underused at all (table 18).

Rolling dips or water bars and culvert upgradingwere by far the most preferred road treatments, withstorm patrol next and other methods ranking lower(fig. 28). No road treatments were named as overused,but culvert upgrading was mentioned often as a treat-ment that should have been used on both project(second highest) and no-action (third highest) fires(table 18). It was the third highest ranked treatmentoverall. Storm patrol ranked eighth overall and wasmentioned twice as a treatment that should havebeen used on no-action fires.


Table 18—Underused treatment on no action fires, underused treatment on BAER project fires, and overusedtreatments on BAER project fires from interviewees.

Underused Underused OverusedTreatment No Action Fires BAER Project Fires BAER Project Fires

Seeding, ground 1 3Contour-felling 2 8 2Straw bale checkdams 2Seeding, aerial 3 1 10Debris basins 1 1 1Silt fence 2 1Tilling/ripping 1 1Geotextile fabrics 1 1Stream bank/channel armoring 1Seeding plus fertilizer 1Culvert upgrading 2 4Storm patrol 2Mulching 1 4Cross drain ditches 2 1Exclusion 1Channel debris clearing 1Outsloping road 1Log dams 1Log grade stabilizers 1Ditch maintenance-cleaning, armoring 1Straw wattles 1Sand, soil, or gravel bags 1

0

10

20

30

40

50

60

straw

bale

chec

k dam

s

log g

rade

stab

ilizer

s

rock

gra

de st

abiliz

ers

chan

nel d

ebris

clea

ring

strea

m b

ank/c

hann

el ar

mor

ing

in-ch

anne

l fellin

g

log d

ams

debr

is ba

sins

straw

wat

tle d

ams

rock

cage

(gab

ion) d

ams

CHANNEL TREATMENTS

CU

MU

LAT

IVE

RA

NK

ING

(Poi

nts)

0

10

20

30

40

50

60

70

80

rollin

g dip

s/wat

er b

ars/c

ross

dra

in

culve

rt up

grad

ing

storm

pat

rol

culve

rt inl

et/o

utlet

arm

oring

/rise

rs

trash

rack

s

culve

rt re

mov

al

ditch

- cle

aning

, arm

oring

arm

ored

ford

cros

sing

outsl

oping

road

othe

r

ROAD TREATMENTS

CU

MU

LA

TIV

E R

AN

KIN

G(P

oin

ts)

Figure 28—Cumulative ranking of treatment effectiveness forroad treatments. Cumulative rankings are taken frominterviewees ranking of their top three treatment preferences.The top 10 preferences are shown out of a total of 15 treatments.

Figure 27—Cumulative ranking of treatment effectiveness forchannel treatments. Cumulative rankings are taken frominterviewees ranking of their top three treatment preferences.Rankings of all preferences are shown.


Discussion _____________________The BAER evaluation process provides a means to

assess the postfire emergency and identify appropri-ate treatments. Although our original intent was toevaluate treatment effectiveness, our efforts to com-pile information on individual fires produced a largedatabase of information on the BAER assessmentprocess itself. Treatment effectiveness depends inpart upon appropriate treatment selection, and thatdepends on accurately identifying the emergency con-dition. Our interviews revealed that some treatmentsare overused and others could be applied more often.We discuss the implications of our findings from re-view of BAER assessments, then evaluate treatmentmethods.

BAER Assessments

Total BAER expenditures during the last threedecades (adjusted to 1999 dollars) were greater than$83 million, with over 60 percent occurring in the1990’s. This was due to several large fires, theirproximity to urban/wildland interface, and increasedvalues at risk, promoting greater protection. Duringthe last three decades, over 3.8 million ac (1.5 millionha) of Forest Service land were burned. Of that, highseverity burned areas has increased from 195,000 ac(79,000 ha) in the 1970’s to over 655,000 ac (265,000ha) in the 1990’s. Flooding and sedimentation risk isgreater from areas with high severity burns. Thusmore money has been spent to try to reduce the threatto downstream values. Most of the increase in spend-ing in the 1990’s was due to high profile fires thatthreatened urban areas (table 9).

BAER teams assign erosion hazard ratings to vari-ous portions of a burned area based on local geology,soil type, topography, burn severity, expected stormduration and intensity, and local experience withpostfire conditions. Improvements in erosion hazardrating could be accomplished by better fire severitymapping with infrared flights and satellite imageryafter the fire (Lachowski and others 1997). Thesemethods, though still in development, have shownpromise for providing better burn area-wide severityassessment. Methods used to calculate erosion poten-tial and sediment yields were not consistent, and insome cases the estimates made could be consideredunreasonable. For example, erosion rates of 1000 t ac–1

(2200 Mg ha–1) and sediment yields of 0.1 millionyd3 mi–2 (0.03 million m3 km–2) were projected onseveral fires. Considering that our review of publishedliterature found reported erosion rates no higher than165 t ac–1 (370 mg ha–1) even from steep chaparralslopes (Hendricks and Johnson 1944). This suggeststhat assumptions about erosion potential used for

those calculations are inaccurate. Uncritical review ofthe erosion potential estimates by the BAER teamleaders must also have occurred. Refinement of thecalculation methods and better training on how to dothese calculations appears warranted.

Most BAER treatments were designed for a 10-yearor 25-year return interval event indicating that treat-ments were designed for major storm events. Thus,the tolerance for high peakflows and excess sedimentwas low. Design storm estimated peakflow changeswere not well correlated with infiltration reduction(fig. 15). A 10 percent reduction in infiltration is notlikely to cause a 10,000 percent or great increase inpeakflows. It is more realistic to expect that magni-tude of increase from infiltration reduction of 80 to100 percent. From our literature review, actual in-creases in peakflows due to wildfires can range over3 to 4 orders of magnitude (Anderson and others 1976,Glendening and others 1961). Hibbart (1971) reporteda 9,600 percent increase in peakflows in chaparralafter a severe wildfire. Although high peakflow in-creases occur due to infiltration reduction and waterrepellent soil conditions in some forest types, designstorm peakflow estimation techniques need to be re-fined and better documented to reflect the realities ofwatershed response to severe wildfire.

According to the Burned Area Reports we collected,water repellent soil conditions are more widespreadafter fire than previously reported (fig. 13; DeBanoand others 1998). Existing research suggests thatwater repellency is usually found on coarse-texturedsoils, especially under chaparral or other vegetationwith high levels of volatile organic compounds in thelitter (DeBano and others 1979b, DeBano and others1998). Our dataset included reports of water repellentconditions across all soil and vegetation types. Unfor-tunately, the information given on the Burned AreaReports did not allow us to analyze what methods wereused to determine soil water repellent conditions (thusassessing the accuracy of the estimates) or how exten-sive the sampling was for the water repellent areadeterminations. These results identify a need for addi-tional research on the extent and severity of waterrepellent soil conditions and its affect on infiltrationafter wildfire in the Western United States.

Quantifying the watershed degradation threat isdifficult. Threats to life and property, water quality,and soil productivity were the main reasons given forproposing BAER treatments. The more urban ForestService Regions listed threats to property as a reasonfor BAER treatment 50 percent of the time. As devel-opment in foothill areas increases, the need to treatburned areas to reduce the risk to property and life willlikely increase as well. The role of flood plains duringflood and debris flows needs to be emphasized.


Water quality issues include effects on aquatic habi-tat, sedimentation in channels and reservoirs, andeffects on drinking water. Several monitoring studiesfound impacts to aquatic ecosystems that occurred inthe first year after the fire or in the first majorstorm. Increases in stream turbidity with high rain-fall were documented in Regions 1 and 6 (Amaranthus1990, McCammon 1980, Story 1994). Large flood anddebris flow events cleared streams of fish after the1984 North Hills fire, Helena National Forest, Mon-tana (Schultz and others 1986) and the 1990 Dudefire, Tonto National Forest, Arizona (Rinne 1996). Inboth cases the populations of at least some speciesrecovered surprisingly quickly, however. Threats todeveloped water sources can be quantified relativelyeasily, because managers know how much it wouldcost to treat turbid water or remove sediment from areservoir.

It is difficult to assess the potential for loss of soilproductivity after fire, because there is no easy way ofcalculating a long-term productivity decline resultingfrom the loss of soil material or nutrients. This isparticularly the case where there are not obviouslosses of large amounts of organic matter and mineralsoil. Depending on fire severity, soil productivitychanges can be either beneficial or deleterious. Short-term increases in plant productivity can occur fromsoil changes such as the mineralization of nutrientstied up in organic matter (DeBano and others 1998,Neary and others 1999). Predicting productivitychanges for long rotation forest stands is difficult,however, because of the many interacting factors whichaffect long-term productivity and the lack of adequateinformation to make long-range predictions (Powersand others 1990).

Site productivity changes can be long-term or tem-porary. If a fire is within the natural range of variationfor an ecosystem, productivity changes should be short-term and acceptable since fire is a natural componentin many ecosystems. If a fire is outside of the naturalrange of variation and intensity, particularly due tohuman interference with forest ecosystems, long-termsoil productivity is more likely to be at risk.

Various methods have been used in the BAER pro-cess to estimate the cost of potential changes in soilproductivity after fire. For example, the value of soilloss has been based on estimated site index changesdue to the fire and the consequent potential loss inharvestable timber during the next regeneration cycle,or based on the cost of replacement top soil if pur-chased commercially. Most Burned Area Reports didnot state how loss estimates were made. Methods thatconsider only the value of harvestable timber mayunderestimate the consequences of site productivityloss to other ecosystem components.

Instead of trying to justify BAER treatments byestimating some future loss in site productivity val-ues (merchantable timber), a better approach wouldbe to identify situations where future productivity ispotentially threatened by the loss of large amounts ofabove ground organic matter in severe fires (Nearyand others 1999) or losses of surface soil horizons(DeBano and others 1998). While both affect long-term productivity, only the latter can be affected inthe short-term by BAER treatments. Until bettermethods can be developed to estimate long-termchanges in productivity after wildfires, the profes-sional judgment of soil scientists is the best tool fordetermining the need for treatments to mitigate soilproductivity losses.

Probability of success stated in the BAER reportswas always high (average 69 percent for hillslopetreatments, 74 percent for a channel treatments, and86 percent for road treatments the first year after fire;table 12). BAER teams are apparently very enthusias-tic and optimistic that the BAER goals can be met andthat the implemented treatments will work—a “cando” attitude, similar to that in fire fighting, prevails.This result should be expected, because only knowneffective treatments are supposed to be used for emer-gency watershed rehabilitation (USDA Forest Service1995).

Results of our interviews suggest that these prob-abilities may be overestimated for some treatments.For example, only 52 percent of interviewees felt thataerial seeding, the most extensively used hillslopetreatment was “good” or “excellent” in effectiveness(a reasonable definition of success), and only 56percent of quantitative monitoring reports consid-ered seeding effective the first year after fire. On theother hand, 79 percent of the nonquantitative reportsconsidered it successful, justifying the high probabil-ity of success. Other treatments fared better in theeffectiveness ratings. “Good” or “excellent” ratingswere given to about 66 percent of contour-felled logprojects, 83 percent of mulch projects, and a whopping91 percent of ground seeding efforts. Monitoring re-ports also found contour-felled logs to be successfulmost of the time. These subjective results suggest thatprobability of success may be overstated for aerialseeding in many reports, but may be more realistic forother hillslope treatments. However, seeding is theonly method for which a significant amount of post-fire research has been conducted (discussed furtherbelow). For other hillslope methods, hard data toevaluate effectiveness—and thus the probability ofsuccess—are scarce.

Among channel treatments, “good” or “excellent”ratings were given to 60 percent of straw bale checkdam and log grade stabilizer projects, while channel


debris clearing was closer to the Burned Area Reportaverage with 71 percent. On the other hand, 69 per-cent of monitoring reports on straw bale check damswere favorable. For major channel treatments, BAERteams appear to be making fairly reasonable projec-tions of success. Too few road treatments were ratedfor effectiveness or evaluated in monitoring or re-search reports to evaluate the reliability of successprojections for those treatments.

Not only were BAER treatments expected to besuccessful, they were projected to save million ofdollars in damages. For every $1 spent on treatments,$10 to $200 in losses was proposed to be saved (fig. 16).These estimates were made with very few data toverify the effectiveness of most BAER treatments.Based on our results, projected benefits from aerialseeding may need to be adjusted downward to reflectlower realistic probabilities of first-year success.Sullivan and others (1987) suggested that a highprobability of success is required for a treatment to beeconomically cost effective.

As the cost of action or no action alternatives arebased on professional judgment and past experience,they are very approximate. It might be better to usethese estimates to rank treatment options. They donot provide real dollar values of what might happen,suggesting that an alternative ranking system mightbe preferable to compare treatment alternativesand no treatment options. Ranking could be basedon actual damages that occurred in nearby similarwatersheds.

BAER teams contained soil scientists and hydrolo-gists most of the time, with a wide range of otherdisciplines represented as needed on particular fires.Although wildlife biologists were often on teams, ecolo-gists were included relatively infrequently except inRegion 4 (table 13). Many monitoring reports andinterviewees identified a need for better informationon the ecosystem impacts of fire and vegetation recov-ery potential (discussed further below) when evaluat-ing the necessity for emergency treatments. In manycases, natural revegetation of burned areas occurredmore quickly than expected. Including ecologists andbotanists on BAER teams more frequently might helpto better assess natural recovery potential.

BAER Project Monitoring

Monitoring of BAER projects has been done for awide variety of reasons. Consequently, there was nostandard format or content to the monitoring docu-ments we collected. The most common type was amemo reporting on a trip to visually assess the resultsof BAER treatments or natural recovery after fire.These reports provided qualitative evaluations oftreatment effectiveness and watershed condition,

but relatively few quantitative data. Until 1998, therewas no funding specifically available for postimple-mentation monitoring of BAER treatments. Any moni-toring had to be done out of Forest Service appropri-ated funds. Thus the trip reports were probably allthat could be squeezed into the normal plan of work onbusy National Forests.

Most of the reports fell into the categories of “imple-mentation” or “effectiveness” monitoring. They as-sessed whether treatments, especially structures suchas straw bale check dams or contour-felled logs, wereproperly installed and operating as designed. In thecase of structures, accumulation of sediment behindthe barrier, structural integrity after the first winter,and lack of flooding or sedimentation problems down-stream were generally regarded as indications of treat-ment success. Seeding operations were regarded assuccessful if the seeded species were observed to begrowing well. Most monitoring was done a few monthsto 1 or 2 years after a fire. Only a few National Forestsmonitored projects lasted longer than that. The im-pacts of treatment on the emergency condition can beevaluated in this time period, but the ecosystemsimpacts of treatments, especially of seeding on nativeplant recovery, may not be adequately assessed.

Where quantitative data were collected, details otherthan the variables being measured were often omittedfrom reports (which were generally intended for inter-nal use). For monitoring results to be informative forothers with similar soils or vegetation types, detailssuch as soil type and texture, slope angle, aspect,watershed or analysis area size, fire severity indica-tors, and other variables should be included in reports.Where treated areas are compared to untreated areas,it is especially important to know how comparable thesites are in other physical and biological attributes.These data are relatively easy to collect in most cases.Quantitative reports also often noted that measure-ments were made in “typical” areas, with no intentionof providing statistical sufficiency. Some descriptionof how “typical” was determined or how representativethe sample plots were of the overall fire area wouldmake the results more useful to future investigators,both on and off the specific National Forest. The lownumber of samples taken in most efforts may haveresulted in overstatement of treatment impacts (aseither effective or ineffective), because inherent sitevariability is not captured in the results. With greaterfunding available for monitoring, this limitation maybe alleviated in the future.

Quantitative monitoring efforts were generally re-stricted to very small areas of a fire, while the qualita-tive trip reports analyzed a much larger proportion ofthe burned area in less detail. Both kinds of reportshave obvious value for assessing the results of BAERprojects. We found few cases where both kinds of


monitoring were done on a given fire. This may be aresult of the incompleteness of our record, or it couldreflect the fact that National Forests could afford to doone or the other kind of monitoring, but not both. Theinterests of the personnel charged with monitoringmay also have determined the type of monitoring thatwas done.

Because BAER treatments are generally designedto reduce erosion, sedimentation, and flooding, themost valuable assessments of treatment effectivenesswould be those that actually quantify sediment move-ment and water yield. Relatively few reports mea-sured sediment movement, and virtually none tried toquantify water yield. Methods used for measuringsediment movement ranged from erosion bridges,which measure change in the distance to groundsurface from a fixed suspended bar, to height of ero-sion pedestals left after sediment movement occurred,to sediment traps such as troughs and silt fencesinstalled below hillsides or in small swales. Erosionbridge results generally proved difficult to evaluate,because sediment was as likely to be deposited on aspot (eroded from above) as removed. Pedestal mea-surement was considered to overstate erosion, be-cause it is measured only in places where sedimentloss has obviously occurred and cannot easily be gen-eralized to a larger area. Traps and silt fences pro-vided the most informative results, although theirtendency to overtop made many measurements mini-mum estimates rather than actual quantities. In addi-tion, it is difficult to determine the size of the areaactually contributing to a trap or fence. If fixed areaplots above a trap are used, the plot boundaries mayaffect sediment movement. Most reports using thesemethods did not tell how contributing area was de-termined for the “tons per acre” sediment outputcalculation.

Because monitoring results can become the basis forfuture management decisions, it is critical that moni-toring efforts and reports be as scientifically credibleas possible. Whether defending a decision to seed orexplaining why a flood occurred despite BAER treat-ments, Forests need to be able to support their workwith good data from their own and other Forests’monitoring efforts. There is little published researchon most BAER treatments. With the limitations ofmonitoring reports mentioned above, we did not feelthat we could evaluate the validity of most reports, letalone generalize the results of monitoring done on oneForest to another area. There is a critical need for moreand better monitoring of BAER treatments (discussedfurther in the Recommendations section).

Most monitoring focused on the most expensive(stream channel treatments, contour-felled logs), wide-spread (seeding), or controversial (seeding) treatmentapplied after a fire. The results from these efforts are

incorporated into our discussions of specific treatmenteffectiveness.

Treatment Effectiveness

The basis for the BAER program is whether treat-ments effectively ameliorate postfire emergency con-ditions without compromising ecosystem recovery.For many treatment methods, effectiveness could onlybe determined qualitatively. From our interviews andthe monitoring reports, it became apparent that treat-ment success often depended on appropriate imple-mentation (see appendix B) and cooperative postfireweather. Quantitative data on effectiveness were avail-able for relatively few treatments. We were able toanalyze hillslope treatments in more detail than chan-nel or road treatments.

Hillslope Treatments—Increasing infiltration ofrain water and preventing soil from leaving the hillslopeare considered the most effective methods to slowrunoff, reduce flood peaks, retain site productivity,and reduce downstream sedimentation. Mulching andgeotextiles were rated the most effective hillslopetreatments by our interviewees, because they provideimmediate ground cover to reduce raindrop impactand hold soil in place. Postfire research and monitor-ing reports showed dramatic decreases in sedimentmovement where mulch was applied (Bautista andothers 1996, Faust 1998). However, both methods arerelatively costly and are difficult or impossible toinstall in remote locations (appendix B). Mulch is mostuseful near roads or in critical areas at the tops ofslopes. Geotextiles are generally applied to small ar-eas, such as road cuts and fills. Aerial seeding andcontour-felled logs are the two most common hillslopetreatments. Their effectiveness in reducing erosionhad mixed reviews from the published literature,monitoring reports, and interview results, even thoughthe Forest Service spent over $25 million in the lastthree decades on each treatment.

Little contour-felling was implemented in the 1970’s,and only $4 million was spent in the 1980’s. Sincethen, however, contour-felled logs have gained inpopularity as a hillslope treatment. Most intervieweesthought the effectiveness was good or excellent. Moni-toring studies did not evaluate runoff, infiltration, orsediment movement changes due to the contour-felledlogs; they only reported sediment storage. Monitor-ing studies indicated that contour-felling could beabout 60 percent efficient (DeGraff 1982) and couldreduce downslope sedimentation by about 40 to 60 per-cent (Griffith 1989a). Maximum trapped sediment of6.7 yd3 ac–1 (13 m3 ha–1) or about 6.8 t ac–1 (17 Mg ha–1)by contour-felled logs was reported by Miles andothers (1989). McCammon and Hughes (1980), on theother hand, estimated storage at about 72 yd3 ac–1


(135 m3 ha–1), using a high density of logs. If first-yearannual erosion rates vary from 0.004 to 150 t ac–1

(0.01 to 370 Mg ha–1), then they could trap 5 to 47percent of 150 t ac–1 (370 Mg ha–1) of sediment,depending on the density of the logs. Beyond that theywould not be cost effective from a sediment-holdingcapacity analysis. This wide range of effectivenessindicates the need of proper estimation techniques ofthe erosion potential, and for properly designing con-tour-felled log installations in terms of log numbersand spacing. For example, if you can trap 60 percent orgreater then they are probably cost effective, but if youare only trapping 5 or 10 percent of the expectedsediment production, then it may not be worth theeffort for such small amount of sediment storageability. Contour-felled logs do provide immediate ben-efits after installation, in that they trap sedimentduring the first postfire year, which usually has thehighest erosion rates.

The ability of this treatment to reduce runoff,rilling, increase infiltration and decrease downstreamtime to peak (slowing velocities) has not been docu-mented, even though these are reasons often givenfor doing contour felling. If contour-felled logs slow oreliminate runoff, sediment movement may not occur.Therefore, measuring sediment accumulation behindthe logs may not be the best method for assessingtheir effectiveness. Quantifying sediment and wateroutput from a watershed are the best ways to trulyevaluate the effectiveness of contour-felled logs, butthis kind of research and monitoring is expensive anddifficult to do.

Contour-felled logs will channel flow if not installedcorrectly on the contour with good ground contact. There-fore, proper training, contract inspections, and close moni-toring during installation are critical to success, as wasrepeatedly pointed out by interviewees (appendix B).

Grass seeding is the most widely used and beststudied BAER treatment. Our interviewees rankedseeding second highest in overall treatment prefer-ence, despite giving it mixed reviews for effectivenessand citing it as overused more often than any othertreatment. Expenditures for seeding declined some-what in recent years (fig. 17). However, seeding re-mains the only method available to treat large areas ata reasonably low cost per acre.

How likely is seeding to increase plant cover orreduce erosion, in either the first growing season orlater? We tabulated results from published studies (intables 6 and 7) to determine rough probabilities ofseeding “success” in the first and second years afterfire (table 19). Only studies that evaluated compa-rable seeded and unseeded plots were included. Dis-tinct research sites within a single paper were treatedas unique “studies” for this comparison. Because fewresearchers measured erosion, we used vegetationcover as an indicator of potential erosion control effec-tiveness. Previous work found that 60 percent groundcover reduced sediment movement to negligibleamounts, and 30 percent cover reduced erosion byabout half compared to bare ground (Noble 1965, Orr1970). We used these levels as indicators of effective orpartly effective watershed protection, respectively,from seeded and/or natural vegetation.

Table 19—Numbers of published studies reporting measures of seeding “success” by native vegetation type during the first 2 years following fire.

Pubs. Showing Those Showing % of Pubs. Showing % of Pubs. Showing Pubs. Showing Those ShowingCover Measure- Seeding Increased >30 % Cover >60 % Cover Erosion Measure- Seeding Reduced

ments1 Cover Seeded Unseeded Seeded Unseeded ments Erosion

- - - - - - - - - - - - - - No. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Percent - - - - - - - - - - - - - - - - - - - - - - - - - - - No. - - - - - - - - - - - -Postfire Year One

Chaparral10 4 50 50 30 20 7 1

Conifer9 5 33 0 22 0 1 0

Combined19 9 42 26 26 10.5 8 1

Postfire Year TwoChaparral

7 2 86 86 86 43 6 1Conifer

11 6 73 55 36 0 3 1Combined

18 8 78 67 56 17 9 2

1All studies contained seeded and unseeded plots and reported plant production as percent cover at the end of the growing season. Only statistically significantincreases in cover or reductions in erosion are tabulated (Amaranthus 1989, Amaranthus and others 1993, Anderson and Brooks 1975, Beyers and others 1998a, Conardand others 1995, Gautier 1983, Geier-Hayes 1997, Griffin 1982, Rice and others 1965, Roby 1989, Taskey and others 1989, Van de Water 1998, Tiedemann and Klock1973, 1976).


Seeding significantly increased total plant cover 47percent of the time by the end of the first growingseason after fire (table 19). Forty-two percent ofseeded sites had at least 30 percent cover, comparedto 26 percent of unseeded. Only 26 percent of seededsites had at least 60 percent cover versus 10.5 percentof unseeded. Using vegetation cover as an indicator,therefore, the probability of seeding providing effec-tive watershed protection by the end of the firstgrowing season was just 26 percent, but that wasmore than twice the probability that an untreatedsite would be stable.

Erosion was decreased by seeding in only one out ofeight first-year studies (12.5 percent). Erosion mea-surements have high variability, and several of thestudies showed a trend toward lower sediment move-ment on seeded plots that was not statistically signifi-cant (e.g., Amaranthus 1989, Wohlgemuth and others1998). The low occurrence of erosion effects is notsurprising, however, considering that much of thesediment movement occurs before plant cover is estab-lished. Krammes (1960), in southern California, foundthat as much as 90 percent of first-year postfire hillslopesediment movement can occur as dry ravel before thefirst germination-stimulating rains even occur.Amaranthus (1989) measured most first-year sedi-ment movement on his Oregon study site during sev-eral storms in December, before the seeded ryegrasshad produced much cover.

In the second year after fire, seeded sites had greatertotal cover (plant and litter) than unseeded 42 percentof the time (table 19). Half of the studies measurederosion, which was significantly lower on seeded sites22 percent of the time. Greater cover, therefore, didnot always produce less erosion. The proportion ofsites with at least 30 percent cover was 78 percent and67 percent of seeded and unseeded plots, respectively.More than half (56 percent) of all seeded sites wereessentially stabilized (at least 60 percent cover), com-pared to only 17 percent of unseeded sites. Thusseeded slopes were three times more likely to be stableafter 2 years than unseeded slopes, though seedingstill had only a 56 percent probability of “success” ifsuccess means “effective” (60 percent) cover.

Published reports from chaparral and conifer sitesdiffered somewhat in response to seeding (table 19).Seeding was less likely to increase cover the first yearon chaparral sites than conifer sites. Half of bothseeded and unseeded chaparral sites had at leastpartially effective cover after 1 year, compared toonly 33 percent of seeded conifer sites and none ofthe unseeded. However, the only study reportingless erosion on seeded plots the first year after firewas from a chaparral site seeded with annual ryegrass(Gautier 1983). The same trend was evident in studiesreporting second-year results (table 19).

The study sites in these publications varied widelyin soil type, percent slope (table 6, 7), annual precipi-tation, rainfall pattern, and prefire plant commu-nity, as well as seeding mix, so that lumping themtogether masks important factors affecting coverdevelopment and erosion. The total cover value tal-lied in the published studies sometimes includedlitter, sometimes not; thus, the number of partiallyand effectively stabilized sites in the second yearcould be underestimated.

We made several generalizations from this tabula-tion. First, plant cover developed relatively rapidly onthe chaparral sites examined, so that seeding was lesslikely to make a difference in total cover in chaparralthan on conifer sites. Second, most of the studiedchaparral sites were seeded with annual ryegrass,while the conifer sites tended to be treated with amixture of perennial pasture grasses, and increasedcover due to seeding was more likely to show up in thefirst year on chaparral sites and in the second year onconifer sites. Third, even if treatment “success” isdefined as at least 60 percent total cover at the end ofthe growing season, rather than as an actual mea-sured reduction in sediment movement, seeding had alow probability of success during the first year afterfire, when most of the erosion occurs (Robichaud andBrown 1999, Wells 1981), and continued to have a lowprobability of success on conifer sites in the secondyear. On the basis of these published results, BurnedArea Reports that project 60 to 80 percent first-yearsuccess for seeding operations are greatly exaggerat-ing the potential benefits of treatment.

A similar tabulation was made from quantitativemonitoring reports, although most of them did notdirectly compare seeded and unseeded plots (table 20).Where they were directly compared, seeded plots hadgreater cover than unseeded plots 64 percent of thetime at the end of the first growing season after fire,though the differences were not tested for statisticalsignificance (table 14). A higher proportion of first-year monitoring studies, compared to published stud-ies, showed apparent reductions in erosion (43 per-cent) as well, although, again, differences in sedimentproduction were not analyzed statistically. Some ofthe comparisons involved only one or two monitoringpoints per treatment. Seeded plots were more likelyto have at least 30 percent cover after one growingseason in the monitoring studies than in the publishedstudies (74 percent vs 42 percent), possibly becausemore of them reported on sites seeded with quick-growing cereal grains or annual ryegrass rather thanperennial pasture grasses. The probability of finding“effective” (at least 60 percent) cover at the end of thefirst growing season was only slightly greater (35 per-cent) than in the published studies.


Interviewees and monitoring reports alike acknowl-edged that the major benefits of seeding are not appar-ent until the second year after fire, because, as notedabove (Amaranthus 1989), most of the growth byseeded grasses takes place after first year damagingstorms have occurred. From the Los Padres NationalForest: “As is typical, the seeding [annual ryegrassand lana vetch] did not significantly control erosionduring the first rainy season. Seeds did not germinateuntil after steady precipitation, and did not growsignificantly until after warm spring weather. Theseeded species are expected to be of greatest valueduring the second and third rainy seasons” (Esplinand Shackleford 1978), when plant litter produced bythe first year’s growth covers the soil. Rainfall thatfirst winter was the second highest on record andresulted in approximately 125 yd3 ac–1 (240 m3 ha–1) ofsoil eroded, despite the fact that seeding was “success-ful” by most criteria, tripling average plant biomasscompared to unseeded areas by the end of the firstgrowing season (Esplin and Shackleford 1978). Onereport suggested that measures other than seedingshould be used in places where first-year control ofsediment movement is critical (Ruby 1997). The in-creased use of contour-felled logs in recent years prob-ably reflects this knowledge.

Seeding is often most successful where it may beneeded least—on gentle slopes and in riparian areas.Janicki (1989) found that two-thirds of plots with morethan 30 percent annual ryegrass cover were on slopesof less than 35 percent. He also noted “observations ofgrass plants concentrated in drainage bottoms sug-gest that seed washed off the slope with the first two

storm events.” Concentration of seeded species at thebase of slopes was also observed by Loftin and others(1998). Some published papers and most monitoringreports did not give slope angles for study sites, mak-ing interpretation of varying success levels difficult.Several interviewees suggested that seeding was un-necessary in riparian areas, because native vegetationthere usually recovers rapidly. On the other hand,other published papers and monitoring reports sug-gested quickly establishing strips of vegetation alongthe margins of streams as one of the best ways toreduce sediment transport into watercourses. Carefulassessment of vegetation regrowth potential duringthe BAER evaluation could help resolve this apparentcontradiction.

Interviewees observed that first-year seeding suc-cess is highly dependent on rainfall pattern. Gentlerain before the first intense storm is needed to stimu-late germination; then enough rain is needed forseeded species to survive. These conditions are morelikely to be met in some areas of the Western UnitedStates than others. Seeding may be particularly riskyin the Southwest (Region 3), where intense monsoonrains follow the early summer fire season. Areaswhere seeding is more often considered “excellent” or“good” may be those where rainfall lasts longer throughthe year (e.g., all but July and August in the PacificNorthwest) or where a significant portion of the an-nual total occurs in summer (e.g., about 30 percent inareas such as Montana, northern Idaho, and north-eastern Washington). California (Region 5) has a longdry season and unpredictable early fall rains, makinggrass establishment less likely to be successful.

Table 20—Numbers of monitoring reports listing measures of seeding “success” by native vegetation type during the first 2 years following fire.

Reports Showing Those Showing % of Reports Showing % of Reports Showing Reports Showing Those ShowingCover Seeding Increased >30 % Cover >60 % Cover Erosion Seeding Reduced

Measurements1 Cover Seeded Unseeded Seeded Unseeded Measurements Erosion

- - - - - - - - - - - - - - No. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Percent - - - - - - - - - - - - - - - - - - - - - - - - - - - No. - - - - - - - - - - - -Postfire Year One

Chaparral7 4 85 25 38 12 3 1

Conifer4 3 60 60 30 0 4 2

Combined11 7 74 38 35 8 7 3

Postfire Year TwoChaparral

2 0 67 67 33 33 2 0Conifer

0 0 80 100 20 100 1 1Combined

2 0 75 75 25 50 3 1

1The first two columns report only studies that contained both seeded and unseeded plots. The middle four columns summarize all studies that contained percentvegetation cover data. The last two columns report only studies that compared erosion between seeded and unseeded plots. Statistical significance was not tested inthese studies.


Research, monitoring reports, and interview com-ments all suggest that “successful” grass establish-ment displaces some native plant regeneration. Thiswas the goal of past range “reseeding” projects—producing useful livestock and wildlife forage on landthat would not contain harvestable timber for decadesand would otherwise produce nothing but “weeds”—and aggressive, persistent grass species were deliber-ately chosen for seeding (Christ 1934, Evanko 1955,Friedrich 1947, McClure 1956, Stewart 1973). Sup-pression of native plant regeneration could potentiallyreduce browse species for wildlife, reduce watershedprotection in chaparral, and limit the seed bank con-tributions of annual and short-lived perennial “fire-followers” in chaparral and Southwestern ecosys-tems (Conard and others 1995, Keeler-Wolf 1995,Keeley and others 1981, Loftin and others 1998).There is no published research that quantifies thelong-term impacts of postfire seeding on native plants,but one monitoring observed that weeping lovegrass(Eragrostis curvula) in the Southwest can effectivelysuppress native vegetation for years (Loftin and oth-ers 1998).

In our interviews, forest silviculturists expressedmajor concerns about the impacts of grass seeding onconifer regeneration. The dilemma between erosionreduction and conifer growth is well recognized: “Sincegranitics are inherently good tree-growing sites, aswell as being extremely erodible when burned, thechoice between immediate reforestation and long-term productivity can be a difficult one” (Van de Water1998, p. 28). Better understanding of the impacts offire and erosion on soil productivity would help ad-dress this problem.

Current USDA Forest Service guidelines promotethe use of native species for revegetation projectswherever practical. Interviewees commented thatnative grasses are expensive and not widely availablein the quantities necessary for postfire seeding projects,and developing seed sources that can provide a rangeof locally adapted genotypes is difficult (Van de Water1998). In addition, well-adapted native perennialgrasses could provide as much or more competitionwith conifers as the non-native species currently inuse. For example, the native Southwestern grass Ari-zona fescue (Festuca arizonica) greatly reduced thegrowth of conifer seedlings (Pearson 1942, Rietveld1975). One BAER team included the cost of usingherbicide for seeded grass control in BAER calcula-tions and, as a result, decided against using a well-adapted native grass and chose to seed cereal barley(Hordeum vulgare) instead (Griffith 1998). The barleydied out after 1 year except where disturbed by salvagelogging (Griffith 1993; tables 14, 15).

Seeded grasses can benefit conifer seedlings if theyexclude more competitive vegetation, such as shrubs

(Amaranthus and others 1993, McDonald 1986).Once conifer seedlings are well established, grasscover is less detrimental to their growth than shrubcompetition (McDonald and Oliver 1984, McDonald1986). If grass cover is not too thick, it can potentiallybenefit tree seedlings. Green (1990) observed over 90percent survival of Douglas fir seedlings on a siteseeded after fire with cereal rye (Secale cereale). Therye, at a density of 9 plants ft–2 (100 plants m–2),provided shade to the seedlings during the first yearafter fire. The rye decreased to less than 3 plants ft–2

(33 plants m–2) the second year and essentially disap-peared in the third.

Cereal grains such as barley, cereal rye, oats (Avenasativa), and winter wheat (Tricium aestivum) appearto show great promise for producing cover that doesnot persist. Annual ryegrass was expected to behavethis way, but it proved persistent beyond a couple ofyears in some ecosystems (Barro and Conard 1987,Griffith 1998) and often produces maximum cover thesecond year after fire, rather than the first (Beyers andothers 1998; compare Janicki 1989 with Conard andothers 1991). A few reports cited initial concerns overthe impacts of cereal grains on native regenerationthat disappeared after further monitoring (e.g.,Callahan and Baker 1997, Hanes and Callahan 1995,1996, Van Zuuk 1997). Some cereal grains may exhibitallelopathy, inhibiting competing plant growth chemi-cally (Went and others 1952), but this has not beeninvestigated under field conditions. Clearly more re-search on and monitoring of postfire cereal grainseeding is needed, especially regarding the impacts onnative herbaceous plants and conifer seedlings.

In many cases natural regeneration provided asmuch cover as seeded species during the first yearsafter fire, but good methods for assessing native seedbank viability are lacking (Isle 1998, Loftin and others1998). One standard test for seed bank viability onlyidentifies large-seeded species (by sieving them frompostfire soil samples) or those that germinate quickly(7 to 10 day greenhouse germination test) (Dyer 1995).Species that will provide cover later in the winter or inthe second growing season—the same time that seededgrasses provide most of their cover—are not detectedby this method if they have tiny seeds or cold require-ments for germination. Better understanding of thenatural range of vegetation response to fire wouldincrease our ability to predict whether seeding isreally necessary (Loftin and others 1998, Tyrrel 1981).

Several interviewees suggested that more flexibil-ity in choosing seed mixes be allowed in BAER projects,including the use of quick-growing annuals for ero-sion control and slower growing native perennials forlong-term ecosystem restoration, particularly nativerange. At present, BAER guidelines stress the use ofonly proven erosion-control species for emergency


rehabilitation. Other interviewees expressed con-cern that grass seeding may introduce noxious weedspecies even in certified seed.

Little evidence suggests that fertilizer applied withseeded grass is effective in increasing cover or reduc-ing erosion after fire. Flight strips from the aircraftthat applied the fertilizer were visible as brightergreen stripes on the ground in some areas, and seededgrasses were twice as tall in the fertilized strips thanin missed areas (Herman 1971). Fertilizer increasednative plant growth on low fertility granite soils inIdaho but did not increase cover of seeded grasses(Cline and Brooks 1979). Other research and monitor-ing studies found no significant effect of fertilizer onplant cover or erosion (Esplin and Shackelford 1980,Tyrrel 1981). After fire, plant growth responds to aflush of readily available nitrogen compounds depos-ited on the soil surface with the ashes (Christensen1973, DeBano and others 1979a, DeBano and others1998). Research that showed increased growth byseeded grass with fertilization was conducted onfirelines, where the nutrient-rich ash layer had beenscraped from the soil (Klock and others 1975). It couldmake more sense to apply fertilizer late in the firstgrowing season or during the second year, after theinitial flush of available nutrients has been used byplants or leached away.

Retention of soil onsite for productivity mainte-nance is an important BAER objective, but almost noevidence indicates whether seeding meets this goal.Two years after a fire, higher available soil nitrogenand higher cation exchange capacity were found inseeded areas than in adjacent swaths that had beenmissed (Griffith 1989b, 1998). Soil retention and nu-trient uptake/release by the seeded grass were cred-ited for the improvement. No other reports addressedsoil fertility. Whether soil nutrient loss from the fireitself and from subsequent erosion are significant tolong-term ecosystem productivity will depend onwhether fire severity was within or far outside thenatural range of variability for a given ecosystem.Although some nutrients are inevitably lost in a fire,they will be made up in time by natural processes(DeBano and others 1998). Loftin and others (1998)pointed out that sometimes postfire conditions aremore “natural,” from a long-term ecosystem perspec-tive, than the prefire condition in many forests thathave been subject to decades of fire suppression. Moreresearch and monitoring are needed to evaluate theneed for and effects of seeding and other BAER treat-ments on soil productivity.

Monitoring reports and interviews noted that occa-sionally seeding is done mostly for “political” reasons,because the public and elected officials expect to seesomething done to restore a burned area “disaster”near their community (Anonymous 1987, Ruby and

Griffith 1994). Smaller fires that burned under condi-tions not far out of the range of natural variation mayhave been seeded unnecessarily for this reason. Betterpublic education on the natural role of fire in ecosys-tems and the inevitability of a postfire sediment pulsecould reduce the need for “political” seeding.

Other hillslope treatments have been used, but littlequantitative information has been published on theireffectiveness. Thus effectiveness ratings are oftenbased on visual assessment with no direct compari-sons with other treatments. Hillslope treatments suchas large contour trenching may increases infiltrationand trap sediment during summer thunderstorms,but they are expensive to install and require machin-ery, thus limiting the slopes that they can work, andthey have long-term impacts on hillslope appearanceand hydrologic function. More recently, hand trencheshave been used. Hand trenches are quicker to installand require less skilled crews. Straw wattles maydetain surface runoff, reduce velocities, store sedi-ment, and provide a seedbed for germination. Cattleexclusion with temporary fencing can be important forthe first 2 years postfire. Ripping/tilling was effectiveon roads, trails, and firebreaks with slopes less than35 percent. Slash spreading is effective if good groundcontact is maintained. Most of these treatments can-not be applied to large areas but may be appropriate incritical areas of high risk. Monitoring of effectivenessis needed to determine if they are cost effective as well.

Channel Treatments—We conclude that channeltreatments should only be used if downstream threatis great. Straw bale check dams are designed to reducesediment inputs into streams. Collins and Johnston(1995) indicated that about 45 percent of the strawbales check dams installed were functioning properlyafter the first 3 months, whereas Miles and others(1989) reported that 87 percent of the straw bale checkdams were functioning and Kidd and Rittenhouse(1997) reported that 99 percent were functioning.They often fill in the first few storms, so their effective-ness diminishes quickly and they can blow out duringhigh flows. Thus their usefulness is short-lived.

Log dams can trap sediment by decreasing velocitiesand allowing coarse sediment to drop out. Fites-Kaufman (1993) indicated only a 3 percent failurerate. However, if these structures fail, they usuallyaggravate erosion problems. Log and rock grade stabi-lizers emphasis stabilizing the channel rather thanstoring the sediment. They tend to work for low andmoderate flows, not high flows. No reports on channelstabilization effectiveness were found.

No estimates of erosion reduction were found bystream bank armoring. Channel clearing—removinglogs and other organic debris—was rated “good” 71percent of the time since it prevents logs from beingmobilized in debris flow or floods. Since the value of


in-stream woody debris to fish has been realized, thistreatment has declined in popularity whereas in-channel felling has increased in popularity. No esti-mates on rock cage dams effectiveness were found butit is known that they provide grade stability andreduce velocities to drop out coarse sediment. Debrisbasins are designed to store runoff and sediment andare often the last recourse to prevent downstreamflooding and sedimentation. They are often designedto trap 50 to 70 percent of the expected flows. Noestimates of sediment trapping efficiencies for debrisbasins were found.

Road Treatments—Road treatments are designedto move water to desired locations and prevent wash-out of roads. There is little quantitative researchevaluating and comparing road treatment effective-ness. A recent computer model, X-DRAIN, can providesediment estimates for various spacings of cross drains(Elliot and others 1998), and the computer model,WEPP-Road, provides sedimentation estimates forvarious road configurations and mitigation treatments(Elliot and others 1999). Thus, effectiveness of variousspacings of rolling dips, waterbars, cross drains, andculvert bypasses can be compared. By shortened flowpaths and route water at specified crossings, erosioncan be reduced. Upgrading culverts to larger sizesincreases their flow capacity, which reduces the risk ofblockage and exceeding capacity. Culvert armoringand adding risers allow sediment to settle out andprevent scouring. Trash racks prevent clogging ofculverts or other structures which keeps the culvertsopening as designed. Culvert removal, when appropri-ate, eliminates the threat of blockage. Storm patrolshows promise as a new cost effective method to keepculverts and drainage ditches clear, provide earlywarnings and close areas that could be threaten by astorm flows. Armoring ford crossings allows for low-cost access across stream channels, with the ability tohandle large flows. Ditch cleaning and armoring pro-vide for drainage of expected flows and reduce scour-ing. Outsloping prevents concentrated flow on roadsurfaces thus reducing erosion. Detail discussion ofroad related treatment effectiveness is beyond thescope of this report. The recent USDA Forest Service,San Dimas Technology and Development Program,Water/Road Interaction Technologies Series (Copstead1997) provides design standards, improvement tech-niques, and evaluations of some surface drainagetreatments for reducing sedimentation.

Conclusions____________________Relatively little monitoring of BAER treatments

has been conducted in the last three decades. Pub-lished literature focused on seeding issues, with little

information on any other treatments. Therefore, in-terview forms and monitoring reports were used todocument our current knowledge on treatment op-tions and their effectiveness. There were at least 321BAER project fires during last three decades thatcost the Forest Service around $110 million to reha-bilitate. Some level of monitoring occurred on about33 percent of these project fires. Our analysis of theliterature, Burned Area Report forms, interview com-ments, monitoring reports, and treatment effective-ness ratings leads us to the following conclusions:

• Existing effectiveness monitoring efforts and re-search are insufficient to accurately compare treat-ment effectiveness and ecosystem recovery.

• Rehabilitation should be done only if the risk tolife and property is high since the amount ofprotection provided is assumed to be small. Insome watersheds, it would be best not to do anytreatments. If treatments are necessary then it ismore effective to reduce erosion onsite (hillslopetreatment) rather than collected it downstream(channel treatment).

• Contour-felled logs show promise as a relativelyeffective treatment compared to other hillslopetreatments. This is considered to be true for areaswhere erosion rates are expected to be high be-cause they provide protection during the first-year postfire which has the highest erosion rates.In areas that do not have available trees, strawwattles may provide an alternative. However, theeffectiveness of contour-felled logs or straw wattleshas not been adequately documented in the scien-tific literature.

• Seeding has a low probability of reducing erosionthe first wet season after a fire. There is a need todo other treatments in critical areas. Seeding canprovide reasonable cover late in first season andin the second year. Most estimates of ground coveroccur at the end of the first growing season, thuscover information is not appropriate for compari-son for first year storm events.

• There is a need to better understand regenerationpotential of natural vegetation. Seeding treat-ment may not be needed as often as currentlythought.

• Because seeding is often not “successful,” it mayhave little impact on natural regeneration. Per-sistent perennials are least effective at providingfirst year cover and most likely to interfere withlater regeneration. Cereal grains (annuals) offerbetter first-year protection than perennials butgenerally do not interfere with later regenerationof natural vegetation. Little is known about theeffectiveness of native annual grasses.

• Evaluating postfire watershed conditions, treat-ment chance of success, cost-benefit ratios, and


risk assessments was difficult because variousmethods were used to estimate their values. Littleinformation and research is available on riskassessment and cost-risk ratios of various BAERtreatments.

• To reduce the threat of road failure, road treat-ments such as rolling dips, water bars, and reliefculverts properly spaced provide a reasonablemethod to move water past the road prism. Stormpatrol attempts to keep culverts clear and closeareas as needed. This approach shows promise asa cost effective technique to reduce road failuredue to culvert blockage.

• Straw bale checkdams, along with other channeltreatments, should be viewed as secondary miti-gation treatments. Sediment has already beentransported from the slopes and will eventually bereleased though the stream system as the balesdegrade, although the release is desynchronized.

Recommendations ______________Based on the findings from this study, we provide

the following recommendations to further our knowl-edge and understanding of the role of emergencyrehabilitation treatments:

• Streamline the Burned Area Report (FS-2500-8)form to address postfire watershed cost-benefit andrisk analysis in an easily understandable manner.Provide information to assist decisionmakers to beable to compare treatment alternatives and under-stand that the consequences are only going to hap-pen if we have storm events.

• Increase training on methods to calculate and usedesign storm intensity and frequency, probabilityof success, and erosion risk estimates. These canbe targeted to soil scientists and hydrologistsbecause they are involved with virtually everyBAER effort.

• Increase the number of quantitative studies todocument contour-felled logs effectiveness in re-ducing erosion. Additional research is needed todetermine whether contour-felling can reducerilling, increase infiltration, and decrease down-stream time to peakflow (slow water velocities).Hand trenching effectiveness is another treat-ment that has not been documented, but may beeffective and should also be evaluated.

• Increase monitoring efforts to determine if treat-ments are performing as planned and designed.Monitoring should include measuring effective-ness in reducing erosion, sedimentation, or down-stream flooding, but may also include changes ininfiltration, soil productivity, ecosystem recoveryand water quality parameters. Two levels ofmonitoring are proposed. Extensive effectiveness

monitoring can be accomplished at the forest levelwith little regional support, thus numerous sites/fires can be evaluated in different climate re-gimes. Intensive performance monitoring wouldneed regional and research support and could bedone on “demonstration” fires for each region(physiographic or Forest Service).

Effectiveness Monitoring: Silt fences placed at thebottom of hillslope plots are an economical method tocompare hillslope treatments by determining howmuch sediment is trapped by each silt fence. Plots canbe established to compare hillslope treatments such asseeding, contour-felled logs, hand trenches, etc. Siltfences have a very high trap efficiency (greater than90-95 percent), and are easily maintained and ser-viced. For maximum information gain, treated repli-cated plots should be compared for physically similaruntreated plots.

Performance Monitoring: To compare sedimenta-tion responses of various treatments, small catchmentsneed to be monitored for runoff and sediment. This isa costly and time-consuming technique but does pro-vide the best results and would need to be conductedin conjunction with research in order to prevent short-coming from past efforts. This method can be used tocompare hillslope or road or channel treatments.

• Support Research efforts to improve methodologiesto assess and predict long-term effects of wildfire onsoil and site productivity.

• Develop a knowledge-base of past and current BAERprojects that is easily accessible to others (i.e., Inter-net). This would include treatment design criteriaand specifications, contract implementation speci-fications, example Burned Area Report calcula-tions, and monitoring techniques.

References _____________________Agee, J. K. 1993. Fire ecology of Pacific northwest forests. Island

Press, Washington, DC. 493 p.Amaranthus, Michael P. 1989. Effect of grass seeding and fertiliz-

ing on surface erosion in two intensely burned sites in south-west Oregon. In: Berg, Neil H., tech. coord. Proceedings of thesymposium on fire and watershed management, October 26-28,1988, Sacramento, California. Gen. Tech. Rep. PSW-109. Berke-ley, CA: U.S. Department of Agriculture, Forest Service, PacificSouthwest Forest and Range Experiment Station: 148-149.

Amaranthus, Michael P. 1990. Siskiyou National Forest wildfireand recovery monitoring report. Unpublished report on file at:U.S. Department of Agriculture, Forest Service, Siskiyou Na-tional Forest, OR. 142 p.

Amaranthus, Michael P.; Trappe, James M.; Perry, David A. 1993.Soil moisture, native revegetation, and Pinus lambertianaseedling survival, growth, and mycorrhiza formation followingwildfire and grass seeding. Restoration Ecology. 1(3): 188-195.

Ambos, Norm. 1992. Summary—Dude fire photo points. Unpub-lished report on file at: U.S. Department of Agriculture, ForestService, Tonto National Forest, AZ. 6 p.

Anderson, William E.; Brooks, Lee E. 1975. Reducing erosionhazard on a burned forest in Oregon by seeding. Journal ofRange Management. 28(5): 394-398.


Anderson, H. W.; Coleman, G. B.; Zinke, P. J., 1959. Summer slidesand winter scour—wet-dry erosion in southern California moun-tains. Gen. Tech. Rep. PSW-36. Berkeley, CA: U.S. Departmentof Agriculture, Forest Service, Pacific Southwest Forest andRange Experiment Station. 12 p.

Anderson, H. W.; Hoover, M. D.; Reinhart, K. G. 1976. Forests andwater: effects of forest management on floods, sedimentation,and water supply. Gen. Tech. Rep. PSW-18. Berkeley, CA: U.S.Department of Agriculture, Forest Service, Pacific SouthwestForest and Range Experiment Station. 115 p.

Anonymous. 1987. Oak Grove fire. Unpublished report on file at:Dixie National Forest, UT. 2 p.

Arndt, Arthur M. 1979. Emergency revegetation of burned water-sheds in Los Angeles County. Riverside, CA: Chaparral Researchand Development Program, CHAPS Newsletter, March 1979: 1-3.

Bailey, Reed W.; Copeland, Otis L. 1961. Vegetation and engineer-ing structures in flood and erosion control. In: Proceedings, 13thCongress, September 1961, Vienna, Austria. International Unionof Forest Research Organization Paper. 11-1: 23 p.

Baron, F. J. 1962. Effects of different grasses on ponderosa pineseedling establishment. Res. Note PSW-199. Berkeley, CA: U.S.Department of Agriculture, Forest Service, Pacific SouthwestForest and Range Experiment Station. 8 p.

Barro, Susan C.; Conard, Susan G. 1987. Use of ryegrass seeding asan emergency revegetation measure in chaparral ecosystems.Gen. Tech. Rep. PSW-102. Berkeley, CA: U.S. Department ofAgriculture, Forest Service, Pacific Southwest Forest and RangeExperiment Station. 12 p.

Bautista, S.; Bellot, J.; Vallejo, V.R. 1996. Mulching treatment forpost-fire soil conservation in a semiarid ecosystem. Arid SoilResearch and Rehabilitation. 10: 235-242.

Benda, L. E.; Cundy. T. W. 1990. Predicting deposition of debrisflows in mountain channels. Canadian Geotechnical Journal. 27:409-417.

Best, D. W.; Kelsey, D. K.; Hagans, D. K.; Alpert, M. 1995. Role offluvial hillslope erosion and road construction in the sedimentbudget of Garret Creek, Humboldt County, California. In: Nolan,K. M.; Kelsey, H. M.; Marron, D. C., eds. Geomorphic Processesand Aquatic Habitat in the Redwood Creek Basin, NorthwesternCalifornia. Professional Paper 1454. Washington, DC: U.S. Geo-logical Survey: M1-M9.

Beyers, Jan L.; Conard, Susan G.; Wakeman, Carla D. 1994. Impactsof an introduced grass, seeded for erosion control, on postfirecommunity composition and species diversity in southern Cali-fornia chaparral. In: Proceedings of the 12th internationalconference on fire and forest meteorology, October 26-28, 1993,Jekyll Island, GA. Bethesda, MD: Society of American Forest-ers: 594-601.

Beyers, Jan L.; Wakeman, Carla D.; Wohlgemuth, Peter M.; Conard,Susan G. 1998a. Effects of postfire grass seeding on nativevegetation in southern California chaparral. In: Proceedings,nineteenth annual forest vegetation management conference:wildfire rehabilitation, January 20-22, 1998. Redding, CA: ForestVegetation Management Conference: 52-64.

Beyers, Jan L.; Wohlgemuth, Peter M.; Wakeman, Carla D.; Conard,Susan G. 1998b. Ryegrass seeding for postfire erosion control inchaparral—does it work? Fire Management Notes. 58: 30-34.

Biswell, H. H.; Schultz, A. M. 1965. Surface runoff and erosion asrelated to prescribed burning. Journal of Forestry. 55: 372-373.

Bitterroot National Forest. 1997. Forest plan monitoring andevaluation report, fiscal year 1997, Bitterroot National Forest,MT. 80 p.

Blanford, Robert H.; Gunter, Louis E. 1972. Emergency revegeta-tion—a review of project evaluations 1956-1970. Forest, Rangeand Watershed Management Section, California Division of For-estry, Sacramento, CA. 21 p.

Blankenbaker, Gene; Ryan, Tom; Graves, Walt. 1985. Aguangaburn soil erosion and vegetation recovery administrative study.Unpublished report on file at: U.S. Department of Agriculture,Forest Service, Cleveland National Forest, CA. 9 p.

Bosch, J. M.; Hewlett, J. D. 1982. A review of catchment experi-ments to determine the effect of vegetation changes on wateryield and evapotranspiration. Journal of Hydrology. 55: 3-23.

Boyd, Anne; McKay, Mary; Dailey, Scott. 1995. Field observationson the Cleveland Fire, January 10, 1995. Unpublished report on

file at: U.S. Department of Agriculture, Forest Service, EldoradoNational Forest, CA. 2 p.

Brooks, K. N.; Ffolliott, P. F.; Gregorsen, H. M.; DeBano, L. F. 1997.Hydrology and management of watersheds. Ames, IA: Iowa StateUniversity Press. 392 p.

Burroughs, Edward R., Jr., ; King, John G. 1989. Reduction in soilerosion of forest roads. Gen. Tech. Rep. INT-264. Ogden, UT: U.S.Department of Agriculture, Forest Service, Intermountain Re-search Station. 21 p.

Callahan, Karen; Baker, Blaze. 1997. Crystal burn analysis. Post-fire succession and seeding monitoring: year three. Unpublishedreport on file at: U.S. Department of Agriculture, Forst Service,Tahoe National Forest, CA. 23 p.

Campbell, G. S.; Jungbauer, J. D., Jr.; Bristow, K. L.; Hungerford,R. D. 1995. Soil temperature and water content beneath a surfacefire. Soil Science. 159(6): 363-374.

Campbell, R. E.; Baker, M. B., Jr.; Ffolliott, P. F. 1977. Wildfireeffects on a ponderosa pine ecosystem: an Arizona case study.Res. Pap. RM-191. Fort Collins, CO: U.S. Department of Agricul-ture, Forest Service, Rocky Mountain Forest and Range Experi-ment Station. 12 p.

Cannon, Susan. 1999. Debris-flow response of watersheds recentlyburned by wildfire. Boulder, CO: University of Colorado. 200 p.Dissertation.

Chiun-Ming, L. 1985. Impact of check dams on steep mountainchannels in northeastern Taiwan. In: El-Swaify, S.A.;Moldenhauer, W. C.: Lo, A., ed. Soil Erosion and Conservation:540-548.

Christ, J. H. 1934. Reseeding burned-over lands in northernIdaho. University of Idaho Agricultural Experiment StationBulletin 201. Moscow, ID: University of Idaho. 27 p.

Christensen, N. L. 1973. Fire and the nitrogen cycle in Californiachaparral. Science. 181: 66-68.

Cline, G. G.; Brooks, W. M. 1979. Effect of light seed and fertilizerapplication in steep landscapes with infertile soils after fire.Missoula, MT: U.S. Department of Agriculture, Forest Service,Northern Region Soil, Air, Water Notes 79-2. 6 p.

Collins, L. M.; Johnston, C. E. 1995. Effectiveness of straw baledams for erosion control in the Oakland Hills following the fire of1991. In: Keeley, J. E.; Scott, T., eds. Brushfires in Californiawildlands: ecology and resource management. Fairfield, WA:International Association of Wildland Fire: 171-183.

Conard, Susan G.; Beyers, Jan L.; Wohlgemuth, Peter M. 1995.Impacts of postfire grass seeding on chaparral ecosystems—whatdo we know and where do we go from here? In: Keeley, J. E.;Scott, T., eds. Brushfires in California wildlands: ecology andresource management. Fairfield, WA: International Associa-tion of Wildland Fire: 149-161.

Conard, S. G.; Regelbrugge, J. C.; Wills, R. D. 1991. Preliminaryeffects of ryegrass seeding on postfire establishment of naturalvegetation in two California ecosystems. In: Andrews, Patricia L.;Potts, Donald F., eds. Proceedings of the 11th conference on fireand forest meteorology, April 16-19, 1991, Missoula, Montana.Bethesda, MD: Society of American Foresters: 314-321.

Conrad, C. Eugene. 1979. Emergency postfire seeding using annualgrass. Riverside, CA: Chaparral Research and DevelopmentProgram, CHAPS Newsletter, March 1979: 5-8.

Cooper, W. S. 1922. The broad-sclerophyll vegetation of California:an ecological study of the chaparral and its related communities.Carnegie Institution of Washington Publication 319. 145 p.

Copstead, R. 1997. The water/road interaction technology series: anintroduction. No. 9777 1805-SDTDC. San Dimas, CA: U.S. De-partment of Agriculture, Forest Service, San Dimas TechnologyCenter. 4 p.

Copeland, Otis L. 1961. Watershed management and reservoir life.Journal of American Water Works Association. 53(5): 569-578.

Copeland, Otis L. 1968. Forest Service Research in erosion controlin the western United States. Presented at: Annual Meeting,American Society of Agricultural Engineers, June 18-21, 1968,Logan, UT. Paper 68-227. St. Joseph, MI: American Society ofAgricultural Engineers. 11 p.

Corbett, E. S.; Green, L. R. 1965. Emergency revegetation to rehabili-tate burned watersheds in southern California. Res. Pap. PSW-22.Berkeley, CA: U.S. Department of Agriculture, Forest Service,Pacific Southwest Forest and Range Experiment Station. 14 p.


Costales, E. F., Jr.; Costales, A. B. 1984. Determination and evalu-ation of some emergency measures for the quick rehabilitation ofnewly burned watershed areas in the pine forest. Sylvtrop Phil-ippine Forestry Research. 9: 33-53.

Croft, A. R.; Marston, Richard B. 1950. Summer rainfall character-istics in northern Utah. Transactions, American GeophysicalUnion. 31(1): 83-95.

Davis, E. A. 1984. Conversion of Arizona chaparral increases wateryield and nitrate loss. Water Resources Research. 20: 1643-1649.

DeBano, L. F.; Conrad, C. E. 1976. Nutrients lost in debris andrunoff water from a burned chaparral watershed. In: Proceedingsof the third interagency sedimentation conference, Denver, CO,March 1976. Water Resources Council. 3: 13-27.

DeBano, L. F.; Eberlein, G. E.; Dunn, P. H. 1979a. Effects ofburning on chaparral soils. I. Soil nitrogen. Soil Science Societyof America Journal. 43: 504-509.

DeBano, L. F.; Ffolliott, P. F.; Baker, M. B., Jr. 1996. Fire severityeffects on water resources. In: Ffolliott, P. F.; DeBano, L. F.;Baker, M. B., Jr.; Gottfried, G. J.; Solis-Garza, G.; Edminster,C. B.;Neary, D. G.; Allen, L. S.; Hamre, R. H., tech. coords. Effects offire on Madrean province ecosystems—A symposium proceed-ings. Gen. Tech. Rep. RM-289. Fort Collins, CO: U.S. Departmentof Agriculture, Forest Service, Rocky Mountain Forest and RangeExperiment Station: 77-84.

DeBano, L. F.; Neary, D. G.; Ffolliott, P. F. 1998. Fires’s effects onecosystems. New York: John Wiley & Sons. 333 p.

DeBano, L. F.; Rice, R. M.; Conrad, C. E. 1979b. Soil heating inchaparral fires: effects on soil properties, plant nutrients, erosionand runoff. Res. Pap. PSW-145. Berkeley, CA: U.S. Departmentof Agriculture, Forest Service, Pacific Southwest Forest andRange Experiment Station. 21 p.

DeBlye, Norbert V. 1970a. Do contour trenches reduce wet-mantleflood peaks? Res. Note INT-108. Ogden, UT: U.S. Department ofAgriculture, Forest Service, Intermountain Forest and RangeExperiment Station. 8 p.

DeByle, Norbert V. 1970b. Infiltration in contour trenches in theSierra Nevada. Res. Note INT-115. Ogden, UT: U.S. Departmentof Agriculture, Forest Service, Intermountain Forest and RangeExperiment Station. 5 p.

DeByle, N. V.; Packer, P. E. 1972. Plant nutrient and soil losses inoverland flow from burned forest clearcuts. In: Csallany, S. C.;McLaughlin, T. G.; Striffler, W. D., eds. Watersheds in transition,proceedings, American Water Resources Association and Colo-rado State University, Fort Collins, CO. Urbana, IL: AmericanWater Resource Association: 296-307.

DeGraff, J. V. 1982. Final evaluation of felled trees as a sedimentretaining measure, Rock Creek Fire, Kings RD, Sierra NationalForest, Fresno, CA. U.S. Department of Agriculture, Forest Ser-vice Special Report, Sierra National Forest. 15 p.

Department of Forester and Fire Warden. 1985. Emergency reveg-etation in Los Angeles County 1919-1984. Unpublished report onfile at: County of Los Angeles, Fire Department, P.O. Box 3009,Terminal Annex, Los Angeles, CA 90051. 93 p.

Doty, Robert D. 1970. Influence of contour trenching on snowaccumulation. Journal of Soil and Water Conservation. 25(3):102-104.

Doty, Robert D. 1971. Contour trenching effects on streamflow froma Utah watershed. Res. Pap. INT-95. Ogden, UT: U.S. Depart-ment of Agriculture, Forest Service, Intermountain Forest andRange Experiment Station. 19 p.

Doty, Robert D. 1972. Soil water distribution on a contour-trenchedarea. Res. Note INT-163. Ogden, UT: U.S. Department of Agri-culture, Forest Service, Intermountain Forest and Range Ex-periment Station. 6 p.

Ducummon, Sheryl L. 1987. Las Pilitas fire rehabilitation assess-ment project. Unpublished report on file at: U.S. Department ofAgriculture, Forest Service, Los Padres National Forest, CA. 13 p.plus maps and photos.

Dyer, David A. 1995. Determining viability of soil seed banks afterwildfires. Tech. Note TN-Plant Materials-36. Lockeford, CA: U.S.Department of Agriculture, Natural Resources ConservationService. 8 p.

Dyrness, C. T. 1974. Grass-legume mixtures for erosion controlalong forest roads. Handout prepared for “Soil erosion preven-tion and sedimentation control short course,” Olympia, Wash-ington, February 6, 1974. 16 p.

Dyrness, C. T. 1976. Effect of wildfire on soil wettability in the HighCascades of Oregon. Res. Pap. PNW-202. Portland, OR: U.S.Department of Agriculture, Forest Service, Pacific NorthwestForest and Range Experiment Station. 18 p.

Edwards, L.; Burney, J.; DeHaan, R. 1995. Researching the effectsof mulching on cool-period soil erosion in Prince Edward Island,Canada. Journal of Soil and Water Conservation. 50: 184-187.

Elliot, Katherine J.; White, Alan S. 1987. Competitive effects ofvarious grasses and forbs on ponderosa pine seedlings. ForestScience. 33(2): 356-366.

Elliot, W. J.; Graves, S. M.; Hall, D. E.; Moll, J. E. 1998. The X-DRAIN cross drain spacing and sediment yields. No. 9877 1801-SDTDC. San Dimas, CA: U.S. Department of Agriculture, ForestService, San Dimas Technology Center. 24 p.

Elliot, W. J.; Hall, D. E.; Scheele, D. L. 1999. WEPP-Road, WEPPinterface for predicting forest road runoff, erosion and sedimentdelivery [Online]. Available: http://forest.moscowfsl.wsu.edu/fswep/docs/wepproaddoc.html.

Esplin, DeLoy H.; Shackleford, James R. 1978. Cachuma burnreseeding evaluation: coated vs. uncoated seed, first growingseason. Unpublished report on file at: U.S. Department of Agri-culture, Forest Service, Los Padres National Forest, CA. 24 p.

Esplin, DeLoy H.; Shackleford, James R. 1980. Cachuma burnrevisited. Unpublished report on file at: U.S. Department ofAgriculture, Forest Service, Los Padres National Forest, CA. 5 p.

Evanko, Anthony B. 1955. Species adaptability and seeding seasonfor increased forage production in northeastern Washington.Res. Pap. INT-37. Ogden, UT: U.S. Department of Agriculture,Forest Service, Intermountain Forest and Range ExperimentStation. 24 p.

Farmer, Eugene E.; Fletcher, Joel E. 1972. Some intra-storm char-acteristics of high-intensity rainfall burst. In: Distribution ofPrecipitation in Mountainous areas, Geilo Symposium, NorwayJuly 31-August 5. Geneva, Switzerland: World MeteorologicalOrganization. 2: 525-531.

Farmer, E. E.; Van Haveren, B. P. 1971. Soil erosion by overlandflow and raindrop splash on three mountain soils. Gen. Tech. Rep.INT-100. Ogden, UT: U.S. Department of Agriculture, ForestService, Intermountain Research Station. 14 p.

Faust, Robert. 1998. Lesson plan: Fork fire soil loss validationmonitoring. Unit VIII, Long term recovery and monitoring. In:Burned area emergency rehabilitation (BAER) techniques, 1998.San Francisco: U.S. Department of Agriculture, Forest Service,Pacific Southwest Region. (course notebook).

Federal Reserve Bank. 1999. Consumer Price Index [online].Available: http://woodrow.mpls.frb.fed.us/economy/calc/cpihome/[June 16, 1999].

Ferrell, William R. 1959. Report on debris reduction studies formountain watersheds. Los Angeles, CA: Los Angeles CountyFlood Control District, Dams and Conservation Branch. 164 p.

Fites-Kaufman, Russell. 1993. Survey of check dams in the Cleve-land Fire Area. Unpublished report on file at: U.S. Department ofAgriculture, Forest Service, Eldorado National Forest, CA. 3 p.

Flanagan, S. A.; Furniss, M. J. 1997. Field indicators of inlet con-trolled road stream crossing capacity. No. 9777 1808-SDTDC.San Dimas, CA: U.S. Department of Agriculture, Forest Service,San Dimas Technology Center. 5 p.

Forsling, C. L. 1931. Artificial reseeding on western mountainrange lands. USDA Circ. No. 178. Washington, DC: U.S. De-partment of Agriculture. 48 p.

Friedrich, C. Allen. 1947. Seeding grass by airplane on westernMontana’s burned-over timberlands. Res. Note No. 52. Missoula,MT: U.S. Department of Agriculture, Forest Service, NorthernRocky Mountain Forest and Range Experiment Station. 5 p.

Furniss, Michael J.; Ledwith, Tyler, S.; Love, Michael A.; McFadin,Bryan C. Flanagan, Sam A. 1998. Response of road-streamcrossings to large flood events in Washington, Oregon, andNorthern California. San Dimas, CA: U.S. Department of Agri-culture, Forest Service, Technology and Development Program,9877 1806-San Dimas Technology Development Center. 14 p.

Gautier, Clayton R. 1983. Sedimentation in burned chaparralwatersheds: is emergency revegetation justified? Water ResourcesBulletin. 19(5): 793-802.

Geier-Hayes, Kathleen. 1997. The impact of post-fire seeded grasseson native vegetative communities in central Idaho. In: Greenlee,Jason M., ed. Proceedings: First conference on fire effects on rare


and endangered species and habitats. Fairfield, WA: Interna-tional Association of Wildland Fire: 19-26.

Gibbons, Boyd. 1995. Remarks on postfire seeding. In: Keeley, JonE.; Scott, Tom, eds. Brushfires in California wildlands: ecologyand resource management. Fairfield, WA: International Associa-tion of Wildland Fire: 211.

Gleason, C. H. 1947. Guide for mustard sowing in burned water-sheds of southern California. San Francisco, CA: U.S. Depart-ment of Agriculture, Forest Service, California Region. 46 p.

Glendening, G. E.; Pase, C. P.; Ingebo, P. 1961. Preliminaryhydrologic effects of wildfire in chaparral. In: Johnson, R., ed.Proceedings of the 5th annual Arizona watershed symposium,September 21, 1961, Phoenix, Arizona. Phoenix, AZ: ArizonaWater Resources Committee and Wathershed ManagementDivision, State Land Department: 12-15.

Goldman, S. J.; Jackson, K.; Bursztynsky, T. A. 1986. Erosion andsediment control handbook. San Francisco, CA: McGraw-Hill.360 p.

Green, L. R. 1981. Burning by prescription in chaparral. Gen. Tech.Rep. PSW-51. Berkeley, CA: U.S. Department of Agriculture,Forest Service, Pacific Southwest Forest and Range ExperimentStation. 36 p.

Green, Pat. 1990. 1989 soil monitoring report, Green Creek Pointfire, Clearwater Ranger District. Unpublished report on file at:U.S. Department of Agriculture, Forest Service, Nez Perce Na-tional Forest, ID. 4 p.

Griffin, James R. 1982. Pine seedlings, native ground cover, andLolium multiflorum on the Marble-Cone burn, Santa LuciaRange, California. Madroño. 29(3): 177-188.

Griffith, Robert W. 1989a. Memo, silt fence monitoring. Unpub-lished report on file at: U.S. Department of Agriculture, ForestService, Stanislaus National Forest, CA. 4 p.

Griffith, Robert W. 1989b. Memo, soil productivity and burned areagrass seeding. Unpublished report on file at: U.S. Department ofAgriculture, Forest Service, Stanislaus National Forest, CA. 2 p.

Griffith, Robert W. 1991. Memo, Arch Rock burn area erosionmonitoring. Unpublished report on file at: U.S. Department ofAgriculture, Forest Service, Stanislaus National Forest, CA. 2 p.

Griffith, Robert W. 1993. Memo, Arch Rock burn area grass moni-toring. Unpublished report on file at: U.S. Department of Agricul-ture, Forest Service, Stanislaus National Forest, CA. 2 p.

Griffith, Robert W. 1998. Burned area emergency rehabilitation.In: Proceedings, nineteenth annual forest vegetation manage-ment conference: wildfire rehabilitation, January 20-22, 1998.Redding, CA: Forest Vegetation Management Conference: 4-7.

Hanes, R. O.; Callahan, K. 1995. Summary of first monitoring.Unpublished report on file at: U.S. Department of Agriculture,Forest Service, Tahoe National Forest, CA. 24 p.

Hanes, Richard; Callahan, Karen. 1996. Crystal burn analysis.Post-fire succession and seeding monitoring: year two. Unpub-lished report on file at: U.S. Department of Agriculture, ForestService, Tahoe National Forest,CA. 23 p.

Hartford, R. A.; Frandsen, W. H. 1992. When it’s hot, it’s hot—ormaybe it’s not (surface flaming may not portend extensive soilheating). International Journal of Wildland Fire. 2: 139-144.

Heede, Burchard H. 1960. A study of early gully-control structuresin the Colorado Front Range. Station Paper 55. Fort Collins, CO:U.S. Department of Agriculture, Forest Service, Rocky MountainForest and Range Experiment Station. 42 p.

Heede, Burchard H. 1970. Design, construction and cost of rockcheck dams. Res. Pap. RM-20. Fort Collins, CO: U.S. Depart-ment of Agriculture, Forest Service, Rocky Mountain Forest andRange Experiment Station. 24 p.

Heede, Burchard H. 1976. Gully development and control: thestatus of our knowledge. Res. Pap. RM-169. Fort Collins, CO: U.S.Department of Agriculture, Forest Service, Rocky MountainForest and Range Experiment Station. 42 p.

Heede, Burchard H. 1981. Rehabilitation of disturbed watershedthrough vegetation treatment and physical structures. In: Pro-ceeding, Interior West Watershed Management Symposium,April 1980, Spokane, WA. Pullman, WA: Washington State Uni-versity Cooperative Extension: 257-268.

Helvey, J. D. 1975. Watershed behavior after forest fire in Washing-ton. In: Proceedings of the irrigation and drainage divisionspecialty conference, Fort Collins, Colorado, April 22-24, 1973.

Washington, DC: United States Government Printing Office:403-422.

Helvey, J. D. 1980. Effects of a north-central Washington wildfireon runoff and sediment production. Water Resources Bulletin.16: 627-634.

Hendricks, B. A.; Johnson, J. M. 1944. Effects of fire on steepmountain slopes in central Arizona. Journal of Forestry. 42:568-571.

Herman, Loren D. 1971. Review of Forks fire, Okanogan N.F., andfire on Wenatchee N.F. Unpublished report on file at: U.S.Department of Agriculture, Forest Service, Okanogan NationalForest, OR. 18 p.

Hibbert, A. R. 1971. Increases in streamflow after convertingchaparral to grass. Water Resources Bulletin. 7: 71-80.

Hibbert, A. R.; Davis, E. A.; Knipe, O. D. 1982. Water yieldchanges resulting from treatment of Arizona chaparral. Gen.Tech. Rep. PSW-58, Berkeley, CA: U.S. Department of Agricul-ture, Forest Service: 382-389.

Hungerford, R. D. 1996. Soils- fire in ecosystem management notes:unit II-I. Marana, AZ: U.S. Department of Agriculture, ForestService, National Advanced Resource Technology Center.

Isle, David. 1988. 2,3445 acre 8/26/88 Powerhouse burn: Isle’sfollow-up observations notes. Unpublished report on file at: U.S.Department of Agriculture, Forest Service, Mendocino NationalForest, CA. 6 p.

Isle, David W. 1998. Erosion control seeding and mulching treat-ments. Unit VIII, “Back to the future”: long term recovery andmonitoring. In: Burned area emergency rehabilitation (BAER)techniques, 1998 (course notebook). San Francisco: U.S. Depart-ment of Agriculture, Forest Service, Pacific Southwest Region.

Janicki, Alex. 1989. Emergency revegetation of the StanislausComplex fire. Unpublished report on file at: U.S. Department ofAgriculture, Forest Service, Stanislaus National Forest, CA. 17 p.

Janicki, Alex. 1992. Memo, Cotton burn area erosion monitoring.Unpublished report on file at: U.S. Department of Agriculture,Forest Service, Stanislaus National Forest, CA. 3 p.

Kay, B. L. 1983. Straw as an erosion control mulch. AgronomyProgress Report No. 140. Davis, CA: University of CaliforniaAgricultural Experiment Station. 11 p.

Keeler-Wolf, Todd. 1995. Post-fire emergency seeding and conserva-tion in southern California shrublands. In: Keeley, J. E.; Scott, T.,eds. Brushfires in California wildlands: ecology and management.Fairfield, WA: International Association of Wildland Fire: 127-139.

Keeley, Jon E. 1991. Seed germination and life history syndromesin the California chaparral. The Botanical Review. 57: 81-116.

Keeley, J. E.; Keeley, S. C. 1988. Chaparral. In: Barbour, M. G.;Billings, W. D., eds. North American terrestrial vegetation.Cambridge: Cambridge University Press: 165-207.

Keeley, Jon E.; Carrington, Mary; Trnka, Sally. 1995. Overview ofmanagement issues raised by the 1993 wildfires in southernCalifornia. In: Keeley, Jon E.; Scott, Tom, eds. Brushfires inCalifornia wildlands: ecology and resource management.Fairfield, WA: International Association of Wildland Fire: 83-89.

Keeley, S. C.; Keeley, J. E.; Hutchinson, S. M.; Johnson, A. W. 1981.Postfire succession of herbaceous flora in southern Californiachaparral. Ecology. 62(6): 1608-1621.

Kidd, Jerry L.; Rittenhouse, Dave. 1997. Evaluation of stormeffects on Eight Street Fire Rehabilitation. Unpublished reporton file at: U.S. Department of the Interior, Bureau of LandManagement, Boise, ID. 8 p.

Klock, Glen O.; Tiedemann, Arthur R.; Lopushinsky, William. 1975.Seeding recommendations for disturbed mountain slopes innorth central Washington. Res. Note PNW-244. Portland, OR:U.S. Department of Agriculture, Forest Service, Pacific North-west Forest and Range Experiment Station. 8 p.

Kozlowski, T. T.; Ahlgren, C. E., eds. 1974. Fire and ecosystems.New York: Academic Press. 542 p.

Krammes, J. S. 1960. Erosion from mountain side slopes after firein southern California. Res. Note PSW-171. Berkeley, CA: U.S.Department of Agriculture, Forest Service, Pacific SouthwestForest and Range Experiment Station. 8 p.

Krammes, J. S.; Hill, L. W. 1963. “First aid” for burned watersheds.Res. Note PSW-29. Berkeley, CA: U.S. Department of Agricul-ture, Forest Service, Pacific Southwest Forest and Range Experi-ment Station. 7 p.


Lachowski, Henry; Hardwick, Paul; Griffith, Robert; Parsons, Annette;Warbington, Ralph. 1997. Faster, better data for burned water-sheds needing emergency rehab. Journal of Forestry. 95(6): 4-8.

Liewer, Marvin. 1990. The big fire—one year later. Unpublishedreport on file at: U.S. Department of Agriculture, Forest Service,Black Hills National Forest, SD. 2 p.

Loftin, Sam; Fletcher, Reggie; Luehring, Peggy. 1998. Disturbed arearehabilitation review report. Unpublished report on file at: U.S.Department of Agriculture, Forest Service, Southwestern Region,Albuquerque, NM. 17 p.

Maloney, P. Cavan; Thornton, John L. 1995. Implementation andeffectiveness monitoring of best management practices and soiland water protection measures within the Foothills Fire salvagelogging area 1992-1995. Unpublished report on file at: U.S.Department of Agriculture, Forest Service, Boise National For-est, ID. 51 p.

McCammon, Bruce P. 1980. Bridge Creek fire rehabilitation statusreport. Unpublished report on file at: U.S. Department of Agri-culture, Forest Service, Deschutes National Forest, OR. 12 p.

McCammon, B. P.; Hughes, D. 1980. Fire rehab in the BendMunicipal Watershed. In: Proceedings of the 1980 watershedmanagement symposium, July 21-23, 1980, Boise, ID. New York:American Society of Civil Engineers: 252-259.

McClure, Norman R. 1956. Grass and legume seedings on burned-over forest lands in northern Idaho and adjacent Washington.Moscow, ID: University of Idaho. 125 p. Thesis.

McDonald, Philip M. 1986. Grasses in young conifer plantations—hindrance and help. Northwest Science. 60(4): 271-278.

McDonald, P. M.; Oliver, W. W. 1984. Woody shrubs retard growth ofponderosa pine seedlings and saplings. In: Coombes, D., ed. Pro-ceedings, fifth annual forest vegetation management conference.Redding, CA; Forest Vegetation Management Conference: 65-89.

McNabb, D. H.; Swanson, F. J. 1990. Effects of fire on soil erosion.In: Walstad, J. D.; Radosevich, S. R.; Sandberg, D. V., eds. Naturaland prescribed fire in Pacific northwest forests. Corvallis, OR:Oregon State University Press: 159-176.

Meeuwig, R. O. 1971. Soil stability on high-elevation rangeland inthe Intermountain area. Res. Pap. INT-94. Ogden, UT: U.S.Department of Agriculture, Forest Service, Intermountain Re-search Station. 10 p.

Megahan, W. F. 1986. Recent studies on erosion and its control onforest lands in the United States. In: Richard, F., ed. Range basinsediment delivery: Proceedings; 1986 August; Albuquerque, NM.IAHS Pub. 159, Wallingford, Oxon, United Kingdom: 178-189.

Megahan, Walter F.; Molitor, Delbert C. 1975. Erosion effects ofwildfire and logging in Idaho. In: Watershed management sym-posium: 1975 August, Logan, UT. New York: American Society ofCivil Engineers Irrigation and Drainage Division: 423-444.

Miles, Scott R.; Haskins, Donald M.; Ranken, Darrel W. 1989.Emergency burn rehabilitation: cost, risk, and effectiveness. In:Berg, Neil H., tech. coord. Proceedings of the symposium on fireand watershed management, October 26-28, 1988, Sacramento,California. Gen. Tech. Rep. PSW-109. Berkeley, CA: U.S. Depart-ment of Agriculture, Forest Service, Pacific Southwest Forestand Range Experiment Station: 97-102.

Nadkarni, Nalini M.; Odion, Dennis C. 1986. Effects of seeding anexotic grass Lolium multiflorum on native seedling regenera-tion following fire in a chaparral community. In: DeVries,Johannes J., ed. Proceeding of the chaparral ecosystems re-search conference, May 16-17, 1985, Santa Barbara, California.Report No. 62, California Water Resources Center, University ofCalifornia, Davis: 115-121.

Neary, D. G.; Hornbeck, J. W. 1994. Chapter 4: Impacts of harvest-ing and associated practices on off-site environmental quality. In:Dyck, W. J.; Cole, D. W.; Comerford, N. B., eds. Impacts of forestharvesting on long-term site productivity. London: Chapman &Hall: 81-118.

Neary, D. G.; Klopatek, C. C.; DeBano, F. F.; Ffolliott, P. F. 1999.Fire effects on belowground sustainability: a review and synthe-sis. Forest Ecology and Management. 122: 51-71.

Niehoff, Jerry. 1995. Mary-mix monitoring trip. Unpublished re-port on file at: U.S. Department of Agriculture, Forest Service,Idaho Panhandle National Forest, ID. 3 p.

Noble, Edward L. 1965. Sediment reduction through watershedrehabilitation. In: Proceedings of the federal inter-agency

sedimentation conference 1963. Washington, DC: U.S. Depart-ment of Agriculture, Misc. Publ. 970: 114-123.

Noble, Edward L.; Lundeen, Lloyd. 1971. Analysis of rehabilitationtreatment alternatives for sediment control. In: Symposium onforest land uses and stream environment, October 19-21, 1971,Corvallis, OR. Corvallis, OR: Oregon State University, School ofForestry and Department of Fisheries and Wildlife, ContinuingEducation Publications: 86-96.

Norris, Logan A. 1990. An overview and synthesis of knowledgeconcerning natural and prescribed fire in the Pacific Northwestforest. In: Walstad, John D.; Radosevich, Steven R.; Sandberg,David V., eds. Natural and prescribed fire in Pacific Northwestforests. Corvallis, OR: Oregon State University Press: 7-22.

Orr, Howard K. 1970. Runoff and erosion control by seeded andnative vegetation on a forest burn: Black Hills, South Dakota.Res. Pap. RM-60. Fort Collins, CO: U.S. Department of Agricul-ture, Forest Service, Rocky Mountain Forest and Range Experi-ment Station. 12 p.

Packer, P. E.; Christensen, G. F. 1977. Guides for controlling sedi-ment from secondary logging roads. Ogden, UT: U.S. Depart-ment of Agriculture, Forest Service, Intermountain Forest andRange Experiment Station, and Missoula, MT: U.S. Departmentof Agriculture, Forest Service, Northern Region. 42 p.

Parsons, Annette. 1994. Field notes from burn rehab and soilerosion monitoring day trip to observe treatments at 2 yearspost-fire. Unpublished report on file at: U.S. Department ofAgriculture, Forest Service, Eldorado National Forest, CA. 5 p.

Parsons, Annette; McDonald, Brandi. 1993. Preliminary results,Eldorado National Forest Cleveland fire emergency watershedtreatment monitoring. Unpublished report on file at: U.S. De-partment of Agriculture, Forest Service, Eldorado National For-est, CA. 5 p.

Pase, P. C.; Ingebo, P. A. 1965. Burned chaparral to grass: early effectson water and sediment yields from two granitic soil watersheds inArizona. In: Cooper, C. C., Jr., ed. Proceedings of the 9th annualArizona watershed symposium, September, 1965, Tempe, AZ.Phoenix, AZ: Arizona Water Resources Committee and WatershedManagement Division, State Land Department: 8-11.

Pase, C. P.; Lindenmuth, Jr., A. W. 1971. Effects of prescribed fireon vegetation and sediment in oak-mountain mahogany chapar-ral. Journal of Forestry. 69: 800-805.

Patric, J. H. 1976. Soil erosion in eastern forests. Journal ofForestry. 74: 671-676.

Pearson, G. A. 1942. Herbaceous vegetation: a factor in naturalregeneration of ponderosa pine in the southwest. EcologicalMonographs. 12: 316-338.

Powers, R. F.; Alban, D. H.; Miller, R. E.; Tiarks, A. E.; Wells, C. G.;Avers, P. E.; Cline, R. G.; Loftus, N. S., Jr.; Fitzgerald, R. O. 1990.Sustaining productivity of North American forests: Problemsand prospects. In: Gessel, S. P.; Lacate, D. S.;Weetman, G. F.;Powers, R. F., eds. Sustained Productivity of Forest Soils, Pro-ceedings of the 7th North American Forest Soils Conference, 24-28 July 1988, Vancouver, BC: University of British ColumbiaFaculty of Forestry: 49-79.

Pyne, S. J.; Andrews, P. L.; Laven, R. D. 1996. Introduction towildland fire. New York: John Wiley & Sons. 769 p.

Radek, K. J. 1996. Soil erosion following wildfires on the OkanoganNational Forest—initial monitoring results. In: InternationalErosion Control Association symposium, February 27-March 1,1996, Seattle, WA. Steamboat, CO: International Erosion Con-trol Association. 7 p.

Ratliff, Raymond D.; McDonald, Philip M. 1987. Postfire grass andlegume seeding: what to seed and potential impacts on refor-estation. In: Proceedings, ninth annual forest vegetation man-agement conference, November 3-5, 1987. Redding, CA: ForestVegetation Management Conference: 111-123.

Reid, L. M. 1993. Research and cumulative watershed effects. Gen.Tech. Rep. PSW-GTR-141. Berkeley, CA: U.S. Department ofAgriculture, Forest Service, Pacific Southwest Research Station.118 p.

Rice, Raymond M. 1974. The hydrology of chaparral watersheds. In:Rosenthal, ed. Symposium on living with the chaparral proceed-ing. San Francisco, CA: Sierra Club: 27-34.

Rice, R. M.; Crouse, R. P.; Corbett, E. S. 1965. Emergency measuresto control erosion after a fire on the San Dimas Experimental


Forest. In: Proceedings of the federal inter-agency sedimenta-tion conference 1963. Washington, DC: U.S. Department ofAgriculture, Misc. Publ. 970: 123-130.

Rich, L. R. 1962. Erosion and sediment movement following awildfire in a ponderosa pine forest of central Arizona. Res. Note 76.U.S. Department of Agriculture, Forest Service, Rocky MountainForest and Range Experiment Station. 12 p.

Rietveld, W. J. 1975. Phytotoxic grass residues reduce germinationand initial root growth of ponderosa pine. Res. Pap. RM-153. FortCollins, CO: U.S. Department of Agriculture, Forest Service,Rocky Mountain Forest and Range Experiment Station. 15 p.

Rinne, J. N. 1996. Short-term effects of wildfire on fishes andaquatic macroinvertebrates in the southwestern United States.North American Journal of Fisheries Management. 16: 653-658.

Robichaud, P. R.; Brown, R. E. 1999. What happened after thesmoke cleared: onsite erosion rates after a wildfire in easternOregon. In: Olsen, D. S.; Potyondy, J. P., eds. Proceeding, Wild-land Hydrology Conference, June 1999, Bozeman, MT. Hernon,VA:American Water Resource Association: 419-426.

Roby, Kenneth B. 1989. Watershed response and recovery from theWill Fire: ten years of observation. In: Berg, Neil H., tech. coord.Proceedings of the symposium on fire and watershed manage-ment, October 26-28, 1988, Sacramento, California. Gen. Tech.Rep. PSW-109. Berkeley, CA: U.S. Department of Agriculture,Forest Service, Pacific Southwest Forest and Range ExperimentStation: 131-136.

Rosgen, D. L. 1996. Applied river morphology. Pagosa Springs, CO:Wildland Hydrology. 364 p.

Ruby, Earl C. 1989. Rationale for seeding grass on the StanislausComplex burn. In: Berg, Neil H., tech. coord. Proceedings of thesymposium on fire and watershed management, October 26-28,1988, Sacramento, California. Gen. Tech. Rep. PSW-109. Berke-ley, CA: U.S. Department of Agriculture, Forest Service, PacificSouthwest Forest and Range Experiment Station: 125-130.

Ruby, Earl. 1995. Tyee BAER report: findings of the post-BAERevaluation team, 8/21-23/95. Unpublished report on file at: U.S.Department of Agriculture, Forest Service, Wenatchee NationalForest, WA. 6 p.

Ruby, Earl. 1997. Observations on the Dome fire emergency rehabili-tation seeding. Unpublished report on file at: U.S. Department ofAgriculture, Forest Service. Tonto National Forest, AZ. 3 p.

Ruby, Earl C.; Griffith, Rob. 1994. First-winter damage survey:Marre fire, burned area emergency rehabilitation, May-June1994. Unpublished report on file at: Los Padres National Forest,CA. 19 p. plus data sheets and photos.

Ryan, K. C.; Noste, N. V. 1983. Evaluating prescribed fires. In: Lotan,J. E.; Kilgore, B. M.; Fischer, W. C.; Mutch, R. W., tech. coords.Proceedings, symposium and workshop on wilderness fire. Gen.Tech. Rep. INT-182. Ogden, UT: U.S. Department of Agriculture,Forest Service, Intermountain Research Station: 230-238.

Sampson, Arthur W. 1944. Plant succession on burned chaparrallands in northern California. California Agricultural ExperimentStation Bulletin. 685. 144 p.

Sartz, R. S. 1953. Soil erosion on a fire-denuded forest area in theDouglas-fir region. Journal of Soil and Water Conservation:279-281.

Schultz, A. M.; Launchbaugh, J. L.; Biswell, H. H. 1955. Relation-ship between grass density and brush seedling survival. Ecology.36(2): 226-238.

Schultz, S.; Lincoln; Cauhorn, J.; Montagne, C. 1986. Quantifica-tion of erosion from a fire and rainfall event in the Big Belt Rangeof the Northern Rocky Mountains. Unpublished report on file at:U.S. Department of Agriculture, Forest Service, Helena NationalForest, MT. 21 p.

Schumm, S. A.; Harvey, M. D. 1982. Natural erosion in the USA. In:Schmidt, B. L., ed. Determinants of soil loss tolerance. Madison,WI: American Society of Agronomy, Special Publication. 45: 23-39.

Souders, C. 1997. Soil loss pin method form, Gila National Forest:HB fire. (data sheets) Unpublished report on file at: U.S. Depart-ment of Agriculture, Forest Service, Gila National Forest, NM. 2 p.

Stewart, William Lemley. 1973. Emergency rehabilitation of water-sheds denuded by wildfire in the ponderosa pine and Douglas-firzones of north-central Washington. Pullman, WA: WashingtonState University. 59 p. Thesis.

Story, Mark T. 1994. Water quality monitoring, Mill Creek timbersales and Thompson Creek fire. Unpublished report on file at:

U.S. Department of Agriculture, Forest Service, Gallatin Na-tional Forest, MT. 10 p.

Story, Mark T.; Kracht, Richard. 1989. Memo: Emergency water-shed stabilization, BMP implementation review Storm Creekfire. Unpublished report on file at: U.S. Department of Agricul-ture, Forest Service, Gallatin National Forest, MT. 16 p.

Sullivan, J.; Omi, P. N.; Dyer, A. A.; Gonzales-Caben, Armando. 1987.Evaluating the economics efficiency of wildfire rehabilitation treat-ments. Western Journal of Applied Forestry. 2(2): 58-61.

Swanson, F. J. 1981. Fire and geomorphic processes. In: Mooney, H.;Bonnicksen, T. M.; Christensen, N. L.; Lotan, J. E.; Reiners, W. A.,tech. coords. Proceedings of the conference; Fire regimes andecosystem properties, December 11-15, 1979, Honolulu, HI. Gen.Tech. Rep. WO-26. Washington, DC: U.S. Department of Agricul-ture, Forest Service: 410-420.

Sweeney, James R. 1956. Responses of vegetation to fire: a study ofthe herbaceous vegetation following chaparral fires. Universityof California Publications in Botany. 28: 143-249.

Taskey, R. D.; Curtis, C. L.; Stone, J. 1989. Wildfire, ryegrassseeding, and watershed rehabilitation. In: Berg, Neil H., tech.coord. Proceedings of the symposium on fire and watershedmanagement, October 26-28, 1988, Sacramento, California. Gen.Tech. Rep. PSW-109. Berkeley, CA: U.S. Department of Agricul-ture, Forest Service, Pacific Southwest Forest and Range Experi-ment Station: 149-161.

Taylor, Carl; Dalen, Ray; Ettner, Bob; Bates, Wes; Little, Harry;Allen, Larry; Knipe, Duane. 1979. Burned area rehabilitationactivity review. Unpublished report on file at: U.S. Departmentof Agriculture, Forest Service, Southwestern Region, Albuquer-que, NM. 68 p.

Tiedemann, Arthur R.; Klock, Glen O. 1973. First-year vegetationafter fire, reseeding, and fertilization on the Entiat ExperimentalForest. Res. Note PNW-195. Portland, OR: U.S. Department ofAgriculture, Forest Service, Pacific Northwest Forest and RangeExperiment Station. 23 p.

Tiedemann, Arthur R.; Klock, Glen O. 1976. Development of vegeta-tion after fire, seeding, and fertilization on the Entiat Experimen-tal Forest. In: Proceedings, annual Tall Timbers fire ecologyconference, number 15: Pacific northwest. Tallahassee, FL: TallTimbers Research Station: 171-191.

Tiedemann, A. R.; Helvey, J. D.; Anderson, T. D. 1978. Streamchemistry and watershed nutrient economy following wildfireand fertilization in eastern Washington. Journal of Environ-mental Quality. 7(4): 580-588.

Tyrrel, Robert R. 1981. Memo, Panorama burn rehabilitation.Unpublished report on file at: U.S. Department of Agriculture,Forest Service, San Bernardino National Forest, CA. 16 p.

U.S. Department of Agriculture, Forest Service. 1995. Burned AreaEmergency Rehabilitation handbook FSH 2509.13.

Van de Water, Roberta. 1998. Post-fire riparian zone management:the Salmon River experience. In: Proceedings, nineteenth annualforest vegetation management conference: wildfire rehabilita-tion, January 20-22, 1998. Redding, CA: Forest Vegetation Man-agement Conference: 25-40.

Van Zuuk, Kathy. 1997. Memo, Crystal burn monitoring. Unpublishedreport on file at: U.S. Department of Agriculture, Forest Service,Tahoe National Forest, CA. 2 p.

Wells, W. G., II. 1981. Some effects of brushfires on erosion pro-cesses in coastal southern California. In: Erosion and SedimentTransport in Pacific Rim Steeplands. Christchurch, New Zealand:IASH Publication No. 132: 305-342.

Went, F. W.; Juhren, G.; Juhren, M. C. 1952. Fire and biotic factorsaffecting germination. Ecology. 33(3): 351-364.

Wohlgemuth, Peter M.; Beyers, Jan L.; Wakeman, Carla D.; Conard,Susan G. 1998. Effects of fire and grass seeding on soil erosion insouthern California chaparral. In: Proceedings, nineteenth an-nual forest vegetation management conference: wildfire rehabili-tation, January 20-22, 1998. Redding, CA: Forest VegetationManagement Conference: 41-51.

Wright, H. A.; Churchill, F. M.; Stevens, W. C. 1976. Effect ofprescribed burning on sediment, water yield, and water qualityfrom dozed juniper lands in central Texas. Journal of RangeManagement. 29: 294-298.

Wright, H. A.; Churchill, F. M.; Stevens, W. C. 1982. Soil loss andrunoff on seeded vs. non-seeded watersheds following prescribedburning. Journal of Range Management. 35: 382-385.


Appendix A—Example Data and Interview Forms


Appendix B—TreatmentEffectiveness Summaries_________

In the course of conducting BAER team memberinterviews and reviewing monitoring reports, we ac-quired considerable information on the factors thatmake the various BAER treatments effective or not, aswell as useful tips for implementation. Most of theinformation was not amenable to tabulation or otherquantitative expression, so it was entered into ourdatabase in “comments” fields. This information hasbeen summarized below, along with the effectivenessratings developed from the interview forms.

The effectiveness and implementation informationin these descriptions comes strictly from the com-ments and monitoring reports collected by us in thisproject. They are not intended to be comprehensiveanalyses of each treatment. Fully describing effectiveinstallation of treatments is beyond the scope of thisreport. The following comments should be used tosupplement other sources of information on the vari-ous treatments.

Hillslope Treatments

Hillslope treatments are implemented to keep soilplace and comprise the greatest effort in most BAERprojects. Consequently, we obtained the most infor-mation on these treatments from our interviews andmonitoring reports.

Aerial Seeding

Purpose: Aerial seeding, usually grasses but occasion-ally also legumes, is carried out to increase vegetativecover on a burn site during the first few years after afire. It is typically done where erosion hazard is highand native plant seed bank is believed to have beendestroyed or severely reduced by the fire. Seed isapplied by fixed-wing aircraft or helicopter.

Relative Effectiveness: Excellent-24% Good-28% Fair-28% Poor-20% (Replies = 83)Interviewees were almost evenly divided on the effec-tiveness rating of aerial seeding, with a slight majorityregarding it as either “good” or “fair” (table 16). Re-spondents in Regions 1, 4, and 6 were more likely torate seeding “excellent” or “good” than respondents inRegions 3 and 5.

Effectiveness depends on timeliness of seed applica-tion, choice of seed, pilot skill, protection from grazing,and luck in having gentle rains to stimulate seedgermination before wind or heavy rains blow or washsoil and seed away. Proper timing of seed applicationdepends on location. In some areas it is best to dropseed directly into dry ash, before any rain falls, to takeadvantage of the fluffy seedbed condition, while inothers seed is best applied after the first snow so that

it will germinate in the spring. Both conditions alsoreduce loss to rodents. Choice of seed determines howeasily it can be applied – some grass species with longawns tend to clog in seeder buckets, and light seedsdrift more than heavy ones – and how well it will grow,how long it will persist, and what impact it will haveon natural regeneration. In general, legumes have notbeen found to be particularly effective at producingcover (there are exceptions). A skillful pilot will applythe seed evenly, rather than in strips with unseededareas in between them, providing better ground coveronce the seed germinates.

A few respondents also mentioned that straw mulch,needle cast, slope barriers such as straw wattles orcontour-felled logs, or ripping the soil enhanced growthof seeded grasses. Maximum cover of seeded species isnot attained until summer. Many respondents re-ported that seeding was not particularly effective atproducing protection from the first year’s storms (es-pecially in the Southwest for fires that occur justbefore the monsoon season with its high intensityrains) but may provide effective cover during thesecond and subsequent years. Several respondentssuggested that waiting to seed onto snow for springgrowth would be the most effective course of action inthe Southwest (Region 3), because they usually endedup having to do a second seeding anyway after thesummer monsoon washed the first application away.Several respondents noted disappointing results fromseeding with relatively expensive native species orRegreen (commercially available sterile wheatgrasshybrid) and would not use them again. On the otherhand, cereal grains were generally reported to per-form well the first growing season. Cereal grains thatdo not germinate in quantity the second year providesoil cover with the mulch from the dead first yeargrowth. Both cattle and elk grazing were reported toreduce the effective cover of seeded grasses. Seededgrass cover tends to be higher on low angle slopes (lessthan 40 percent) than steep ones.

Implementation and Environmental Factors: Manyrespondents reported difficulties in contracting forseed and aircraft operators which, especially after fallfires, resulted in seed being applied too late for opti-mum conditions. Ground sampling, with sticky papersor by visual inspection, should be done to monitor seedapplication rate and evenness. Fixed wing aircraftmay be less expensive per application but can be lessaccurate at directing seed than helicopters.

The use of native seed is a major issue on manyForests. Native grass seed can be hard to acquire inlarge quantities or in a timely manner compared tocereal grains or pasture grasses; it is also generallymore expensive. Native seed should come from anearby source area to preserve local genetic integrity.Cereal grains will germinate and grow the second year


if the ground surface is disturbed by salvage logging orgrazing. Many monitoring studies have found lowercover of native plants in areas with high seeded grasscover, even where seeding increased total cover. Some-times this resulted in lower total cover after the seededgrass decreased in abundance. On the other hand,seeded grass may also inhibit growth of noxious weedsthat invade sites after fire, a beneficial outcome. Rhi-zomatous (sod-forming) grasses make reforestationmore difficult if they achieve significant cover. It isimportant to know the composition of prefire vegeta-tion when proposing to seed – if the vegetation in-cluded many annuals or lots of perennial grass orsedge, there will usually be considerable cover estab-lished naturally after a fire.

Other factors: Many respondents noted that grassseeding was sometimes done primarily for “political”reasons, especially at the wildland-urban interface.

Ground Seeding

Purpose: Ground seeding is done in localized areas ofhigh burn intensity where reestablishing plant coverquickly is essential, such as riparian areas, abovelakes and reservoirs, or highly productive forest land.Annual or perennial grasses, usually non-native pas-ture grasses or cereal grains, and non-native legumi-nous forbs, are typically used. Ground seeding assuresmore even seed application than aerial seeding andsometimes includes treatments to cover the seed,which enhances germination. Seed is applied from all-terrain vehicles or by hand.

Relative Effectiveness: Excellent-9% Good-82% Fair-9% Poor-0% (Replies = 11)

Ground seeding was judged “good” in effectiveness bymost interviewees. As with aerial seeding, the post-fire weather pattern frequently determines the effec-tiveness of cover production by seeded grass. Highwinds may blow seed off site. First rains can wash ashand seed from the hillslope, or they may be gentleenough to stimulate germination. Use of a rangelanddrill, raking, or mulch to cover seed increases success.One forest used cattle to trample seed into the groundand break up a hydrophobic layer. Non-native species,especially perennial grasses, grow well, sometimes toowell, and provide persistent cover. Cereal grains dis-appear in a few years.

Implementation and Environmental Factors: Timing ofseed application is essential to success; optimum timingdepends on local weather pattern. The seed mix must beadapted to the soil type. Awned or very light seedsspread more easily if rice hulls (or similar material) areincluded in the mix. Grass growth is best on lower angleslopes (less likely to wash away). Protection from cattlegrazing the first year is considered by some to be the

biggest factor in success; protection for 2 or 3 years isgood. Elk may have a negative effect on seeded grassesas well.

Seeding Plus Fertilizer

Purpose: Seeding plus fertilization is done to increasetotal vegetation cover quickly on a burned slope. Occa-sionally fertilizer alone is applied to enhance naturalregeneration.

Relative effectiveness: Excellent-25% Good-0% Fair-50% Poor-25% (Replies = 4)

Fertilization received mixed reviews among the fourrespondents. As with seeding, timing of applicationand post-fire weather pattern are important to suc-cess. Fertilization is mainly done in the Northwest andammonium sulfate is most commonly used. One re-spondent reported that greener strips were apparentin the seeded area where the fertilizer had beenapplied. Pelleted seed, containing a small amount offertilizer, may be easier to apply than uncoated seed.

Implementation and Environmental Factors: Alongriparian areas slow release fertilizer has been used tominimize leaching into waterways. There is evidencethat fertilizer may inhibit or depress mycorrhizaeformation.

Contour-Felled Logs (Log Erosion Barriers, Log Ter-races, Terracettes)

Purpose: Contour-felled logs reduce water velocity,break up concentrated flows, and induce hydraulicroughness to burned watersheds. Sediment storage isa secondary objective. The potential volume of sedi-ment stored is highly dependent on slope, the size andlength of the felled trees, and the degree to which thefelled trees are adequately staked and placed intoground contact.


The effectiveness of contour-felling covered the spec-trum from “excellent” to “poor,” although more ratingswere “excellent” or “good” (66 percent) than “fair” or“poor” (34 percent) (table 16). Some personnel re-ported 100 percent of logs functioning, while othersreported 0 percent functioning. Site conditions, instal-lation quality, climate, and the quality of materialsare major factors in determining relative effective-ness. In some instances contour-felled log barriershave filled with sediment following the first stormevent after installation, while others have taken 1 to2 years to fill.

Implementation and Environmental Factors: Goodplanning, proper implementation, and knowledge ofenvironmental factors are crucial to the success of


contour-felling. This BAER treatment is expensive,technically demanding, and dangerous work, so crewskill and experience and good supervision are impor-tant. Attention to felling and delimbing safety rules isparamount. Logs must be placed on the contour, put incontact with the ground, and properly anchored. Ifthese three items are ignored, failure is assured. Thistreatment needs to be implemented in a very methodi-cal and meticulous manner. Increased installationspeed or area covered will not make up in effectivenessthat can be lost by poor installation. Ground contactcan be assured by adequate delimbing beneath eachlog, leaving branches downhill, trenching, and back-filling. In some instances machinery has been used tomake ground contact trenches, but the usual methodis to excavate with hand labor due to equipment andslope limitations. Trenching to seat contour-felled logshas an additional benefit in that it can help to break uphydrophobic layers in the soil. Anchoring can be donewith wooden or re-bar stakes where slopes are steeper,but should be of sufficient frequency and depth toprevent movement of the logs.

Shallow, rocky soils that are very uneven are prob-lematic for anchoring, so care must be taken to ensurethat logs are adequately secured to the slope. Overlyrocky and steep slopes should be avoided, becausebenefits gained from contour-felling treatment can beeasily offset by extra implementation time requiredand limited stabilization of small amounts of soil.Gentler slopes and finer textured soils (except clayeysoils) lead to better installation and greater sedimenttrapping efficiency. Slopes less than 40 percent arerecommended for successful contour-felling. Slopesgreater than 75 percent present significant installa-tion safety hazards and should be avoided. In someinstances, only the lower portions of slopes near ephem-eral or perennial channels have been treated. In highlyerosive soils derived from parent material such asgranitics or glacial till, so much sediment can bemobilized that it might overwhelm small contour-felled logs.

Availability of adequate numbers of straight treesalso affects this treatment. Specifications require logsfrom burned trees 15 to 20 ft (4.5 to 6 m) in length withdiameters of 4 to 12 in (100 to 305 mm). Placing treestems 10 ft (3 m) apart on slopes over 50 percent, 15 ft(4.5 m) apart for slopes of 30 to 50 percent, and 20 ft (6 m)apart for slopes less than 30 percent would require2000 to 4000 linear ft ac–1 (1500 to 3000 linear m ha–1)of tree bole on some sites. A shortage of dead timber orlarge numbers of small diameter trees could placelimitations on the contour-felled treatment area.Crooked stems, such as oak, are often readily avail-able, but they are not useable or cost-effective forcontour-felling treatment. Cutting trees for contour-felled log barriers reduces the number of snags for

birds to use. However, it often increases vegetationcover when plants become established in fine sedi-ments trapped on the uphill sides of the felled logs.Contour-felled logs should be placed in a randompattern to ensure a more “natural” appearance andavoid patterns which might aggravate runoff.

Mulch

Purpose: Mulch is used to cover soil, reducing rainimpact and soil erosion. It is often used in conjunctionwith grass seeding to provide ground cover in criticalareas. Mulch protects the soil and improves moistureretention underneath it, benefitting seeded grasses inhot areas but not always in cool ones.


Mulch was judged “excellent” in effectiveness by mostinterviewees, although many also noted that it is quiteexpensive and labor-intensive (table 16). It is mosteffective on gentle slopes and in areas where highwinds are not likely to occur. Wind either blows themulch offsite or piles it so deeply that seed germina-tion is inhibited. On very steep slopes, rain can washsome of the mulch material downslope. Punching itinto the soil, use of a tackifier, or felling small treesacross the mulch may increase onsite retention. Mulchis frequently applied to improve germination of seededgrasses. In the past, seed germination from grain orhay mulch was regarded as a bonus, adding cover tothe site. Use of straw from pasture introduces exoticgrass seed. Forests are now likely to seek “weed-free”mulch such as rice straw. Mulch is judged most valu-able for high value areas, such as above or below roads,above streams, or below ridge tops.

Implementation and Environmental Factors: Mulchcan be applied most easily where road access isavailable because the mulch must be trucked in,although for critical remote areas it can be applied byhelicopter or fixed wing aircrafts. Hand application islabor-intensive and can result in back or eye injuriesto workers. Using a blower to apply the mulch re-quires considerable operator skill to get uniformdistribution of the material. Effectiveness dependson even application and consistent thickness. Ricestraw is not expected to contain seeds of weeds thatcould survive on a chaparral or forested site (too dry);however, weeds do germinate sometimes and couldresult in introducing new exotics to wildland areas.Other certified “weed-free” straws sometimes con-tain noxious weeds. There is concern that thick mulchinhibits native shrub or herb germination. Shrubseedlings have been observed to be more abundant atthe edge of mulch piles, where the material was lessthan 1 in (25 mm) deep. Because of the weed and


germination concerns, mulch should not be used inareas with sensitive or rare plants. Mulch can beapplied in 100 to 200 ft (30 to 60 m) wide strips on longslopes, saving labor costs and also reducing the po-tential impact of the mulch on native plant diversity.

Slash Spreading

Purpose: Slash spreading covers the ground with or-ganic material, interrupting rain impact and trappingsoil. It is a common practice after timber sales, but canalso be used on burned slopes where dead vegetationis present. Slash is more frequently used on firebreaksand dozer firelines.

Relative Effectiveness: Good–50% Fair–50% (Replies = 2)

Interviewees that used this treatment rated the effec-tiveness “good” and “fair.” It is more effective on gentleslopes than steep ones. In accessible areas, the mate-rial can disappear as people collect it for firewood. Onerespondent was disappointed that not much sedimentwas trapped by spread slash.

Implementation and Environmental Factors: Slashneeds to be cut so it makes good contact with theground. It can be used in a moderately burned area,where there is more material to spread, or below anintensely burned slope or area of water repellent soil.There is concern that slash will attract or harborinsects, and it could act as fuel for a reburn.

Temporary Fencing

Purpose: Temporary fencing is used to keep grazinglivestock and/or vehicles off of burned areas and ripar-ian zones during the recovery period. Resproutingonsite vegetation and seeded species attract grazinganimals and are initially very sensitive to distur-bance. Fencing can speed up the recovery process byremoving post-fire disturbance from grazers and ve-hicles.


Temporary fencing was evaluated as “good” or “fair”by the limited number of interviewees that rated it(table 16). They noted that the effectiveness is depen-dent on the extent to which grazers are excluded fromthe burned areas. In some areas, elk grazing is asproblematic as cattle grazing, and the use of the morecostly high fences that exclude elk needs to be consid-ered. The presence and intensity of native ungulategrazing will definitely affect the success of fencing.Elimination of grazing for 2 years was judged to bevery important for achieving hillslope stability. Oneperson noted that temporary fencing could have ex-cellent effectiveness when done before winter, butthe chance of fencing being completed before winter

is often low due to the extensive time requirements offence construction.

Implementation and Environmental Factors: SomeBAER personnel recommend cattle exclusion if morethan 50 percent of an allotment is burned. If a decisionis made to employ temporary fences, installation needsto be timely and proper. Fence construction is slowrelative to other BAER treatments so it is importantthat fence installation is not delayed. It is important tokeep cattle out of burned areas before and during fenceconstruction. Incursions by cattle can slow fence con-struction. Consideration should be given to installa-tion of big game/elk exclosures where these animalshave a significant impact on burned area recovery.The location of temporary fences should be coordi-nated with existing allotment fences.

Other Factors: Some personnel liked using BAERfunds with Forest funds to achieve long-term fencinggoals. Others apparently have had problems gettingfencing put in with BAER funds. Electric fence is anoption for excluding cattle. This option needs to beconsidered more in the future. It may be more cost-effective, easier, and quicker to install just after aerialseeding than other types of fences. Fencing is also agood tool for excluding off-road vehicles from sensitiverecently burned areas.

Straw Wattles

Purpose: Straw wattles are permeable barriers used todetain surface runoff long enough to reduce flow veloc-ity. Their main purpose is to break up slope length.They have also been used in small drainages or on sideslopes for detaining small amounts of fine suspendedsediment.


The effectiveness rating of straw wattles ranges from“excellent” to “fair” depending on the circumstances inwhich they were used and the quality of the installa-tion. Comments within one Region on straw wattleeffectiveness ranged from being an “excellent” treat-ment at a reasonable cost and still functioning after 2years, to that of exhibiting pronounced undercuttingimmediately on the downhill side. Visual monitoringhas noted that straw wattles usually remain in placeand often fill with soil material on the uphill side.Where that happens, good seed germination occurs.Straw wattles have been placed onto specific sites andrandomly located on slopes. Some monitoring observa-tions have noted that there does not appear to be adifference in overall vegetative recovery between con-tour-felled log areas and straw wattle treatment ar-eas. Overall effectiveness can be affected by break-down of the wattles and release of built-up sedimentonto the rest of the slope or into drainages.


Implementation and Environmental Factors: Correctinstallation of straw wattles is crucial to their effec-tiveness. They are labor intensive because they needto have good ground contact and anchoring. Wattlescan be anchored to the ground by trenching andbackfilling or staking. An effective anchoring tech-nique is to use “U” shaped 1/8 in (3 mm) re-bar. Re-barcan hold wattles to the ground without trenching andis less likely to break than wood stakes in shallow soils.Straw wattles can work well on slopes greater than 40percent but they are difficult to carry and hard toinstall on steep terrain. Spotting the wattles withhelicopters can solve some of this problem.

Other Factors: The cost of straw wattle installation isabout one half that of contour-felled logs.

Tilling/Ripping

Purpose: Tilling and ripping are mechanical soil treat-ments aimed at improving infiltration rates in ma-chine-compacted or water repellent soils. Both treat-ments may increase the amount of macropore space insoils by physical breakup of dense or water repellentsoils, and thus increase the amount of rainfall thatinfiltrates into the soil.


Tilling and ripping was judged to be an “excellent”treatment for roads, firebreaks, and trails but lesseffective on hillslopes (table 16). These techniquesmay add roughness to the soil and promote infiltra-tion. They may be successful for site-specific circum-stances like compacted or water repellent areas, butnot economically feasible on large areas or safe to do onslopes greater than 30 to 45 percent. Size of theequipment and crawler tractor operator skill are alsoimportant effectiveness factors. Up- and down-hilltilling/ripping needs to be avoided because it candiminish the effectiveness of the treatment in reduc-ing soil erosion by promoting rilling in the furrows.According to some personnel, this type of treatmentwas the most effective when done in combination withbroadcast seeding. Others indicated that tilling/rip-ping can be successful accomplished at a high produc-tion rate on non-timbered areas without seeding.

Implementation and Environmental Factors: Shallowsoils, rock outcrops, steep slopes, incised drainages,fine-textured soils, and high tree density create sig-nificant problems for tilling and ripping. These treat-ments work best where there is a good soil depth, thesoils are coarse textured, slopes are less than 30percent, and woody vegetation density is low. Thistype of treatment has a high logistics support require-ment (fuel, transport carriers, access, and drainagecrossing).

Other Factors: Since tilling and ripping are ground-disturbing activities, cultural clearances are required.Obtaining proper cultural clearances may significantlyslow accomplishment of tilling/ripping projects.

Contour Trenching and Terraces

Purpose: Contour trenches are used to break up theslope surface, to slow runoff and allow infiltration, andto trap sediment. Rills are stopped by the trenches.Trenches or terraces are often used in conjunctionwith seeding. They can be constructed with machinery(deeper trenches) or by hand (generally shallow). Widthand depth vary with design storm, spacing, soil type,and slope.


Two of the three interviewees who rated trenchingconsidered its effectiveness “excellent;” the otherthought it “good” (table 16). Trenches trap sedimentand interrupt water flow, slowing runoff velocity.They work best on coarse granitic soils. When in-stalled with heavy equipment, trenches may resultin considerable soil disturbance that can createproblems.

Implementation and Environmental Factors: Trenchesmust be built along the slope contour to work properly;using baffles or soil mounds to divide the trenchreduces the danger of excessive flow if they are notquite level. Digging trenches requires fairly deep soil,and slopes of less than 70 percent are best. Trenchesare hard to construct in heavy, clay soils and are notrecommended for areas prone to landslides. Handcrews can install trenches much faster than log ero-sion barriers (a similarly effective hillslope treat-ment), and crew skill is not quite as important toeffective installation. Trenches have high visual im-pact when used in open areas (and thus may be subjectto controversy), but tend to disappear with time asthey are filled with sediment and covered by vegeta-tion. On the other hand, more extreme (wide, deep)trenches installed several decades ago are still visibleon the landscape in some areas.

Geotextiles, Geowebbing

Purpose: Matting is used to cover ground and controlerosion in high risk areas where other methods willnot work, such as extremely steep slopes, above roadsor structures, or along stream banks. It is usually usedin conjunction with seeding. Geotextiles come in dif-ferent grades with ultraviolet inhibitors that deter-mine how long they will last in the field.

Relative Effectiveness: No interviewees rated thistreatment.


When geotextiles mats are applied over seed andmulch, they are very effective in stopping erosion.Because the cost is very high, they are used only whereimmediate ground cover is needed; large areas cannotbe covered by this method. Geotextiles are particu-larly effective for steep upper slopes where othermaterials (seed, mulch alone) will blow off. Materialmust be anchored securely to remain effective, espe-cially along streambanks.

Implementation and Environmental Factors: An expe-rienced crew is needed to ensure that good contact isestablished between the fabric material and the ground,and that the fabric is securely anchored. Fabric mat-ting is difficult to apply on rocky ground. Plasticnetting on some geotextiles material can trap smallrodents and birds. Jute netting does not provide com-plete ground cover but it has not been reported to trapanimals. The complete cover provided by somegeotextiles can reduce native plant establishment.

Silt Fences

Purpose: Silt fences are installed to trap sediment inswales, small ephemeral drainages, or along hillslopeswhere other methods cannot be used. They providetemporary sediment storage. Silt fences are also in-stalled to monitor sediment movement as part ofeffectiveness monitoring.

Relative Effectiveness: Excellent–38% Good–62% Fair–0% Poor–0% (Replies = 8)

Silt fences were considered “good” or “excellent” byinterviewees (table 16). Most respondents felt theyworked well in ephemeral channels, but not all. Thesize of the watershed above the fence may be impor-tant, and silt fences cannot handle debris flows orheavy sediment loads. They work better on gentlerslopes, such as swales. Silt fences can be installed onrocky slopes where log erosion barriers would notachieve good ground contact. Sealing the bottom of thefence to the ground well is critical to effectiveness andseems to work best if a trench is dug behind the fenceto trap sediment. Silt fences also effectively catchsmall rocks and ravel on slopes above buildings.

Implementation and Environmental Factors: As notedabove, silt fences must be anchored and sealed to theground to be effective. Sandbags can be used as an-chors. Burying the bottom of the fence in a trench isalso useful. Rockiness of the soil affects how well thetoe of the fence can be buried. When used in ephemeralchannels, silt fences must be cleaned out or they canfail and release the stored sediment all at once. Theyare useful for monitoring sediment movement, andcan last several years before failing.

Sand, Soil or Gravel Bags

Purpose: Sand, soil or gravel bags are used in smallchannels or on hillslopes to trap sediment and inter-rupt water flow.

Relative effectiveness: No interviewees rated thistreatment.

Comments indicate that bags are useful in ephemeralchannels or on slopes, where they are placed in stag-gered rows like contour felling in areas where thereare no trees. Rows of bags break water flow andpromote infiltration. They store sediment temporarily,then break down and release it.

Implementation and Environmental Factors: Whenbags are used in channels, installation sites must beselected by an experienced person. They are not appro-priate for use in V-shaped channels. Installation of soilbags is labor-intensive, but they can be a relativelycheap treatment if volunteer labor is used. The bagsare easy for volunteers to fill and install.

Channel Treatments

Channel treatments are implemented to modifysediment and water movement in ephemeral or small-order channels, and to prevent flooding and debristorrents that may affect downstream values at risk.

Straw Bale Check dams

Purpose: Straw bale check dams are used to prevent orreduce sediment inputs into perennial streams duringthe first winter or rainy season following a wildfire.Straw bales function by decreasing water velocity anddetaining sediment-laden surface runoff long enoughfor coarser sediments to deposit behind check dams.The decreased water velocity also reduces downcuttingin ephemeral channels.


Straw bale check dams were judged to cover the rangefrom “good” to “poor” effectiveness. They often fill inthe first few storms, so their effectiveness can dimin-ish rapidly. However, channel gradients can be easilystabilized, and sediment is stored and released at aslower or diminished rate. They appear to work well infront of culverts, and in semi-arid environments re-quire little maintenance. Structural survival rates of90 percent have been reported after 1 year with 75 to100 percent sediment storage, and 95 percent survivalafter rainfall of 2.4 in hr–1 (60 mm hr–1) for a 10-minduration. However, a common negative comment wasthat straw bale check dams tend to blow out in largestorms. Failure can occur if the dams are poorly


installed or put in locations where they can not containrunoff. Straw bale check dams are considered by manyBAER project coordinators to be effective emergencyrehabilitation treatments. Straw bale check damsappear to work better than contoured felled logs. SomeForests use straw bales below culverts to disperse flowand trap sediments. They appear to the most success-ful in channels small enough to require only threebales, but in narrow, steep drainages two-bale widestructures do not function as well.

Others do not recommend use of straw bales becausethey fill to capacity after small storms. They can bewashed out later even when anchored with “U” shaped1/8 in (3 mm) re-bar are useful only in the upperreaches of watersheds (1st or 2nd order drainages)that are often difficult to access, and can be easilyundercut if energy dissipators are not installed. One ofthe comments on straw bale check dams was there isalways a risk of failure in large events. These damscannot be designed for large storms, and will failduring significant runoff events.

Implementation and Environmental Factors: A largenumber of comments were made about importantimplementation and environmental factors that affectthe success of straw bale check dams. Regardingimplementation, a key factor is having a skilled imple-mentation leader and trained, experienced crews.Straw bale check dams are costly and labor intensive.With such a high investment, the dams must be well-designed, properly placed, and well built.

Generally speaking, straw bales work best in drierregions, on small drainage areas that have low gradi-ents (less than 30 percent), and in channels that arenot incised. The bales need to be placed so that theycontact the channel bottom, are curved up to andkeyed into banks, and are adequately staked or wiredto stay in place. Inter-bale spaces need to be filed sothat channelized flow does not occur. “U” shaped re-barseems to work well in stabilizing bales but don’tguarantee that the bales will remain in place. Geotextilefabric works well as an energy dissipator and shouldbe placed starting on the uphill side running over thebales in the center of the channel and downstream ina splash pad. Chicken wire and staking should be usedto keep the geotextile in place. Rock, wood, or otherstraw bales can also be used as energy dissipators butmust be large enough or well-anchored to preventmovement during runoff. Straw bale check dams seemto work better and survive longer than silt fences,especially when reinforced with wire on the upstreamside.

Other Factors: Because straw bales will break downover time and fail in high flows, maintenance duringthe first year is very important. Straw bales are notreadily available early in the year. After August they

are very available. Rice straw bales should be consid-ered because they usually do not contain noxiousweeds, and weeds associated with rice crops do not dowell on dry hillslopes and ephemeral channels. Strawbale check dams can be destroyed by grazing animalssuch as cattle and elk. Bears also have a peculiartendency to indulge in ripping straw bale check damsapart.

Log Grade Stabilizers

Purpose: The purpose of log grade stabilizers is muchthe same as log dams, except that the emphasis is onstabilizing the channel gradient rather than trappingsediment.


Interviewees rated log grade stabilizers about equallyacross the spectrum from “excellent” to “poor.” Like logdams, these structures are expensive and time-con-suming. In situations where log grade stabilizers wererated “excellent,” 70 to 80 percent of the structureswere still functional after 1 year. “Poor” ratings usu-ally resulted where the log grade stabilizers did notmake a difference or they were lost to high stormflows.

Implementation and Environmental Factors: Log gradestabilizers have many of the same design, implemen-tation, and environmental factors considerations thatlog dams do. Proper design and crew experience arecritical in making these structures last and functioneffectively. Numerous small log grade stabilizers arepreferable to a few larger ones. In some locations,there might not be adequate, straight, woody materialleft after a fire to build log grade stabilizers with onsiteresources.

Rock Grade Stabilizers

Purpose: The purpose of rock grade stabilizers is thesame as log grade stabilizers, except that they aremade of rock. The emphasis is on stabilizing thechannel gradient rather than trapping sedimentalthough some sediment will be trapped by thesestructures.


Only a few interviewees commented on rock gradestabilizers. They rated this technique as “good” to“fair.” There were not many comments about thistechnique.

Implementation and Environmental Factors: Manycomments on implementation and environmental fac-tors pertaining to log grade stabilizers, apply to rockgrade stabilizers. Proper design, adequate planning,and experienced crews often make the difference


between “good” and “fair” effectiveness. Like log gradestabilizers, this technique is expensive and time con-suming. A key implementation factor is the availabil-ity of rock for the grade stabilizers. A couple of impor-tant implementation factors that affect effectivenessare: (1) the use of rocks that are large enough to resisttransport during runoff events, and (2) placement oforganic debris or sediment screening on the upstreamside of the grade stabilizer.

Channel Debris Clearing

Purpose: Channel clearing is the removal or size re-duction of logs and other organic debris or the removalof sediment deposits to prevent them from being mo-bilized in debris flows or flood events or alteringstream geomorphology and hydrology. This treatmenthas been done to prevent creation of channel debrisdams which might result in flash floods, or aggravateflood heights or peakflows. Organic debris can lead toculvert failure by blocking inlets culverts, or reducechannel flow capacity. Excessive sediments in streamchannels can compromise in-channel storage capacityand the function of debris basins.


Channel debris clearing was rated as “good” in effec-tiveness by the majority of the interviewees, but nearlya third rated its effectiveness to be “poor.” The latterrating came from situations where there was notenough post-fire organic debris in riparian areas or thechannels to cause debris dam problems or streamhydrology was adversely altered by clearing. Becausemuch of the debris from fire-killed trees does not enterchannel system until 2 or 3 years later, this treatmentwas not considered by some to be a useful BAERtreatment. Also, there has been a significant improve-ment in the understanding of the positive role of largewoody debris in trapping sediment, dissipating theenergy of flowing water, and providing aquatic organ-ism habitat. In some instances the channel clearinghas been more disruptive than the wildfire. So, insome areas the policy now is to avoid channel clearing.

Channel clearing is definitely an expensive, time-consuming operation, but it has been successful incertain situations such as locations where trash rackscannot be used to protect road culverts, where woodydebris might move into reservoirs, and where sedi-ment must be removed from debris basins and chan-nels to provide adequate sediment storage capacity.Important factors in the relative effectiveness ofchannel clearing, when it is used, include a goodanalysis of risk and the value of resources at risk,knowledge of the size and quantity of material toremove, the clearing distances above roads needed toprotect culverts, and understanding of the physical

characteristics of the channels which might aggra-vate or reduce stormflows.

Implementation and Environmental Factors: Timingis an important factor which affects both the effective-ness and the assessment of the value of channelclearing. When sediment removal is the objective ofchannel clearing, operations must be done before sea-sons (usually winter) that produce the first or mostsignificant stormflows. For large woody debris, thekey question is if and when inputs of woody debris arelikely to occur. In some areas, woody debris recruit-ment (greater than 2 years) may be beyond thetimeframe of BAER projects. Crews conducting chan-nel clearing must be well trained in order to recognizewoody material that is too large to float or be firmlyanchored, is part of the natural instream coarse woodydebris load, or is a natural grade stabilizer. Wherewoody debris is cut up it must be sufficiently short topass through culverts.

Other Factors: Channel debris clearing may producesignificant, adverse riparian area impacts, destabi-lize the channel, reduce aquatic habitat, and alterstream hydrology. These side effects may negate anypositive benefits derived from channel clearing insome situations.

Stream Bank Armoring/Channel Armoring

Purpose: Stream bank and channel armoring is doneto prevent erosion of channel banks and bottomsduring runoff events. In some hydrologic systemsstream banks are a major source of sediment.

Relative Effectiveness: Not enough interviewees ratedthis treatment.

Comments on armoring indicated that it functionswell in small, ephemeral drainages or near the headsof larger ephemeral drainages, and lower gradientareas. In steep terrain, sloughing of upslope materialscan bury the bank armoring.

Implementation and Environmental Factors: Streambank armoring requires proper design, a well-devel-oped implementation plan, and experienced crews formaximum effectiveness. Other implementation fac-tors that contribute to success include proper sizedmaterials, use of geotextile fabric, avoiding overlysteep areas, and the use of energy dissipators.

In-Channel Felling

Purpose: This BAER channel treatment is designed toreplace woody material in drainage bottoms that havebeen consumed by wildfire. It is intended to traporganic debris and temporarily detain or slow downstorm runoff. Woody material felled into channels willultimately alter channel gradient, and may causesediment deposition and channel aggradation.



It is difficult to assess the relative effectiveness of thistreatment because no monitoring information wasavailable and few visual observations have been made.Log jams created by felling trees into channels havethe potential to detain sediment, but there is littlecredible confirmation of this potential.

Implementation and Environmental Factors: Goodpre-planning, supervision, and a high level of crewexperience are crucial to successful implementation ofthis treatment. Skilled tree fellers and chainsaw crewsare vital to implementation of this treatment. Crewsneed to be able to judge correct log spacing, position-ing, and adequate contact with the streambed. Thistreatment can be implemented only where there is agood supply of dead trees near the channels. Also, caremust be taken not to use large trees because they donot work as well as smaller ones. Snags can be felledparallel to channels to support channel banks or in v-shaped or other patterns to retain woody debris aboveroad culverts.

Log Dams

Purpose: Log dams, like straw bale check dams, areused to prevent or reduce sediment inputs into peren-nial streams during the first winter or rainy seasonfollowing a wildfire. They are constructed of moredurable material than straw bale dams. Log damsfunction by decreasing water velocity and detainingsediment-laden surface runoff long enough for coarsersediments to deposit behind check dams. Decreasedwater velocity also reduces downcutting in ephemeralchannels.


Log dams were rated “excellent” and “good” in theireffectiveness as a BAER treatment by the limitednumber of interviewees who commented on log dams.Well-built log check dams can be 70 to 80 percenteffective in trapping sediment and last 15 to 30 years.The amount of sediment trapped is highly variabledepending on the size of the dam. In one location,individual log dams were reported to trap up to 40 yd3

(40 m3) of sediment without failure. They can be veryeffective in adding to channel stability and keepingsediment onsite. On the negative side, failures due toundercutting, bypassing, and complete blowout haveaggravated erosion problems by producing deep scour-ing at dam sites and release of large amounts ofsediment in pulses. Despite these potential problemsand situations where 25 percent failed in the firststorm, no one rated log dams as “fair” or “poor” ineffectiveness.

Implementation and Environmental Factors: Likestraw bale check dams, a key factor in log dam con-struction is having a skilled implementation leaderand trained, experienced crews. Log check dams arecostly and labor intensive, requiring six to eight timesthe labor for installation than straw bale check dams.Design features such as appropriate size of watershed,dam orientation, log sizes, lateral keying (1.5 to 3 ft(0.4 to 1 m) into banks), spillways, contact with thestream bed, plugging of gaps, and energy dissipatersare important implementation considerations. SomeBAER coordinators recommend that log dams neverbe put in fully functional channels, but others recom-mend that log dams can be used to replace coarsewoody debris burned out of small perennial channels.Often rocks are used in conjunction with log dams.

Other Factors: In some locations, there might not beadequate woody material after a fire to build log dams.

Debris Basins

Purpose: Debris basins are constructed to treat eitherthe loss of control of runoff and deterioration of waterquality, or threats to human life and property. Thedesign of debris basins must be to a standard that theyprovide immediate protection from flood water, float-able debris, sediment, boulders, and mudflows. Theyare usually constructed in stream systems with nor-mally high sediment loads. Their purpose is to protectsoil and water resources from unacceptable losses or toprevent unacceptable downstream damage. Debrisbasins are considered to be a last resort because theyare extremely expensive to construct and require com-mitment to annual maintenance.


In order for debris basins to function they must be ableto trap at least 50 percent and preferably 70 to 80percent of 100-year flows. A spillway needs to beconstructed in the debris basin to safely release flow inexcess of the design storage capacity. The downstreamchannel should be lined to prevent scour. In someinstances excavated pits in ephemeral channels havebeen used as debris basins. These must be largeenough to trap 50 to 90 percent of flood flow. They needto be cleaned annually until abandoned.

Implementation and Environmental Factors: Becausedebris basins are rather large, they require design byqualified engineers. They are built in depositional orrunout areas that have large storage capacity. Duringconstruction it is important to maintain the channelgradient. Head cutting can result from improperlylocated or constructed debris basins.

Other Factors: Debris basins must be designed withlarge vehicle access to the basins so they can be


cleaned out periodically. Maintenance is a key factorin effectiveness of this treatment. Although protectionis immediate, maintaining debris basins may be along-term commitment.

Straw Wattle Dams

Purpose: Straw wattle dams work on the same princi-pal as straw bale check dams. They trap sediment onside slopes and in the upper ends of ephemeral drain-ages by reducing channel gradient. Straw wattles areeasy to place in contact with the soil and provide a lowrisk barrier to soil movement.


The limited number of interviewees that rated thistreatment scored straw wattle dams as “excellent” or“good” in terms of controlling movement of sediment inchannels. In one instance, only 10 of 3,300 wattlesfailed during the first storm after installation. An-other reported an 80 percent first-storm survival rate,and excellent channel energy dissipation and trappingof sediment.

Implementation and Environmental Factors: Like anyother channel treatment, good plans, designs, andexperienced crews go a long ways to ensure successfulimplementation. Straw wattles work best on firstorder ephemeral channels with slopes less than 45percent gradient. They can be easily placed by rela-tively untrained crews since they conform to the soilsurface very well. This is a distinct advantage overrigid barriers like logs. Placement of straw wattlecheck dams is easiest on loamy sand soils that can bereadily excavated. The closer together straw wattlesare placed in steep terrain the more effective they arein detaining sediment. “U” shaped re-bar is very effec-tive in keeping straw wattles fastened down but isanother factor to consider in the logistics plans for thistype of BAER project. Shallow or rocky soils can causeproblems with re-bar usage, but hard pans can bepenetrated by driving the re-bar. Straw wattle damsare a good alternative in burned areas where logs areabsent, poorly shaped, or scarce. Wattles can be usedquite effectively in combination with straw balecheck dams. They also can be easily prepositionedby helicopters.

Other Factors: Straw wattles are relatively cheap tobuy. They can be disturbed by grazing animals, decom-pose, and catch fire. Although the wattle netting isphotodegradable, there are concerns that it persistslong enough to pose hazards for small animals. Supplyis a major problem, particularly for a large project.There are concerns among some users about the costeffectiveness of straw wattle dams since the materialand labor costs are quite high.

Rock Cage (Gabion) Dams

Purpose: Also known as rock fence check dams, thesestructures are used in intermittent or small perennialchannels to replace large woody debris that may havebeen burned out during a wildfire. The rock cage damsprovide a degree of grade stability and reduce flowvelocities long enough to trap coarse sediments.


Comments by some individuals indicated favorableresults. On mild gradients these structures work well.Some failures occurred on steeper slopes when highvelocity flows are greater than 3 ft s–1 (1 m s–1). This isa common theme for all channel treatments. Most ofthe failures occur where treatments are imposed onsteep gradient sections of ephemeral or first to secondorder perennial channels. Rock cage dams often lastlong enough and trap enough fine sediments to pro-vide microsites for woody riparian vegetation to getreestablished. Rock cage dams on the Wenatchee Na-tional Forest were very successful, trapping 2000 to10,000 yd3 (1500 to 7600 m3) of material after just onestorm.

Implementation and Environmental Factors: Like mostother BAER channel treatments, proper dam designand installation by experienced crews are crucial tosuccess. The rock cage dams must be properly placed,keyed in, and anchored to stay in place during runoffevents. Downslope energy dissipators are recom-mended because they reduce the risk of the rock cagedams being undercut.

Other Factors: Construction of these structures isdependent on the availability of adequate amountsand sizes of rocks. Rock cage dams need to becleaned out periodically if they are to maintain theireffectiveness.

Road Treatments

Road treatments are implemented to increase thewater and sediment processing capabilities of roadsand road structures. They are not meant to retainwater and sediment, but rather to manage its erosiveforce.

Rolling Dips/Waterbars/Cross Drain/Culvert Overflow/Bypass

Purpose: These treatments are designed to providedrainage relief for road sections or water in the insideditch to the downhill side of roads especially when theexisting culvert is expected to be overwhelmed.

Relative Effectiveness: No interviewee rated thistreatment.


Environmental/Implementation Factors: Rolling dipsare easily constructed with road grader or dozer.Rolling dips or waterbars need to be deep enough tocontain the expected flow and location carefully as-sessed to prevent damages to other portions of theroad. Waterbars can be made out of rocks or logs.Armoring of fillslope at the outlet is often needed toprevent gullying.

Culvert Upgrade

Purpose: Culvert improvements increase the flow ca-pacity which will prevent damage to roads.


When sized properly and installed correctly, the re-sults were rated “good.” The “poor” rating was fromculverts that were still not large enough and failed.

Environmental/Implementation Factors: Upgradedculverts need to be sized properly based on expectedincreased flows. They should be installed at the properslope with appropriate approaches and exits. To beeffective, upgraded culverts need to be installed beforethe first damaging rainfall. Flexible down spouts andculvert extensions often are needed to keep exitingwater from highly erodible slopes.

Storm Patrol

Purpose: Patrol during storm events provides immedi-ate assessment of flood risk, clear blocked culvertentrances, and drainage ditches and close access (gates)to areas that are at risk.


Several interviewees indicated that storm patrol wasa cost effective alternative to installing trash racks, orremoving culverts.

Environmental/Implementation Factors: This treat-ment can include early warning systems such as radio-activated rain gauge or stream gauge alarms whenflows are increasing. Storm patrols remove floatingwoody debris near culvert inlets and clean inlets aftereach storm event. Storm patrols can be activatedduring forecast events of weather which may triggerlarger than normal water, sediment or woody debrisflows.

Culvert Inlet/Outlet Armoring/Risers

Purpose: These treatments reduce scouring aroundthe culvert entrance and exit. They allow heavy par-ticles to settle out of sediment laden water and reducethe chance of debris plugging the culvert.

Relative Effectiveness: Not enough interviewees ratedthis treatment to make any statements about itseffectiveness.

Environmental/Implementation Factors: Sometimesculvert risers can clog and may be difficult to clean.

Trash Racks

Purpose: Trash racks are installed to prevent debrisfrom clogging culverts or down stream structures.


Comments included that in one watershed the thirdwinter after the fire, a large storm detached consider-able debris which blocked trash rack, causing com-plete culvert failure.

Environmental/Implementation Factors: These struc-tures are generally built out of logs, but occasionallythey are from milled lumber or metal. Sizes vary fromsmall culverts to 30 ft (9 m) diameter. Several cagedesigns have been used with most of them allowingdebris to ride up and to the side of the cage. Some cageshave been set in concrete. Trash racks generally per-form better in smaller drainages. They need to becleared after each storm to be effective.

Culvert Removal

Purpose: This procedure removes undersized culvertswhich would probably fail due to increased flows, in acontrolled fashion.


Environmental/Implementation Factors: Removalneeds to be completed before the first damaging storms.It is often done in conjunction with road obliteration.

Ditch Improvements: Cleaning/Armoring

Purpose: Cleaning and armoring provides adequatewater flow capacity and prevents downcutting ofditches.


Environmental/Implementation Factors: When main-tenance does not occur, high water levels can overtoproadways leading to gully development in the roadbed.

Armoring Ford Crossing

Purpose: Armored crossings provide low-cost accessacross stream channels that are generally capable ofhandling large flows.



Environmental/Implementation Factors: Largeriprap is placed upstream and downstream of actualroad crossing area. Armoried crossings are often usedfor low traffic volume roads. Low water crossing werenot used on one fire because they could attract anendangered toad species that would inhabit the cross-ing when wet and be killed by vehicle traffic.

Outsloping

Purpose: Outsloping prevents concentration of flowon road surfaces that produces rilling, gullying, andrutting.


One interviewee commented that this is one of the fewtreatments that has both immediate and long-termfacility and resource benefits.

Environmental/Implementation Factors: Sometimesafter regrading, compaction does not occur due to lowtraffic volume which may cause sheet and rill erosion.Both public and administrative traffic should be cur-tailed during wet road conditions to prevent ruttingand road sub-grade damages.

Trail Work

Purpose: BAER treatments on trails are designed toprovide adequate drainage and stability so trails re-main functioning.

Relative Effectiveness: Not enough interviewees ratedthis treatment to make any statements about itseffectiveness.

Environmental/Implementation Factors: Crew skillis important for this labor intensive treatment. Waterbars need to be installed correctly, proper slope anddepth, to be effective.

Other Treatments

Purpose: This category consists of various treatmentsolutions to specific problems. It includes wettingagents to reduce water repellency on high erosionhazard areas, gully plugs to prevent headcutting inmeadows, flood signing installation to warn residentsand visitors of flooding potential, and removal of looserocks above roadways that were held in place by roots,forest debris, duff and were now in a precarious posi-tion due to the fire.

Relative Effectiveness: No interviewees rated thesetreatments.

Environmental/Implementation Factors: Treatmentspecific.

Federal Recycling Program Printed on Recycled Paper

The U.S. Department of Agriculture (USDA) prohibits discrimination in all itsprograms and activities on the basis of race, color, national origin, sex, religion,age, disability, political beliefs, sexual orientation, or marital or family status. (Notall prohibited bases apply to all programs.) Persons with disabilities who requirealternative means for communication of program information (Braille, large print,audiotape, etc.) should contact USDA’s TARGET Center at (202) 720-2600 (voiceand TDD).

To file a complaint of discrimination, write USDA, Director, Office of Civil Rights,Room 326-W, Whitten Building, 1400 Independence Avenue, SW, Washington,DC 20250-9410 or call (202) 720-5964 (voice or TDD). USDA is an equalopportunity provider and employer.

ROCKY MOUNTAIN RESEARCH STATIONRMRS

The Rocky Mountain Research Station develops scientific informa-tion and technology to improve management, protection, and use ofthe forests and rangelands. Research is designed to meet the needsof National Forest managers, Federal and State agencies, public andprivate organizations, academic institutions, industry, and individuals.

Studies accelerate solutions to problems involving ecosystems,range, forests, water, recreation, fire, resource inventory, land recla-mation, community sustainability, forest engineering technology,multiple use economics, wildlife and fish habitat, and forest insectsand diseases. Studies are conducted cooperatively, and applicationsmay be found worldwide.

Research Locations

Flagstaff, Arizona Reno, NevadaFort Collins, Colorado* Albuquerque, New MexicoBoise, Idaho Rapid City, South DakotaMoscow, Idaho Logan, UtahBozeman, Montana Ogden, UtahMissoula, Montana Provo, UtahLincoln, Nebraska Laramie, Wyoming

*Station Headquarters, Natural Resources Research Center,2150 Centre Avenue, Building A, Fort Collins, CO 80526

United States Department Evaluating the …...United States Department of Agriculture Forest Service...

Documents

Transcript of United States Department Evaluating the …...United States Department of Agriculture Forest Service...