Rules of the road: A qualitative and quantitative ... · Rules of the road: A qualitative and...

Geomorphology xxx (2016) xxxndashxxx

GEOMOR-05740 No of Pages 24

Contents lists available at ScienceDirect

Geomorphology

j ourna l homepage wwwe lsev ie r com locate geomorph

Rules of the road A qualitative and quantitative synthesis of large wood transportthrough drainage networks

Natalie Kramer Ellen Wohl Dept of Geosciences Colorado State University Fort Collins CO 80123 United States

Corresponding authorsE-mail addresses nkramerandersongmailcom (N K

EllenWohlcolostateedu (E Wohl)

httpdxdoiorg101016jgeomorph2016080260169-555Xcopy 2016 Elsevier BV All rights reserved

Please cite this article as Kramer NWohl Enetworks Geomorphology (2016) httpdx

a b s t r a c t

a r t i c l e i n f o

Article historyReceived 14 January 2016Received in revised form 18 August 2016Accepted 22 August 2016Available online xxxx

To effectively manage wood in rivers we need a better understanding of wood mobility within river networksHere we review primarily field-based (and some numerical) studies of wood transport We distinguish smallmedium large and great rivers based on wood piece dimensions relative to channel and flow dimensions anddominant controls on wood transport We suggest further identification and designation of wood transport re-gimes as a useful way to characterize spatial-temporal network heterogeneity and to conceptualize the primarycontrols on wood mobility in diverse river segments We draw analogies between wood and bedload transportincluding distinguishing Eulerian and Lagrangian approaches exploring transport capacity and quantifyingthresholds of woodmobilityWe identifymobility envelopes for remobilization of woodwith relation to increas-ing peakdischarges stream size and dimensionless log lengthsWood transport in natural channels exhibits highspatial and temporal variability with discontinuities along the channel network at bankfull flow and when loglengths equal channel widths Although median mobilization rates increase with increasing channel size maxi-mum mobilization rates are greatest in medium-sized channels Most wood is transported during relatively in-frequent high flows but flows under bankfull can transport up to 30 of stored wood We use conceptualmodels of dynamic equilibrium of wood in storage and of spiralling wood transport paths through drainage net-works as well as a metaphor of traffic on a road to explore discontinuous wood movement through a river net-work The primary limitations to describingwood transport are inappropriate time scales of observation and lackof sufficient data onmobility from diverse rivers Improvingmodels of wood flux requires better characterizationof average step lengthswithin the lifetime travel path of a piece ofwoodWe suggest that future studies focus on(i) continuous or high-frequency monitoring of wood mobility (ii) monitoring changes in wood storage (iii)using wood characteristics to fingerprint wood sources (iv) quantifying volumes of wood buried within rivercorridors (v) obtaining existing or new data from unconventional sources such as citizen science initiativesand (vi) creating online interactive data platforms to facilitate data synthesis

copy 2016 Elsevier BV All rights reserved

KeywordsInstream woodWood transportLarge woodRiver network

1 Motivation

The need is strong to understand large wood transport dynamics inthe context of flooding hazards (eg Mazzorana et al 2012Ruiz-Villanueva et al 2013 2014a Luciacutea et al 2015) and global nutri-ent fluxes (Bilby 1981 Elosegi et al 2007 Hilton et al 2008 West etal 2011Wohl et al 2012) Bywood dynamics wemean the processesassociated with the recruitment storage and transfer of dead woodthrough drainage basins Specifically we focus on the transport oflarge wood (ge1 m in length and 10 cm in diameter)

Conceptual and quantitative models of river adjustment typicallyrely on two primary driving variables water and sediment (Knighton1998) However instreamwood can be as important for channel change

ramer)

Rules of the road A qualitativdoiorg101016jgeomorph

as sediment (eg Massong andMontgomery 2000 Brooks and Brierley2002 Montgomery and Abbe 2006 Le Lay et al 2013) Early scientificand historical writings onwoodweremainly inspired by awed observa-tions of immense volumes of wood and associated channel change(Kindle 1919 Kindle 1921 Bevan 1948 Triska 1984) Reviews of his-torical accounts document the enormous quantities and landscape-scale impacts of wood from headwater channels to large rivers priorto extensive wood removal (Triska 1984 Sedell et al 1988 Wohl2014a) and most forested or historically forested catchments have veg-etation- or wood-driven morphologies (Hickin 1984 Collins et al2012 Gibling and Davies 2012 Polvi and Wohl 2013 Gurnell et al2015)

Despite this evidence for the importance of wood as a driving geo-morphic variable descriptions of geomorphic effects of wood on chan-nel processes were largely absent in the mid-1900s (Hickin 1984)when researchers were building foundational conceptual models(Grant et al 2013 Wohl 2014b) - such as graded rivers (Mackin

e and quantitative synthesis of largewood transport through drainage201608026

Table 1Functional classification of river size based on wood dynamics Categorization based oncharacteristic dimensions of wood pieces relative to channel size and patterns of recruit-ment storage and transport of wood after Keller and Swanson (1979) Church (1992)Nakamura and Swanson (1993) Pieacutegay and Gurnell (1997) Gurnell et al (2002) andWohl (2016 this issue)

Small rivers

Size 1st to 2nd order key piece wood length N channel width diameterof logs N flow depth

Recruitment Characteristics of riparian woodland of overriding importancehillslope instability blowdown snow avalanches individualriparian tree mortality

Storage Individual wood pieces form stable features that control sedimentdeposition channel morphology and gradient pieces typically spanand are suspended above the active channel jams may be presentbut most wood stored as individual pieces distribution of woodclose to random and governed by sites of input wood causes localwidening where channel boundaries are erodible wood conditiondepends on input mechanisms and forest disturbance history lowmobility can result in high levels of wood decay

Transport Wood not reorganized or transported except during unusual floodsor debris flows with recurrence intervals N decadal large piecesmostly immobile long residence times regulated by decay andphysical breakdown rather than fluvial transport

Medium rivers

Size 2nd to 4th order log diameter ~ flow depth length of key logs N or ~channel width

Recruitment Hillslope riparian and bank source areas tributary inputstransport from upstream

Storage Wood commonly stored in non-random spatially discontinuousjams that partly or completely span the channel jams comprised ofmixed sizes of wood pieces with larger key pieces trapping smallermore mobile pieces jams form at roughness elements such asboulders and planform irregularities jams force bank erosion andavulsion can create multithread planform high inter-reachvariation in wood loads wood regulates sediment transport

Transport Hydrologic regime is dominant control drives periodic transport ofstored and newly recruited pieces during high flows key pieces andjams remain in place during smaller floods accumulating smallerpieces larger more infrequent floods break up and rearrange jams

Large rivers

Size ge5th order log diameter ≪ depth of flooded channel all woodlengths b channel width transition between medium and largerivers at channel widths ~20ndash50 m wide because longest woodpiece lengths are in this range

Recruitment Transport of wood from upstream lateral channel erosionexhumation of previously buried wood on floodplains

Storage River morphology dictates wood storage sites wood generally hasless decay because of greater mobility wood influences bar andchannel sedimentation and regulates formation of secondarychannels channel planform and widening of valley floor reaches

2 N Kramer E Wohl Geomorphology xxx (2016) xxxndashxxx

1948) and Lanes balance (Lane 1955) and quantitative understandingof river systems - such as hydraulic geometry relations (Leopold andMaddock 1953) The large-scale removal of wood from rivers duringthe 20th century (Montgomery and Pieacutegay 2003 Wohl 2014a) likelyled to a neglect of the geomorphic effects of wood on river processand form during this formative period of fluvial geomorphology

As long as rivers are wood-depleted models of river form adjustingprimarily to sediment and water are adequate for managing riversHowever in order to improve valued ecosystem services such as de-ni-trification sustainable fisheries improved water quality and enhancedphysical and mental health of human communities (eg Wohl et al2015)managers now focus on reintroducingwood as engineered struc-tures (Abbe and Brooks 2011 Gallisdorfer et al 2014) leaving mobileun-engineered wood in place on the floodplain (Pieacutegay et al 2005Wohl et al 2015) andmanaging riparian forests to increase the amountof wood that can be recruited (Kail et al 2007 Wohl et al 2015)

In Europe afforestation of river corridors (Lieacutebault and Pieacutegay 2002)has led to increased wood in transport resulting in wood impound-ments against bridge piers during floods (eg Luciacutea et al 2015) in-creasing flood levels and forcing large-scale channel changes beyondwhat can be predicted through existing models (Pieacutegay et al 2005Ruiz-Villanueva et al 2014b) Inability to plan for the impacts and haz-ards of large wood on channels is a management concern and stemslargely from the fact that current models of river form and processused by managers do not include wood dynamics

In the past few years substantial progress in numerical simulation ofwood transport has come from flume studies (Bertoldi et al 2014Davidson et al 2015) and incorporation of woodmodules into comput-er simulations such as the reach scale channel simulator for habitatmodelling (Eaton et al 2012 Davidson and Eaton 2015) and iberwood for hydrodynamic modelling of wood transport (Ruiz-Villanuevaet al 2014a 2015c) Simulations from these models have led to betterunderstanding of wood-related flooding hazards channel change andaquatic habitat Howevermodels are simplified versions of the complexinteractions found in the field and they require information from fieldstudies to constrain variables and assess reliability of scenario responses(Ruiz-Villanueva et al 2015b) Recent rapid growth in numerical andmodelling publications onwood transport is notmatched by equivalentgrowth in field studies (Fig 1)

In this paper we first qualitatively and then quantitatively summa-rize analyze and synthesize literature on wood transport within theframework of a functional classification of small medium large andgreat rivers (Table 1 Fig 2) We focus on presenting a thorough reviewof case studies that have directly measured wood transport We exam-ine existing transport premises present new conceptual frameworks

Flume or Modelling Studies

Field Studies

1

2

7

5

1

2

4

Fig 1 Comparison of field and modelling publications that report direct measurement ofwood transport by year Bar heights equal total number of articles found

with greater sinuosity more bars and lower gradients typicallyhave higher wood loads wood accumulates within active channelat sites such as apex of bars outer upstream facing margins ofchannel bends backwaters bank benches and as individual piecesalong channel margins during flood recession large proportion ofwood may be buried in floodplains and channel bars woodaccumulates on upstream side of living vegetation and influencesmorphologic evolution of stable vegetated islands and formation ofmultithread planform wood does not form channel spanning jamsbut can form channel spanning dynamic wood rafts that persist fordecades and cause channel avulsions

Transport Wood exported downstream laterally onto floodplains or buriedwood exported regularly during high flow amount of wood transferhighly variable and largely dependent on pattern of antecedentpeak flows high variability in water levels during flooding createsmany opportunities for wood sequestration in long-term storage onthe floodplain causing large variability in wood residence time

Great riversa

Size ~106 km2 or larger mean annual discharge N 103 m3s perennialflow commonly have vast seasonally inundated floodplains greathydraulic diversity large fine sediment loads and deep channels

Recruitment Mostly via inputs from upstream local inputs include lateralexchange with floodplain exhumation of buried wood inputs from

Please cite this article as Kramer NWohl E Rules of the road A qualitative and quantitative synthesis of largewood transport through drainagenetworks Geomorphology (2016) httpdxdoiorg101016jgeomorph201608026

Table 1 (continued)

Small rivers

bank failures and mass recruitment such as during hurricanes(Phillips and Park 2009)

Storage Pieces stranded along banks during receding flows larger woodloads downstream from tributary junctions that deliver woodfloodplain accumulations at long lateral distances from channelaccumulations commonly decay in place or are partly buried raftsthat completely plug the channel are rare except where amultithread planform occurs

Transport During most flows rapid transfer of wood to deposition zones suchas deltas estuaries or the ocean lesser rapid fluctuations indischarge stage than large rivers create fewer opportunities fortrapping of wood within channel or overbank deposition woodtransfer largely controlled by the spatial distribution and timing ofwood recruitment from large tributaries floodplain wood likelytransported and redeposited within the floodplain rather thanretransported to main channel wood buried within channel bedmay be transported downstream annually as part of the bed load

a Very little study has been done onwood in great rivers This summary is based on ourcurrent understanding of wood dynamics in great rivers from personal field experienceThe ideas presented should be tested and refined with further study

Fig 2 Streamsizewith relation towood storagepatterns in afluvial system Schematic afterNaiRandom distribution of pieces on Ouzel Creek Colorado through a burned area (~6 m channejamming on North St Vrain Creek Colorado (~25 m channel width and average piece lengthisland formation in Hyland River British Columbia (channel width is ~100 m wide) (D) Buriaat 6618510 N minus14811925 E)

3N Kramer E Wohl Geomorphology xxx (2016) xxxndashxxx

Please cite this article as Kramer NWohl E Rules of the road A qualitativnetworks Geomorphology (2016) httpdxdoiorg101016jgeomorph

of wood transport through drainage networks and make suggestionsfor avenues of further research

Although other reviews have included wood transport as a subcate-gory (Keller and Swanson 1979 Harmon et al 1986 Sedell et al 1988Gurnell 2003 Le Lay et al 2013Wohl 2016) this is the first review ofwhich we are aware that focuses solely on transport We intend the re-view to be a thorough summary to acquaint newwood researcherswiththe existing literature a compilation of field-based constraints to com-pare to future numerical models and as a collection of thought-provok-ing ideas intended to push researchers to develop new testablehypotheses thus rapidly advancing wood transport research

2 Qualitative summary

Wood in storage is commonly referred to as instreamwood Howev-er once wood becomes mobile especially in larger channels re-searchers tend to prefer the terms floating wood (eg MacVicar andPieacutegay 2012 Ruiz-Villanueva et al 2014b) or drift wood (egSchmocker and Hager 2011 Kramer and Wohl 2014) Denotingwood as floating or drift (rather than instream)makes sense in the con-text of the transfer of wood through drainage networks especiallywhen wood is referenced along channels and in nonchannelized set-tings such as lakes estuaries or oceans (eg Kramer and Wohl 2015)

man et al (2002) Keller and Swanson (1979) Le Lay et al (2013) and Schumm(1977) (A)l width average piece length is 27 m with maximum pieces N width of channel) (B) Logis 33 m maximum piece size b channel width) (C) Wood accumulations and vegetatedl of wood accumulations on side bars of the Yukon River Alaska (~500 m channel width


Table 2Synthesis of main findings and references for influence of flow characteristics on woodtransporta

Flow characteristics

Magnitude

bull Flow magnitude strongly influences whether wood is in transport because itdictates the flooded cross-sectional area velocity and depth of flow

bull The largest wood fluxes on rivers of all sizes occur during infrequent high flowsbull Background wood mobility rates under 30 exist for wood movement initiationin ordinary floods for many rivers

bull Wood transport responds non-linearly to increases in flow magnitudes and ishighly variable

bull Hysteresis in wood transport flows of equal magnitude transport much lesswood on the falling limb than the rising limb

bull Wood mobilization threshold exists at magnitudes less than bankfull before thethreshold transport is negligible after the threshold transport is possible andincreases linearly with discharge until an upper wood transport threshold asso-ciated with overbank flows is reached at which point wood transport suddenlydecreases or levels off

bull Most wood is deposited near peak flow magnitudebull Floods strand wood at discrete elevations and locations depending on stageheight elevations of stored wood can be used indirectly to assess wood flux forparticular flows

bull Newly recruited wood can re-organize into jam stable states after only onebankfull flow jams are re-mobilized and re-organized during exceptional flows

bull Wood transport velocity is more significantly related to log volume than mag-nitude of floods

Duration

bull The amount of time between initiation of wood transport and peak flow is abetter predictor of wood displacement downstream than magnitude of flowalone

bull Newly recruited wood during floods moves farther distances downstream thanpreviously recruited wood

Rising and falling limbs

bull Mobilization of wood occurs on the rising limb and is comparably negligible onthe falling limb

bull New wood on the falling limb originates from morphological changes of thechannel

bull Wood in transport on the falling limb is rapidly retained and entrappedbull Flashier more steeply rising hydrographs mobilize more woodbull Transport and storage patterns respond nonlinearly to wood input rates (rate ofrising limb) with a threshold input rate governing transition between congestedand uncongested transport and corresponding depositional pattern as jams orsingle pieces (high or low wood loads)

bull Shorter pieces (ie broken pieces and branches) are transported earlier on therising limb and follow a consistent relation with discharge larger pieces aremobilized after small pieces and do not correlate well with discharge

bull Short pieces are transported on the rising and falling limbs whereas the largestpieces are mainly transported on the rising limb

Flow history

bull The amount of wood available for transport for any flood is a function of pastflow history and non-fluvial recruitment since the last wood transporting flow

bull Flood peaks of similar magnitude will have varying wood loads based on theirposition in a sequence of floods

bull In small streams the frequency of extreme events governs wood transferdownstream with cycling of wood storage related to recurrence intervals ofdebris flows

bull In medium rivers low recurrence yearly flows re-organize individual pieces ofwood into jam stable states whereas exceptional floods re-organize jams

bull Newly recruited wood is less stable than previously transported woodbull In large rivers reach-scale wood storage can be highly variable year to year anddepends on flows during prior years

a Field references Keller and Swanson (1979) Lienkaemper and Swanson (1987) Berg et al(1998) Keim et al (2000) Haga et al (2002) Angradi et al (2004 2010) MacVicar et al(2009) MacVicar and Pieacutegay (2012) Bertoldi et al (2013) Turowski et al (2013) Kramer andWohl (2014) Schenk et al (2014) Iroumeacute et al (2015) Jochner et al (2015) Ravazzolo et al(2015ab) Kramer et al (in press)Modelling references Braudrick et al (1997) Bertoldi et al (2014) Ruiz-Villanueva et al(2014a 2015c 2016b) Davidson et al (2015)


Wood flux is governed by the unique interactions among hydraulicswood pieces and channel morphology across space and time (Gurnellet al 2015 Ruiz-Villanueva et al 2015c) Thus we summarize theexisting knowledge of wood transport from field and numerical studiesin three categories flow characteristics (Section 21 Table 2) woodcharacteristics (Section 22 Table 3) and reach characteristics (Section23 Table 4) Tables 2ndash4 are used to summarize the literature In the fol-lowing text we propose a series of ideas hypotheses and suggested fu-ture research directions based on the findings from the paperspresented in Tables 2ndash4 We summarize some of the main hypothesespresented in the following text in Table 5

21 Flow characteristics

The primary flow characteristics that influence wood mobilizationand transport include general shape of the hydrograph magnitude du-ration and rate of rise and fall and the sequence of flows through time(Table 2) Flow magnitude influences the areas inundated and the hy-draulic roughness and retention capacity of the flooded area Althoughwood discharge generally increases with water discharge on the risinglimb the relationship is nonlinear and highly variable (MacVicar andPieacutegay 2012 Kramer and Wohl 2014 Ruiz-Villanueva et al 2014aRavazzolo et al 2015b) This is logical given the potential for multipleand complex interactions among piece size and shape hydraulic forcesand channel characteristics such as boundary irregularities that can trapwood

The rate of discharge increase may be important for estimatingwood transport rates and rapid changes in thehydrographmay accountfor much of the variability in wood discharge on the rising limb(Braudrick et al 1997 Angradi et al 2004 MacVicar et al 2009)Some of the observed variability in wood transport may reflect woodinput rates associated with processes such as bank erosion floating ofpreviously stable pieces and piece-to-piece interactions among woodin transport (Braudrick et al 1997 Keim et al 2000 Bertoldi et al2014 Wohl 2016) Thus we propose that transport distance will bemore limited for flashier floods because more wood will be mobilizedin shorter amounts of time leading to congested transport and in-creased wood deposition especially as jams (Braudrick et al 1997Bertoldi et al 2014) Higher densities of jams deposited near peakflows then limit transport distance of uncongested single pieces onthe falling limb (Davidson et al 2015)

We hypothesize that the duration of flows near or just underbankfull exerts the greatest influence on wood transport distance inmeandering rivers Transport capacity is likelymaximized near bankfullwhen smaller-scale channel roughness features occupy a smaller pro-portion of the total flow and overbank roughness and wood trappingare not yet factors Longer duration floods just under bankfull mayalso result in greater bank erosion from maximization of streampower directed at banks during these flows Catastrophic bank failuresduring floods mostly occur on the falling limb (Rinaldi et al 2008) orbetween long duration moderate floods (Luppi et al 2009) from ero-sion of the bank on the outside of a bend (Pizzuto 2009)Wood that en-ters the river from these bank collapses falls directly into the thalwegand moves longer distances downstream than previously fluvially de-posited instream wood (MacVicar and Pieacutegay 2012) Consequentlytransport distances may also be longer for flows with longer durationof stage just under bankfull on the falling limb

Larger wood transport rates have been observed when water levelsincrease rapidly (MacVicar et al 2009 Ruiz-Villanueva et al 2016b)but not enough studies have focused onwood transport rates in relationto hydrograph characteristics specifically rates of hydrograph rise andfall to provide robust constraints or thresholds for initial downstreamor lateral mobility Thus uncertainty remains regarding what types offlood hydrographs facilitate the greatest downstream versus lateralmovements of wood and whether threshold rates of change in waterdischarge exist that dictate how and where wood will be deposited for


Table 3Synthesis of main findings and references for influence of wood characteristics on woodtransporta

Wood characteristics

Anchoring

bull How well a piece of wood is anchored is the most important variable governinginitial wood mobilization anchoring types include burial bracing against otherpieces or instream obstructions ramping or bridging onto banks roots of livingvegetation growing from logs rootwads

bull The single most important anchoring mechanism for initial mobilization ofnon-rooted pieces is burial

bull Bracing by other pieces of wood is more common and more effective at limitingmobilization than bracing against other in-stream elements such as boulders

bull Wood loads and jams are related to spacing of mobile ramped piecesRootwads

bull Presence of a rootwad limits travel distance and initial mobilizationbull Shorter pieces with rootwads may travel the shortest distancesbull Rootwads influence log steering in transport and pivoting during entrainmentand entrapment

Length

bull A length mobility threshold exists near or above bankfull width wood less thanbankfull width is more easily mobilized and travels farther

bull Piece length is a better predictor for whether a piece will move out of a reachthan volume or diameter for nonbraided rivers

bull Shorter pieces (ie broken pieces and branches) are transported earlier on therising limb because they are preferentially transported longer distances on thefalling limb of previous floods and deposited at lower elevations

bull Shorter pieces of wood may not always move before longer pieces of wood dueto shielding and bracing against larger pieces

bull In medium rivers shorter pieces generally travel longer distances than longerpieces but there are instances when the longest pieces travel the longest dis-tances

bull In large rivers there is no consistent relation between transport distance andlength of wood

bull In braided rivers medium sized logs travel the farthest distancesDiameter

bull A threshold for transport exists when flow depths = a critical floating depthrelated to the diameter and density this is ~05 log diameter

bull Diameter is important for controlling the timing and elevation of wooddeposition especially in streams where inundated depth contracts rapidly withsmall changes in discharge such as in broad braided rivers or floodplains

bull There is weak predictive power between log diameter and distance travelled insingle thread channels and high predictive power in braided channels

bull Most instream logs float more than half way under water and are denser than inthe forest environment

Orientation

bull Single pieces with rootwads are commonly deposited oriented parallel to flowwith the rootwad on the upstream side which can then entrap more wood

bull Flume studies indicate perpendicular pieces are mobilized sooner than parallelpieces while field studies indicate the opposite natural complexities in pieceshapes and anchoring are likely the source of the discrepancy

bull Wood travels parallel to flow unless interacting with other pieces (thenperpendicular) ruddering from branches or rootwads or travelling throughvelocity heterogeneities

bull Despite smaller forces acting on the parallel pieces they are more mobile duringfloods than perpendicular or oblique pieces which are commonly anchored orbraced against other wood or banks

bull Loose unanchored pieces originally stored oriented at an angle to flow travelfarther distances most likely because once mobile they ferry across the currentaway from snags on the bank and enter the swifter velocities of the thalwegsooner

Species (also decay density and branching complexity)

bull Tree species is a master variable that governs many other mobility predictors(ie length breakage density branching complexity shape rate of decay)

bull Field studies find species to be the best predictor of variance in transport dis-tances

bull Travel distance decreases with increasing densitybull Density and shape strongly influence travel speed and entrapment during floodrecession

Table 3 (continued)


bull Newly recruited green wood moves differently than fluvial wood due to differ-ences in abrasion water absorption and buoyancy

a Field references Harmon et al (1986) Lienkaemper and Swanson (1987) Gippel et al (1996)Pieacutegay and Gurnell (1997) Berg et al (1998) Jacobson et al (1999) Keim et al (2000) Haga etal (2002) Lassettre and Kondolf (2003) Daniels (2006) Millington and Sear (2007) Warrenand Kraft (2008) Wohl and Goode (2008) MacVicar et al (2009) Cadol and Wohl (2010)Iroumeacute et al (2010 2015) Merten et al (2011 2013) MacVicar and Pieacutegay (2012) Bertoldi etal (2013) King et al (2013) Beckman and Wohl (2014) Dixon and Sear (2014) Schenk et al(2014) Shields et al (2006) Ravazzolo et al (2015ab)Modelling references Bocchiola et al (2006ab) Braudrick and Grant (2000 2001)Welber et al (2013) Bertoldi et al (2014) Davidson et al (2015) Ruiz-Villanueva et al(2015a 2016a)



different channel types Also more work should be done to characterizethe variability in speed of wood transport related to piece type and flowmagnitude for diverse river morphologies In meandering rivers logsfloatwith few obstructions whereas in braided rivers numerous obsta-cles cause logs to stop and be reentrained several times reducing the av-erage velocity MacVicar and Pieacutegay (2012) found that uncongestedtransport of logs moved at approximately the surface velocity of thewater in a single-thread meandering channel whereas Ravazzolo etal (2015b) found that wood moved at approximately half the celerityof peak flow for a braided gravel-bed river

Although higher floods mobilize wood that remains stable duringlower floods (MacVicar and Pieacutegay 2012) the relationship betweenflood magnitude and large wood flux is highly variable (Kramer andWohl 2014 Iroumeacute et al 2015 Kramer et al in press) and dependson flow history (Haga et al 2002 Kramer et al in press) The lack ofconsistent relations between flow magnitude and large wood flux be-cause of the influence of priorflows is analogous to the lack of consistentcorrelations between flowmagnitude and channelmorphologic change(Harvey 1984 Desloges and Church 1992 Cenderelli andWohl 2003)Because wood deposition mostly occurs near peak water discharge(MacVicar and Pieacutegay 2012 Ravazzolo et al 2015b) we suggest thatprior patterns of flood peak magnitudes may have high predictivepower regarding how much wood is available for downstream trans-port in future floodsWe hypothesize that the frequency of wood redis-tributionwithin the channel banks is largely scale-dependent such thatwood is redistributed more frequently with increasing river size Wepredict that further research will reveal a discontinuity in transportflux through drainage networks simply because the frequency of redis-tribution and delivery is dissimilar between river types and sizes

Turowski et al (2013) have shown that for the steep step-poolheadwater Erlenbach River in Switzerland coarse particulate organicmatter (CPOM) transport rates (including finer material such as twigsleaves and large wood) consistently scale with discharge Furthermorethey suggest that consistent power relations exist for transported piecesize distributions of all organic matter and that if transport rates of finerfractions of CPOMcan bemonitored transport rates of largewood couldbe estimated and vice versa This idea holds some promise asMacVicar and Pieacutegay (2012) found that smaller pieces of largewood followed consistent relations with discharge on the risinglimb of a flood whereas larger pieces did not Turowski et al(2013) may have found more consistent relations between woodtransport and discharge magnitudes compared to other large woodstudies (MacVicar and Pieacutegay 2012 Kramer and Wohl 2014) be-cause they included the more abundant finer fraction of CPOM andor because a stronger relation between discharge and wood exportexists in steep headwater channels with limited floodplains Wesuggest further testing the broad applicability of estimating wooddischarge based on monitoring subsets of smaller size fractionsthat have consistent relationships with flood magnitudes and thenback-estimating quantity of larger wood based on knowntransported size distributions for a particular river


Table 4Synthesis of main findings and references for influence of reach characteristics on woodtransport

Reach characteristics

Channel morphology

bull Wood will be routed more quickly and stored for less time in reaches that areconfined versus unconfined multi thread versus single thread higher slopesversus lower slopes flow regulated versus unregulated smaller variability inchannel depths of the flooded cross-section

bull Best predictors for wood mobility and transport distance change based onchannel type in reaches with high connectivity to floodplains wood character-istics are better predictors whereas in channelized reaches hydrologic variablesare better predictors

bull Wood transport distances and mobilization are substantially reduced on flood-plains versus the main channel

bull For steep coarse substrate confined channels roughness during low flowslimits transport mobility and distance travelled as flows overtop roughnesselements transport capacity increases and wood flux is limited by supply

bull For alluvial channels with floodplains where low flow roughness is low theduration and depth of flow over sand bars and supply of wood limit transport atflows under bankfull whereas vegetated bars and floodplains limit transport asflows begin to overtop the regularly flooded channel

bull Degree of sinuosity may not limit transport distanceWood abundance and jams

bull Large wood and wood jams are commonly the dominant entrapment site forfluvially deposited wood

bull Wood pieces incorporated into jams are harder to mobilize than individualpieces

bull Storage frequency of jams decreases overall reach-scale mobility of woodbull Single pieces of wood can travel past jams during moderate and high flowbull Channel spanning wood jams may have a limited impact on the travel distanceof wood because most wood is transported during very high flows when jamsare floated water flows around them or over them or the jams are transported

bull Log jams are frequently mobile and often exchange pieces or are transportedand re-form in the same location with new pieces but similar structure

bull Most jams are mobilized by channel change during high flows or failure of keypieces due to breakage or decay

bull Jam frequency and location are related to long ramped piecesbull Increased wood abundance in storage relates to increased wood recruitmentduring floods

Live-wood

bull Living vegetation within the flow (live-wood) is extremely effective at trappingand anchoring wood especially key pieces that later form jams

bull Floods create heterogeneity in riparian vegetation patterns based on locations ofdeposited wood piles that re-sprout or provide nursery sites for new live-wood

Forest disturbance history

bull Forest disturbance history is closely related to dead and live wood abundancewhich in turn influences mobility

bull Extreme floods in channels with high wood loads recruit more wood than floodsin channels with low pre-flood wood loads

bull Wood mobilization is lower in old-growth reaches due to associated increase inchannel complexity and supply of large key pieces that form jams

bull Wood is more mobile with higher export rates and greater travel distances inrecently burned versus unburned catchments

bull Large infrequent catastrophic disturbances re-set the template for wood distri-bution patterns and channel change in subsequent years

bull Peak wood loads occur decades after catastrophic forest mortality events

a Field references Murphy and Koski (1989) Benke and Wallace (1990) Young (1994)Pieacutegay and Gurnell (1997) Berg et al (1998) Jacobson et al (1999) Johnson et al(2000) Gurnell (2003) Wyżga and Zawiejska (2005) Shields et al (2006) Pettit et al(2006) Millington and Sear (2007) Oswald and Wohl (2008) Curran (2010) Sear et al(2010) Iroumeacute et al (2010) Wohl and Goode (2008) Wohl and Cadol (2011) Collinset al (2012) MacVicar and Pieacutegay (2012) King et al (2013) Le Lay et al (2013) Beckmanand Wohl (2014) Dixon and Sear (2014) Boivin et al (2015) Jackson and Wohl (2015)Luciacutea et al (2015) Wohl (2016) Modelling references Braudrick et al (1997) BraudrickandGrant (2001) Bertoldi et al (2014 2015) Davidson et al (2015) Ruiz-Villanueva et al(2015b 2016b)

Table 5Hypotheses regarding controls on wood mobilization and transport drawn from the liter-ature review

Hypotheses and associated assumptions

Flow characteristics-The duration of flows near bankfull exerts the greatest influence on woodtransport

bull transport capacity is likely maximized near bankfullbull transport distances may be longer for flows with longer duration of dischargenear bankfull

-The frequency of wood redistribution within the active channel isscale-dependent such that wood is redistributed more frequently with increasingriver sizeWood characteristics-Diameter is a strong predictor variable for wood mobility in reaches with highdepth variability during wood-transporting flows-Diameter becomes increasingly significant with greater proportion of flow depthsnear half the diameter of wood pieces-In small to medium rivers orientation and position of wood do not stronglyinfluence travel distance

bull orientation and position are likely to be overshadowed by the influence of floodmagnitude and reach-scale roughness elements that can deflect pieces

-Wood that spends a greater portion of time oriented parallel to flow in small tomedium rivers will spend more time in transport

bull wood moving parallel to flow is less likely to become lodged or braced en routeTree type is the most important global predictor variable for how wood movesthrough drainage networks

bull tree type significantly influences piece size shape and wood density-For a particular tree species increasing density correlates positively withresidence time and inversely with travel distance-Wood pieces recruited from tree species that maintain a more complex branchedshape during river transport have lower mobility than pieces that weather tocylindrical shapeReach characteristics-Reach retention capacity is smallest within the range of discharges above a woodmobilization threshold and below bankfull stage (retention capacity is greatest forvery low flows and overbank flows)-Log jam spacing only limits wood piece transport distance during low flows withlimited wood mobility-During infrequent high flows when most wood is transported the frequency andlength of inundation during which living vegetation obstructs flow is a betterpredictor of wood mobility than the downstream spacing of log jams-Entrapment of wood by living vegetation may explain why wood flux peaksbefore water flux

bull live wood can snare dead wood on the rising limb of the hydrograph beforedeposition at flood recession due to decreasing flow depths



22 Wood characteristics

Five characteristics of wood pieces exert an important influence onwood mobilization and transport anchoring length diameter orienta-tion and tree type (species which governs decay rates abrasion inten-sity density and branching complexity) Table 3 summarizes how thesecharacteristics are linked to wood mobilization in the literature Al-though burial is widely recognized as the most important anchoringmechanism (Berg et al 1998 Wohl and Goode 2008 Merten et al2011) and rootwads are often noted as substantially less mobile thanpieces without rootwads (eg Braudrick and Grant 2000 Bocchiola etal 2006b Daniels 2006 Cadol and Wohl 2010 Welber et al 2013Davidson et al 2015 Iroumeacute et al 2015) anchoring thresholds arenot yet well understood Information is very limited for example onthe amount of burial needed to effectively keep pieces stationary forfloods of variable magnitude and the stability of different types ofrootwads Surprisingly Davidson et al (2015) recently found thatsmall pieces with simplified square rootwads in flumes were actuallymore stable than longer pieces with rootwads We suggest further



exploring the relative stabilizing effect and likelihood of subsequent jamformation for different types of rootwads from different species withvarying levels of complexity and varying bole lengths Rootwads arecommonly used in stream restoration projects (eg Shields et al2004) Thus managers could use information on the minimum degreeof root complexity and length required tomeet stability criteria becauseof preference for smaller less complex pieces that can more easily passbuilt structures if mobilized Detailed field measurements of rootwadscoupled with three-dimensional printing technology could be used tomore accurately simulate wood characteristics in future flume experi-ments to test anchoring processes

Case studies conducted on small to large rivers have found that lon-ger pieces are either less likely to be moved during floods or are mobi-lized at higher discharges (Jacobson et al 1999 King et al 2013) Athreshold for piece mobility exists based on the ratio of piece length tobankfull width with logs less than bankfull width substantially moremobile than logs greater than bankfull width (Lassettre and Kondolf2003 Shields et al 2006 Warren and Kraft 2008 Wohl and Goode2008 Merten et al 2011 MacVicar and Pieacutegay 2012 Dixon and Sear2014) Longer pieces are more likely to be braced or buried againstchannelmargins formkey pieces in jams and be anchored by rootwadsall of which give the longer pieces greater stability (Merten et al 2011)However much variability exists For example wood shorter thanbankfull width does not always move when larger pieces become mo-bile (Warren and Kraft 2008) and tracking studies of loose wood inlarge rivers have noted no relationship between size of wood andwhether a piece is mobile during a flood (Schenk et al 2014) or thetiming ofmobilization (Ravazzolo et al 2015b) This lack of relationshipbetweenwood size andmobilization in large rivers ismatched byobser-vations in theflume that length does not influence thresholdmovementfor logs shorter than bankfull width (Braudrick and Grant 2000)

Case studies on medium rivers show that wood pieces shorter thanbankfull width travel greater distances than longer pieces (egLienkaemper and Swanson 1987 Berg et al 1998 Daniels 2006Millington and Sear 2007 Warren and Kraft 2008 Iroumeacute et al2010 Dixon and Sear 2014) However transport distances are highlyvariable and the longest logs are capable of travelling the farthest dis-tance despite their size (Dixon and Sear 2014) Also this relationshipbetween piece length and travel distance does not consistently holdtrue for large rivers Even though bigger pieces are less likely to bemoved by floods (Jacobson et al 1999) once mobilized no consistentrelation between transport distance and piece size is common(Jacobson et al 1999 MacVicar and Pieacutegay 2012 Schenk et al 2014Ravazzolo et al 2015b) One flume study simulating a large rivershowed that mobilization and travel distance actually increase withlength (Davidson et al 2015) Another flume study showed that traveldistance peaks withmedium sized logs that are long enough to steer offobstacles and have enough momentum to scrape past shallow deposi-tional zones but are short enough not to be deflected by banks(Welber et al 2013) Trends of decreasing transport with larger piecesize are present in single-thread rivers but the effects of log size on trav-el distance are more complicated in braided and multi thread channelsof similar size (Welber et al 2013 Ruiz-Villanueva et al 2015c)

Existing observations leave several questions unanswered If largerlogs are mobilized later and are trapped sooner than shorter logs(MacVicar and Pieacutegay 2012) why can they have equal or longer trans-port distances Is this solely because they travel at faster velocities orcould this be a result of sampling bias because it is easier to tag trackand identify larger pieces Does a threshold piece size for equal down-stream mobility exist Does this threshold vary by stream and woodtype How does piece length impact the effectiveness of hydraulicsteering to avoid obstacles and remain in swift current

Diameter ismore easilymeasured in thefield thanwooddensity andthus is often used in field studies of mobilization as a surrogate for flo-tation depth (ie buoyant depth) whichflume studies indicate is an im-portant variable in predicting wood mobilization and entrapment


(Braudrick and Grant 2000) For single pieces diameter controls thetiming and elevation of deposition on the falling limb of floods especial-ly for braided reaches or floodplains where the inundated depth altersrapidly with small changes in river stage (Bertoldi et al 2013) Asdepths decrease larger diameter pieces are slowed and stopped earlieron the receding limb (Bertoldi et al 2013)

Although diameter can be the most important variable for mobilityin braided channels (Welber et al 2013) for single-thread channelswood diameter appears to be a good predictor of mobilization only inthe absence of other more primary predictor variables such as anchor-ing and length (Haga et al 2002) Field observations of natural wood inmedium sized mountain and meandering channels have found no sig-nificant relationship (Iroumeacute et al 2010) or weak predictive power(Dixon and Sear 2014) between log diameter and distance travelledThus although diameter appears to be important as a secondary thresh-old condition for initial mobilization its influence on travel distance inthese channels may be limited to conditions in which greater propor-tions of flow depth are similar to log diameter such as during lowflows and through shallow braided or unconfined reaches

A flume study of braidedmorphology has shown thatwoodmobilitydecreases at threshold flows equal to half the diameter of logs (Welberet al 2013) This threshold is also supported by work in headwater riv-ers of Chile which showed that the pieces moved in average floods haddiameters less than half bankfull depth (Iroumeacute et al 2015) Thus wehypothesize that diameter is a stronger predictor variable for woodmo-bility in reaches with high depth variability during wood-transportingflows and that diameter becomes increasingly significant with greaterproportion of flow depths near half the diameter Although half-diame-ter appears to be equivalent to the flotation depth of wood recent ex-periments examining relationships between water saturation densityand buoyancy for instream wood suggest that most logs actually floatmore than halfway underwater (Ruiz-Villanueva et al 2016a) Thusthe half-diameter threshold for wood mobility identified in the fieldandflume (Welber et al 2013 Iroumeacute et al 2015)may bemore closelyrelated to conditions when flow depths are slightly less than flotationdepths

Variability in results relatingwood size tomobility likely result from(i) Equal mobility conditions when small pieces are shielded frommovement by larger instream structures (such as boulders and largerlogs or logjams) This scenario is similar to the concept of armouringin gravel- and cobble-bed streams (Parker and Toro-Escobar 2002)(ii) Local anchoring or bracing of pieces in natural channels exertinggreater controls on wood mobilization than wood size (MacVicar andPieacutegay 2012) (iii) Smaller pieces preferentially transported earlier onthe rising limb of floods because they are transported longer duringthe falling limb of prior floods and stranded at lower bank heights(MacVicar and Pieacutegay 2012) (iv) Flashy hydrographs mobilizingwood in short time intervals potentially reducing any significant differ-ences in initial mobilization between large and small pieces (Ravazzoloet al 2015b) (v) Variability in piece shapes and densities (Merten et al2013) (vii) Local and reach scale variability in channel retentivenessand characteristics (Wyżga and Zawiejska 2005 Wohl and Cadol2011)

Flume studies indicate that pieces perpendicular to flow are less sta-ble than those oriented parallel to flow (Bocchiola et al 2006a) but thisrelationship is unclear in the field (Iroumeacute et al 2015) In natural chan-nels individual loose pieces of wood are commonly deposited alongchannelmargins parallel to flow whereaswood oriented perpendicularis usually anchored or braced This complicates fieldmobility studies onlog orientation If flow is sufficient to float most loose unjammed woodon the channel margins a greater amount of wood oriented parallelmay be floated simply because there are greater amounts of loose easilytransported pieces with this orientation

The orientation of pieces greatly impacts drag force which impactswhen a piece is mobilized and its downstream velocity (Gippel1995) Thus orientation should play a role in total travel distance of a



piece of wood during floods However the influence of floating woodpiece orientation on travel distance is not well characterized in thefield Video monitoring of a large river found that uncongested woodtransport is oblique to flow commonly zig-zagging across the channel(MacVicar et al 2009) This is in disagreement with flume studiesthat show uncongested transport parallel to flow (Braudrick andGrant 2001 Welber et al 2013) Flume studies typically use smoothdowels whereas wood pieces in natural channels have irregularitiesin shape that could work as passive rudders making parallel travelless likely in natural settings

We could not find any studies that investigated travel orientation insmall tomedium channelsWe hypothesize that for these channels ori-entation and position ofwood in the river are not dominant explanatoryvariables for travel distance because they are likely to be overshadowedby the influence of flood magnitude and reach-scale roughness ele-ments that can deflect pieces However if orientation does play a rolein these streams we hypothesize that wood that spends a longer por-tion of time oriented parallel to flow will travel longer because thewood is less likely to become lodged or braced en route This may beone reason why shorter wood travels longer distances shorter woodpieces are generally more similar to cylindrical rods and their smallsize allows them to spend more time travelling parallel to flow andless time being deflected by instream roughness elements

Species and level of decay dictate how prone wood pieces are tobreaking en route and can impact travel distances by changing thesize and shape of pieces (Lassettre and Kondolf 2003 Wohl andGoode 2008 MacVicar and Pieacutegay 2012 Merten et al 2013) Prelimi-naryfindings from the few studies to explicitly look at influences of den-sity tree species and shape on travel distance suggest that tree type isone of the best predictors of downstream transport distance (Mertenet al 2013 Schenk et al 2014 Ravazzolo et al 2015bRuiz-Villanueva et al 2016a) We hypothesize that tree type is themost important global predictor variable for how wood moves down-river and through drainage networks explaining variance associatedwith transport distance residence times travel paths and depositionalpatterns

Recent advances have been made in understanding howwood den-sity relates to water content and buoyancy (eg Ruiz-Villanueva et al2015a) and how variation in buoyancy among andwithin different spe-cies relates to travel distance (Ruiz-Villanueva et al 2016a) Howevermorework is needed onhowwater content of logs changes as they trav-el through drainage networks and how this relates to depositional pat-terns such as the probability of becoming buried which can lead tolonger residence times We hypothesize that for a particular tree spe-cies increasing density (from increasing water content) correlates pos-itively with residence time and negatively with travel distance Inrelation to wood type we also hypothesize that wood pieces recruitedfrom tree species thatmaintain amore complex branched shape duringriver transport have lower mobility than pieces that tend to approxi-mate cylinders

23 Reach characteristics

Although general patterns relate hydrology (see Section 21) andwood characteristics (see Section 22) to woodmobility these relationsare highly variable (Welber et al 2013 Iroumeacute et al 2015) and signif-icant differences in mobility and deposition exist among contrastingchannel types sizes and forest disturbance regimes (Young 1994Wyżga and Zawiejska 2005) Much of this variability is likely relatedto local channel and floodplain characteristics that promote reach-scale retention (Braudrick and Grant 2001 Wyżga and Zawiejska2005 Pettit et al 2006) because during floods wood is preferentiallydeposited and stored in hydraulically rougher (Ruiz-Villanueva et al2015b) unconfined (Dixon and Sear 2014 Luciacutea et al 2015) reaches

The degree of channel roughness is commonly described using rela-tionships between wood characteristics and channel morphology


(Braudrick and Grant 2001) abundance of dead wood and jams(Beckman and Wohl 2014) and presence of live-wood within theflow (Jacobson et al 1999) Forest history governs the recruitment ofnew wood (King et al 2013) and catastrophic events can drasticallychange wood loads channel morphology and distribution of livewood (Johnson et al 2000 Pettit et al 2006 Oswald and Wohl2008) Thus we have grouped reach-scale characteristics that governwood transport into five categories channel morphology wood abun-dance and jams live wood and forest disturbance history (Table 4)

Wood pieces not travelling in the thalweg long compared tobankfull width or thicker than flow depth are more likely to interactwith in-channel obstructions channel margins and the floodplain lim-iting mobility and transport distances (Braudrick et al 1997) The pro-portion of wood carried in the thalweg versus near-bank regions fordifferent levels of flow and diverse river types is unknown Preliminarystudies suggest that in large and great rivers wood is predominantlycarried in the thalweg for flows less than bankfull (MacVicar et al2009 MacVicar and Pieacutegay 2012 Kramer and Wohl 2014) Howeveras water levels reach overbank flows wood is quickly routed to andtrapped on the floodplain (MacVicar and Pieacutegay 2012) For flows thatexceed bar heights on gravel-bed braided rivers GPS tracking of logpaths during transport showed that most logs are transported abovebars rather than travelling in the thalweg and as flows begin to recedewood is quickly deposited (Ravazzolo et al 2015b)

Reach characteristics contributing towood retention change asfloodmagnitude increases (MacVicar and Pieacutegay 2012 Ruiz-Villanueva et al2015b) because deposition is dominantly controlled by the availabilityand type of trapping sites at peak flow (Millington and Sear 2007MacVicar and Pieacutegay 2012) For steep confined small- to medium-sized mountain channels retentiveness is highest at lower flows anddecreases as flood magnitude increases (Wohl and Goode 2008 Wohland Cadol 2011) However floods with peak flows greater thanbankfull will be more effective at trapping wood than floods belowbankfull because wood is more likely to interact with roughness ele-ments such as riparian vegetation and shallow depths in overbankareas (Wohl et al 2011 MacVicar and Pieacutegay 2012) For any streamreach we hypothesize that reach retention capacity is smallest withinthe range of discharges above wood mobilization thresholds andunder bankfull and retention capacity is greatest for very low andoverbank flows

Retention rates of reaches have been estimated using exponentialdecay models of transport distances for experimentally introducedleaves (Larrantildeaga et al 2003) and small dowels le 106 m in lengthand le0035 m in diameter (eg Millington and Sear 2007) Whetherthese models can be scaled up to model large wood retention has yetto be tested but emergent properties such as congested transport(Braudrick et al 1997) are unlikely to be adequately represented in ex-perimental additions of small materials

Wood transport dynamics on floodplains are less understood thantransport in channels particularly with respect to the timing and trans-fer of wood between channels and floodplains Limited studies suggestthat despite having higher initial mobility floodplain wood does notmove very far and floodplains are a net sink for wood that becomes animportant part of the floodplain ecosystem (Benke andWallace 1990)

Large wood channel-spanning jams and wood rafts can be very ef-ficacious at trapping wood and when present are commonly the domi-nant entrapment site for fluvially transported wood (Pettit et al 2006Millington and Sear 2007Warren andKraft 2008 Beckman andWohl2014 Dixon and Sear 2014Wohl 2014a Boivin et al 2015 Iroumeacute etal 2015 Jackson and Wohl 2015) and smaller coarse particulate mat-ter (Jochner et al 2015) Higher amounts of wood in storage especiallyas jams have been related to higher amounts of wood recruited duringfloods (Johnson et al 2000) lower wood export rates (Bertoldi et al2014 Davidson et al 2015) and longer residence times (Wohl andGoode 2008) However several studies highlight that jams are fre-quently mobilized often reforming in the same locations with new



pieces (Pieacutegay and Gurnell 1997 Gurnell et al 2002Wohl and Goode2008 Curran 2010 Sear et al 2010 Collins et al 2012 Dixon and Sear2014) Pieces that are trapped within and behind jams can be releasedand replaced by other pieces during high flows so that the overall archi-tecture of the jam appears similar even though the internal pieces aredifferent (Lienkaemper and Swanson 1987 Dixon and Sear 2014) Di-rect observations indicate that entire logjams can be floated up and thenset back down during high flows (S Gregory pers comm July 2015) orcan break apart as individual basal wood pieces break or are dislodged(Wohl and Goode 2008) Although factors such as proportion of chan-nel cross-sectional area obstructed by the jam potential for overbankflow and dissipation of hydraulic force (Wohl 2011) porosity of thejam and cause of jam formation likely influence the relative stabilityof individual jams little is known of the relative importance of diverseprocesses by which jams become mobile

Because the storage frequency of large pieces of in-channel woodhas a negative relationship to piece mobilization (Merten et al 2011Wohl and Beckman 2014 Davidson et al 2015) channels with morestored wood and greater numbers of jams have so far been assumedto have lower overall woodmobility and to transport wood shorter dis-tances (Warren andKraft 2008) However destruction and reformationof jams in the same location may create the impression that a reach istransport limited because of a high density of wood in storage whenin fact the site favors repeated formation of jams Thus field observa-tionsmade during low flow of changes in wood storagemay not actual-ly reflect howmuchwoodmoved during a flood because overall storagevalues within a reach can remain the same despite high wood flux(Lienkaemper and Swanson 1987 Keim et al 2000 Wohl andGoode 2008 Dixon and Sear 2014) And although jams and large keypieces of wood form obstacles they do not impede the ability of somepieces to flow over around or through the jams during high flows(Lassettre and Kondolf 2003 Millington and Sear 2007 Warren andKraft 2008 Dixon and Sear 2014 Schenk et al 2014) Several ques-tions remain such aswhat storage density of wood is needed to greatlyimpact overall wood mobility within a reach what is the impact of re-moving wood from streams on the downstream transport of otherpieces of wood and can turnover of wood be estimated based on stor-age density

The proportion of the channel cross-sectional area obstructed by ajam as well as the porosity of the jam likely influences the ability of in-dividual wood pieces to move downstream past the jam Jams can alsocause avulsion during high flow (Keller and Swanson 1979 OConnoret al 2003 Collins et al 2012 Phillips 2012) creating a transport by-pass chute forwoodWe suggest that the presence of jams does not nec-essarily cap wood transport distances at one to two jam spacingintervals as previously presumed (Warren and Kraft 2008) andmodelled (Martin and Benda 2001 Lassettre and Kondolf 2003) Wehypothesize that logjam spacingmay only limit transport distances dur-ing low flowswith limitedwoodmobility This is supported by observa-tions by (Lassettre and Kondolf 2003) who found that jams onlyimpede wood travel distance for floods with recurrence intervals b 6ndash15 years Relationships between transport distance and jam spacing re-main unclear for varying flow magnitudes and should be an avenue offuture research Complicating analysis of jam mobility and jam influ-ence on piece transport in the existing literature is that the word jamis used to refer to 2ndash3 pieces of wood in contact (eg Wohl and Cadol2011 Bertoldi et al 2013) as well as large channel-spanning struc-tures (eg Beckman and Wohl 2014) and commonly the criteria fordesignating a jam are not clearly stated

Of particular importance to trapping efficacy during major floods isthe presence of rooted living vegetation (live-wood after Opperman etal 2008) within the flow and on banks (Jacobson et al 1999 Pettit etal 2006 MacVicar and Pieacutegay 2012 Bertoldi et al 2015 Ravazzoloet al 2015b) Living vegetation is particularly effective at trapping mo-bile wood in overbank areas (Wohl et al 2011) and on mid-channelbars (Jacobson et al 1999) We hypothesize that during infrequent


high flows when most wood is transported the frequency and lengthof inundation during which living vegetation obstructs flow is a betterpredictor of wood transport distance than the downstream spacing oflogjams Furthermore the abundance and distribution of specific livewood species that are singularly effective trappers (Jacobson et al1999) may account for most of the variance If this is the case then alarge river with flow paths that route wood over vegetated bars andfloodplains and through overbank channels may bemore transport lim-ited thanmedium channels with a high density of logjams even thoughchannel widths on the large river are greater than log lengths

We also hypothesize that entrapment of wood by living vegetationmay be the reason why wood flux peaks before water flux (MacVicarand Pieacutegay 2012 Ruiz-Villanueva et al 2016b) because live woodcan snare dead wood on the rising limb of the hydrograph before depo-sition during flood recession due to decreasing flow depths If this istrue then the lag between the wood flux peak and the discharge peakcould be used as a measure of overall entrapment efficiency of thatchannel for a particular flood Episodic reorganization of wood andmass deposition of wood from exceptional flooding can have lasting ef-fects on channel morphology and overbanks and greatly impact theavailability mobility and transport paths of wood (Oswald and Wohl2008)

Fluvially deposited wood piles create heterogeneity in riparian veg-etation patterns by providing nursery sites for new live-wood and bycreating hard points protecting new growth from rapid erosion ofbanks and bars at lower flows (Hickin 1984 Collins et al 2012Gurnell et al 2015) Even where input and transfer of fluvial woodare rare and wood piles rapidly decay or burn their influence on vege-tation germination can be substantial (Pettit et al 2006) Thus floodingdisturbance history can have a feedback loop governing forest structureheterogeneity and age which in turn influence wood loading channelcomplexity and wood transport Flume studies show a threshold re-sponse in patterns of wood storage (Bertoldi et al 2014) and woodtransport to input rates of wood (Braudrick et al 1997) These resultssuggest the possibility of using estimates of wood input rates from epi-sodic events such as floods or hillslope failures to predict different al-ternate stable states of reach-scale wood transport (congested versusuncongested) and wood storage patterns (single piece versus jamdominated)

Although channels draining old-growth forests tend to have lowerwood mobility than younger forests attributed to higher instreamwood loads (Gurnell 2003 Iroumeacute et al 2010 Collins et al 2012Beckman andWohl 2014 Jackson andWohl 2015) some field studiesof transport distances in relation to forest history yield mixed resultswith Berg et al (1998) documenting greater transport distances in anold-growth watershed compared to a logged watershed and Young(1994) and King et al (2013) documenting greater export in burnedwatersheds compared to unburned watersheds Differences betweendisturbance types and subsequent impact on transport dynamics inchannels likely influence these contrasting findings

Legacy impacts from past land use also influence contemporarywood transport Studies from forest ecology indicate that rates ofwood recruitment vary on time frames from decades to centuries(Murphy and Koski 1989 Wohl 2016) Two centuries or more arecommonly required to reach old-growth forest conditions and peakwood loads occur decades after catastrophic forest mortality eventssuch as pine beetle outbreaks (Bragg 2000) Consequently observedwood transport could reflect past rather than current conditions of thewatershed (Wohl and Cadol 2011) but the effect on interpretation ofwood transport processes remains unclear

3 Quantitative summary

Field researchers havemade advances in understandingwood trans-fer and transport through natural drainage networks during the last40 years but this has been on a case-by-case basis and generalizations


Table 6Summary of field studies by location used in the quantitative analysis of wood mobilityTd are the number of downstream travel distances and Mi are the number of initial mo-bility measurements used from each study at each location

Location River Td Mi Study


have not yet been drawn from the data In this section we focus on an-alyzing and synthesizingfieldmeasurements ofwood transport in orderto highlight general patterns of wood mobility across diverse sizes andtypes of rivers Appendix A provides a summary of field-derived empir-ical equations

North AmericaCalifornia Central Sierra Streams 2 23 Berg et al (1998)California Soquel Creek ndash 2 Lassettre and Kondolf

(2003)California Sacramento River 6 6 MacVicar et al (2009)Colorado 5 Rocky Mountain

Streams-Coloradondash 48 Wohl and Goode (2008)

Georgia Oggeechie River ndash 7 Benke and Wallace(1990)

Illinois Poplar Creek 1 1 Daniels (2006)Illinois Upper Mississippi Missouri

Ohio Riversndash 6 Angradi et al (2010)

Minnesota Streams draining into LakeSuperior

ndash 1 Merten et al (2010)


1 ndash Merten et al (2013)

Mississippi Little Topshaw Creek ndash 4 Shields et al (2008)New York Rocky Branch 1 3 Warren and Kraft (2008)NorthCarolina

Lower Roanoke 2 1 Schenk et al (2014)

NorthDakota

Upper Missouri ndash 1 Angradi et al (2004)

Oregon Bark Buttermilk and HudsonCreeks

ndash 8 Keim et al (2000)

Texas San Antonio River ndash 1 Curran (2010)Washington Salmon Creek ndash 4 Bilby (1984)Washington Various ndash 1 Grette (1985)Washington Queets River 1 ndash Latterell and Naiman

(2007)Washington Streams in Lookout Basin 1 5 Lienkaemper and

Swanson (1987)Wyoming Crows and Jones Creek 2 4 Young (1994)

Central and South AmericaCosta Rica 6 Costa Rican Streams ndash 12 Cadol and Wohl (2010)Chile Vuelta de la Zorra 1 1 Iroumeacute et al (2010)Chile Vuelta de la Zorra ndash 1 Ulloa et al (2011)Chile Buena Esperanza and Tres

Arroyos3 4 Mao et al (2013)

Chile 4 Chilean Mountain Streams ndash 18 Iroumeacute et al (2015)Chile Blanco River ndash 4 Ulloa et al (2015)

EuropeBasque 12 Iberian Streams ndash 19 Diez et al (2001)Basque Agueara Basin ndash 4 Elosegi et al (1999)England Highland Water 1 4 Dixon and Sear (2014)England Main Highland Water ndash 1 Pieacutegay and Gurnell

(1997)England Upper and Main Highland

Waterndash 2 Gurnell (2003)

England Upper Highland Water ndash 9 Millington and Sear(2007)

France Ain River ndash 11 MacVicar et al (2009)France Ain River ndash 4 MacVicar and Pieacutegay

(2012)Italy Tagliamento River ndash 3 Bertoldi et al (2013)Italy Tagliamento River 1 ndash Ravazzolo et al (2015a)Italy Tagliamento River ndash 8 van der Nat et al (2003)Italy Piave River 1 1 Pecorari (2008)Switzerland Erlenbach River 1 ndash Jochner et al (2015)

Africa Asia and AustraliaNamibia Kuisab River 1 2 Jacobson et al (1999)Japan Oyabu Creek 4 4 Haga et al (2002)Australia Daly and Katherine Rivers ndash 1 Pettit et al (2013)

31 Methods

We conducted a literature search for field studies that measured themobility of wood via change in storage in a reach (Eulerian) and studiesthat measured the travel distance of wood pieces (Lagrangian) Hence-forth we use the term initial mobility to refer to observations of changein storage downstream mobility to refer to data that record travel dis-tance andmobility to refer generally to both categories In the literaturethe termmobility is used interchangeably when referring to initial mo-bility of wood in storage and downstream transport mobility of floatingwood This ambiguity can create confusionWedistinguish between ini-tial and downstream transport mobility because factors that influenceinitial mobilization may not be as important for transport of the samepiece downstream

We only included studies with direct mobility measurements in thefield We did not include studies that back-calculated mobility fromwood budgets or recruitment measurements or results from numericalmodels We located 40 studies with such measurements in a broadrange of peer-reviewed publications (17 journals) as well as one tech-nical report and two theses About 57 of studies came from physicalscience publications while about 43 came from ecology- and forest-ry-related publications The largest proportion of studies from a singlejournalwas 20 inGeomorphology followed by 10 in Earth Surface Pro-cesses and Landforms and in River Resources Research Table 6 summa-rizes the studies by continent and by number of data valuescontributed from each study We did not conduct a full quantitativeanalysis on total wood flux because this information was very limitedmost flux values were back-calculated rather than directly measuredand reconciling values from different studies proved too difficult be-cause of different reporting metrics Wohl (2016 this issue) summa-rizes some wood flux values but a more thorough compilation ofwood budget fluxes would be useful

Wood mobility is typically monitored via tagging pieces or remotesensing of large accumulations along a reach (Δx) and then notinghow much and how far this wood moved within a given time frame(Δt) typically one year Consolidatingdata frommanydisparate sourceson wood mobility proved difficult Most studies do not report the samevalues because study questions differ For example some studies inves-tigated jammobility whereas others looked at mobility of loose piecesSome studies analyzed only wood that was exported whereas othersconsidered only newly recruited wood

From each initial mobility study we recorded what was measured(typically pieces or volume of wood) as well as the following catego-ries (if reported) the starting amount in storage (start) the amountthat left the reach (left) the amount that entered the reach (came)the amount that stayed (stayed) and the final value in storage (end)When possible the stayed variable was subdivided into an amountthat stayed but moved within the reach (repositioned) and the amountthat was immobile (immobile) For comparison with the widely usedwood budget equation (Benda and Sias 2003) came is Qi andor Li leftis Qo andor Lo Start end and repositioned are all different forms of ScWhen studies did not differentiate between pieces that wererepositioned locally versus pieces that were exported downstream weassumed that the pieces were exported downstreamWhen a study dif-ferentiated between repositioned pieces and exported pieces the au-thors distinguished two populations of mobile wood one that movedminimal distances and was locally retained and one that was exportedlonger distances downstream In most cases we were able to back-cal-culate categories not supplied Sometimes we calculated values based


on a starting value and a reported mobilized If we were unable toback-calculate based on given information the category was left empty

In addition to these values we recorded potential explanatory vari-ables stream width (w) recurrence interval of largest event in Δt (RI)fraction bankfull of largest event inΔt (Fb) maximum log length in stor-age (Lmax) study time frame (Δt) and reach length (Δx) We chose not



to focus on diameter because the range of flows and channel typeswould hide any relationships related to diameter Also diameter wasnot as consistently reported as length When available we used report-ed bankfull widths otherwise we used reported width reported meanormedianwidth or used the center of a given rangeWewere not over-ly concerned about the mix of types of width values because the errorintroduced from this is much smaller than the range of stream widths(15 mndash927 m) When possible we used the fraction bankfull reportedfor each study If this value was not reported we calculated the valuebased on reported discharge values for floods and bankfull or we esti-mated the value based on text descriptions assigning a value of 05 forlow flows 1 for near bankfull flows 15 for overbank flows and 2 forvery high flows In some cases too little informationwas given to deter-mine fraction bankfull and this category was left blank We did not at-tempt to estimate recurrence intervals and only used values for thiscategory when reported The final data set included 36 studies with229 data entries (Supplemental data file initialmobilitycsv) Channelwidths (w) ranged from 15 to 927 m (median (M) = 72 n = 223)RI from 029 to 50 yr (M = 15 n = 44) Fb from 04 to 54 (M = 1n = 169) Lmax from 1 to 60 m (M = 20 n = 229) Δt from 001 to13 years (M = 1 n = 211) and Δx from 004 to 240 km (M= 100 mn = 193)

We computed dimensionless wood lengths as Lmaxw (henceforthL) The L values ranged from 002 to 11 (M = 3 n = 223) Based ondata availability we simply used the maximum length of wood foundin storage for the river to calculate L When maximum log lengthswere not reported we used maximum log lengths from other woodstudies on the same river or from similar regions We chose not to useaverage length of wood because the size of the largest pieces reflectskey pieces for jams and accumulations that moderate wood mobilityAnother option would have been to use riparian tree heights butthese are commonly larger than wood found in the stream becausemost trunks break during the recruitment process Maximum woodlength might not be the best metric to calculate L because maximumlength can be highly variable depending on howmany pieces are mea-sured No studies have thoroughly investigated how L should be calcu-lated so uncertainty remains regarding which log lengths should beused for comparison with stream width to maximize predictive poweron wood mobility Should maximum log length in storage be used orL90 or L85 (analogous to D90 or D85 in sediment research) We recom-mend that future studies more fully characterize the distribution oflog lengths in storage Additional reporting of other size fractions suchas L90 and L85 might provide better metrics than Lmax to use in calcula-tions of L Also channel width is generally assumed to be width atbankfull but there is no reason to limit L to just onewidth and thismet-ric could be calculated at various stages of flooding

In order to compare studies we calculated the net change in storage((end-start)start) redeposition rate ( end value that came) remobiliza-tion rate ( start value that left) stability rate ( of start that stayed orwasrepositioned) reposition rate ( start value that was repositioned) and totalmobile pieces (100 lowast (came+ left+ repositioned) (start+ came))in Δt We suggest that total mobile pieces as defined here is the bestoverall metric for mobility because this includes imported exportedand locally transported wood as a fraction of all the wood that movedthrough storage during Δt In many studies it is unclear whether mo-bilized pieces were locally mobile or left the reach because no differ-entiation is made between repositioned pieces and downstreammobile pieces The most commonly reported mobility variable wasthe remobilization rate because most studies were focused on ex-ploring factors that influenced the relative stability of wood alreadyin storage not fluvial redeposition of previously transported ornewly recruited wood

We analyze transport distance by stream size using all reportedvalues for distance travelled minima maxima means and medians(Supplemental data file distancecsv) We used all reported values inorder to assess range of variability within travel and because travel


distance from field data is limited Although we would have liked tocompare transport distance to transportedwood lengths and diametersmany studies did not report transported log dimensions separately fromstored wood dimensions Thus we were unable to extract this informa-tion for enough studies to achieve meaningful results

32 Results

To obtain an overall sense for initial mobility during varying flowswe plotted mobility values against fraction bankfull and against recur-rence interval (RI) of the highest flow within the study period (Fig 3)Most of the data are from flows under twice bankfull Of the studiesthat reported RI most were ge1 year and b10 years There is a strongstepped threshold for maximum mobility at bankfull discharge or theyearly recurrence interval Below bankfull or for minor floods withless than yearly RI the maximum possible mobility of stored wood is~30 For bankfull or higher discharges maximum possible mobilityjumps to ~80 These thresholds bound a mobility envelope for storedwood The outliers not contained within this envelope include mobilitymeasurements after large morphological changes (van der Nat et al2003) or measurements of mobilization of newly recruited wood onan actively erodingmeander bend (MacVicar et al 2009) and in a braid-ed river (Bertoldi et al 2013)

The 30 percent low flow maximum mobility threshold (Fig 3) isslightly lower than typical values for of wood stored as individualpieces and of wood in the regularly flooded channel (Table 7) Thisis expected because some of the individual pieces are probably abovebankfull and some of the pieces below bankfull are probably within sta-bilizing jams or anchored Comparing loose unattached single pieces atvarying stage heights to low flow mobility rates would be useful If aconsistent relationship holds it may be possible to determine yearlylow flow background wood flux based on wood elevations akin tobase flow on hydrographs

To relate mobility to change in storage (ΔS) we plotted the totalmobility in relation to net change in storage (Fig 4) Our compileddata show that a net change in storage of zero is associated with thewidest range in mobility from zero to ~80ndash90 A definable lowerthreshold for total mobility based on net change in storage is appar-ent in Fig 4 This lower bound resembles a funnel that meets at a pointand is nearly symmetrical for negative and positive net change for lowerrates of mobility This means that on average rivers are in equilibriumwith wood inputs equal to outputs This observation is also supportedby Fig 3 which shows little difference in the ranges of mobility ofwood that came or went However mobility between 30 and 80 ismore commonly related to larger net gains in storage than net loss(Fig 4) This probably reflects the influence that high flows have on re-cruitment of new wood All mobility rates under maximum thresholdsare possible for all flows (Fig 3) The largest flows do not always trans-port the most wood from a reach and can even have zero mobility be-cause mobility depends not only on absolute flow magnitude but alsoon deposition patterns set by the sequence of prior high flows (egHaga et al 2002) and reach-scale biogeomorphic characteristics Lowreach-scale remobilization rates can occur alongside high reach-scaleredeposition rates or vice versa resulting in increasing or decreasingtrends in total reach wood storage Whether a reach has net increasingor net decreasing trends in stored wood impacts wood export yields forfuture floods The asymmetry toward net accumulation of wood in stor-age for events with N30 mobility (Fig 4) could reflect larger scale pat-terns of global wood accumulation in rivers as a result of afforestationeffects of disturbances such as fire on wood loads or decreasing fre-quency of high peak flows from flow regulation

To understand mobility across stream sizes we plotted percent ini-tial mobility and transport distance against channel width and L(Lmaxw) (Fig 5) The explanatory variables w and L were plotted on alog scale as well as untransformed to better display the results fromthe 900-m range in channel widths Maximum potential woodmobility


mobility envelope~30 low flow

max mobility threshold

~80 high flow

max mobility

threshold

Bertoldi et

al(2013)

Van der Nat

et al(2007)MacVicar

and Pieacutegay

(2009)

mobility envelope

remobilized

redeposited

repositioned

Fraction bankfull

M

ob

ile

Recurrence interval (year)

Fig 3Mobility of wood in relation to flow LeftMobility related to fraction of bankfull from the highest discharge in the study period RightMobility related to highest recurrence intervalflood within study timeframes Dashed lines annotate possible mobility thresholds outlining a mobility envelope Data are provided in the supplemental file initialmobilitycsv


(drawn as the upper dashed threshold in Fig 5) increases as channelsbecome wider up to ~3 m at which point maximummobility stabilizesat ~80ndash90 up to ~20 m channel width then decreases until channelsare N100 m wide beyond which mobility likely stabilizes again al-though we lack the data to definitively show this

If L is used rather than absolute stream widths this relationship ofincreasing maximum mobility from small to medium channels(L b 5) constant maximum mobility for medium channels(5 ge L ge 1) and then decreasingmobility for large channels (L b 1) be-comes more clear with fewer outliers (Fig 5) Outliers not containedwithin the mobility envelope (Fig 5) are from a study of remobilizationof recently recruited wood from a volcanic eruption (Ulloa et al 2015)mobilization of wood from one actively eroding meander bend(MacVicar and Pieacutegay 2012) and remobilization of wood due to largemorphological changes in braided rivers (van der Nat et al 2003Bertoldi et al 2013)

The braided river data may be plotting outside mobility envelopesbecause of inappropriate values for streamwidth used in the calculationof L The estimated channel width carrying water was not reported fordifferent flows so we had to use the channel width of the entire valleybottom Channel widths change substantially with small changes instage along braided rivers As a result many of these outlier valuesmay actually plot closer to a medium-sized river for low flows largeriver for high flows and possibly a great river for extremely high

Table 7Case study values for percent pieces in storage that could be potentially mobile in lowerflows Ind is the percent of wood surveyed as individual pieces and ubf is the percentof wood in the regularly flooded channel between low flow and bankfull

River (reference) Ind ubf

East Fork (Berg et al 1998) 23 ndashEmpire (Berg et al 1998) b1 ndashLavezolla (Berg et al 1998) 27 ndashBadenaugh (Berg et al 1998) 2 ndashSagehen (Berg et al 1998) 25 ndashPauley (Berg et al 1998) 9 ndashVeulta de la Zorra (Ulloa et al 2011) 50 47Pichuacuten (Ulloa et al 2011) 20Tres Arroyos (Mao et al 2013) 88 ndashSacramento (MacVicar et al 2009) 66 40Kuisab River (Jacobson et al 1999) 54 ndashLower Roanoke (Schenk et al 2014) 50 44Kochino-tani Creek (Haga et al 2002) 66 ndashCrows Creek (Young 1994) 25 55Jones Creek (Young 1994) 37 41Upper Missouri (Angradi et al 2004) 39 ndashUpper Missouri (Angradi et al 2010) 86 30Mean 43 40St dev 27 21


flows Also diameter is an important mobility variable in braided sys-tems where depth and width change rapidly with small fluctuationsin flow (Welber et al 2013) Thus similar graphics as in Figs 3 and 5that replace L with a dimensionless log diameter D (Diamflowdepth) may highlight mobility patterns in braided rivers and on flood-plains that we were unable to capture in this analysis

Transport distances plotted against channel width and L fall intotwo main groups which are depicted as two separate shaded boxeson Fig 5 corresponding to medium and large rivers A discontinuity ex-ists between 05 and 1 L when typical maximum log lengths are lessthan the channel width but greater than half the channel width Thisdiscontinuity occurs at the transition between medium and large rivers(between ~20 and 50 m channel width) The mean recovery rate fromstudies for trackedwoodwas ~50ndash70 and ranged as low as 0 for stud-ies that reported transport distance at reach length Even though not allwood was recovered and thus maxima are not true maxima transportdistances appear to be about two orders of magnitude higher for largerivers than for medium rivers The median travel distance (calculatedfrom all reported values max means and mins) for each group isdrawn as a solid black line on Fig 5 (bottom row) and is close to100 m for small-medium rivers and one to two orders of magnitudehigher for large rivers The one outlier is from experimental releaseand tracking of smaller pieces of wood in a cleaned mountain channelwith no obstructions (Haga et al 2002) Because of the small size ofwood pieces used in the experiment this study plots as a mediumriverwhen plotted by channel width but a large riverwhen consideringthe size of wood in the experiment in relation to channel width

-100 -50 0 50

6080

4020

0

100

~30 low flow

maximum mobility

threshold

MacVicar and Pieacutegay

(2009)

mobility

envelope

threshold

T

ota

l m

ob

ility

Net change in storage

~80 high flow

maximum mobility

Fig 4 Total mobility of wood in relation to percent net change in storage Dashed linedefines a mobility envelope marked outliers are discussed in the text Low and highflow thresholds are from Fig 3 Total mobility was calculated as the amount of woodthat was mobile (redeposited remobilized and repositioned) divided by the totalamount of wood in storage within the timeframe (original amount plus the amount thatwas newly deposited) Data are provided in the supplemental file initialmobilitycsv


Haga et al

(2002)

Millington and

Sear (2007)

MacVicar

et al (2009)

Ulloa et al (2015)

Jacobson et al (1999)

MacVicar

et al (2012)

Van der Nat

et al (2003)

Small

Medium

Large Large

amp Great

Bertoldi et

al(2013)

Van der Nat

et al (2003)

Bertoldi

et al(2013)

Ulloa et al

(2015)

MacVicar et al

(2009)

high mobility from

bull extreme events

bull recently recruited wood

bull large morphological

changes

mobility envelope for

previously recruited wood in

stable channels

mobility envelope

remobilized

redeposited

repositioned

M

ob

ile

M

ob

ile

Stream width (m)

Stream width (m)

Stream width (m)

Tra

ve

l d

ista

nce (m

)

L

L

L

Fig 5Mobility of wood in relation to stream size as width (left) and dimensionless length L (right) L is maximumwood length in storage divided by channel width Log scales are usedto better show patterns across the wide data range Top row presentation of initial mobility data dashed lines define mobility envelopes Middle row interpretation of initial mobilitydata Bottom presentation of downstream mobility data Solid lines mark the median transport value within each shaded grouping Marked outliers outside of mobility envelopes arediscussed in the text Data are provided in the supplemental files initialmobilitycsv and distancecsv


One analysis that wewould like to conduct is to compare flow (as RIor fraction bankfull) to the ratio Lmax of transportedwoodLmax of woodin storage Unfortunately few studies reported the length characteris-tics of transported wood and the length characteristics of wood in stor-age in a manner that facilitated comparison Or studies did not includevalues of flows that transported wood compared to either bankfull or tothe gage record We recommend that future studies of wood mobilityreport size characteristics of transported wood size characteristics ofthe nontransported wood and recurrence interval or proportion ofbankfull for flows

The transition betweenmedium and large rivers is usually placed atL= 1 Our plots of mobility against L (Fig 5) reveal that most obser-vations of mobilization have been conducted in medium-sized riverswith maximum log length between two and five times channel widthOur plots of mobility and transport distance suggest that the transitionbetween medium and large rivers occurs between L = 1 and L = 05Although maximum initial mobility may be maximized for mediumchannels (Fig 5) median mobility measurements increase with


increasing channel size (Fig 6) increasing from 9 in small channels(n = 23) to 14 in medium channels (n = 136) 32 in large channels(n = 45) and 80 in braided large rivers (although this sample size issmall n = 5)

Several field studies have empirically modelled probability of move-ment and downstream transport and we summarize their approachesand equations in Appendix A The probability of movement typically ismodelled from change in storage using logistic functions with woodcharacteristics as explanatory variables (Lassettre and Kondolf 2003Wohl and Goode 2008 Merten et al 2010) Variables typically includelength or length related to bankfull width categorical variables for an-choring or rootwads and diameter related to flow depth Numericalsimulations of wood flux based on physical modelling of a 10 yearflood on the Czarny Dunajec in Poland present transport ratios (theamount of wood exported divided by the amount imported) as linearfunctions of wood volume effective depth and dimensionless woodlength (L) as well as exponential functions of wood density separatelyfor single- and multithread reaches (Ruiz-Villanueva et al 2015c)


Lgt5 1leL le 5

Small Medium

Llt1

Large

Large

Braided

Med=9

Med=14

Re-m

obiliz

ation

ra

te (

)

Med=32

Med=80

Min=25

Max=95

Max=83Max=80

Max=33

n=45 n=5n=23 n=136

020

40

60

80

100

Fig 6 Remobilization rates by stream class Extent of bars is the range of the data Theshaded grey portion extends from the minimum to the median values


Nonlinear functions of transport ratios related to varying discharge arealso presented in which Ruiz-Villanueva et al (2015c) show that thetransport ratio response is different between more (RI b 10ndash15 years)and less frequent events Iroumeacute et al (2015) found linear relationshipsbetween percentmobility of stored pieces and flow characteristics suchas the unit stream power at maximum stage height and the ratio ofmaximum stage over bankfull stage

Transport distance has been modelled using exponential functionsto reach retention rates (Millington and Sear 2007) and water depthat peak flow (Haga et al 2002) Transport distance has also beenmodelled as a function of jam spacing (Martin and Benda 2001Lassettre and Kondolf 2003) although this may have limited applica-tion to low flow conditions when jams are completely obstructing thechannel and are not mobile

Although variablesmodelled in empirical field-derived equations forwood mobility (Appendix A) can be similar most of the studies listedmodelled different response variables (wood velocity percent mobileprobability of key piece mobility travel distance) because each studywas focused on different goals To facilitate comparison of mobility be-tween case studies and between rivers of differing size we suggestthat field scientists focus effort on empirical formulation of mobilityand travel distance As we have shown in Section 3 mobility can becalculated in many ways (eg start that left end that came) Wesuggest developing equations of mobilization that predict the ofstarting value that leave the reach or are repositioned within thereach because thiswill facilitate estimation ofwood flux frommeasuredstorage values However we highly recommend always reporting theraw volume or count values of changes in storage so that other re-searchers can recalculate mobility in other forms

4 Discussion

41 Mobile wood and management

The origin of instream wood research as a discipline can be tracedback to the Pacific Northwest (Oregon Washington and BritishColumbia) during themid-1970s to early 1990swhere government for-esters and fish ecologists became interested in the role of wood forimproving fish habitat Early research into wood transport was primar-ily motivated by the desire to understand the stability of wood instreams because investigators thought that the longevity of instreamwood influences the quality of fish habitat (Anderson and Sedell1979 Keller and Swanson 1979 Hogan 1984 Harmon et al 1986Bilby and Ward 1989 Nakamura and Swanson 1993)

But even the largest pieces ofwood and jams in natural channels canmove When engineered structures are anchored in mobile systemsrestoration projects are sometimes declared unsuccessful overtimeframes of 5+ years because most of the wood structures are


destroyed by large floods or bank erosion (eg Shields et al 2006)Engineered wood structures can be ecologically significant even ifthey do not stay in place (Choneacute and Biron 2015) In a compellingpaper on wood mobility and ecological function Daniels (2006) foundthat although wood in a low gradient meandering river was too mobileto havemore than a short-term impact on themorphology or hydraulicsof the channel wood did directly impact bed storage of organicmaterialin amanner that outlasted the residence of thewood andprovided valu-able ecosystem services to benthic communities Thus the mobility ofwood allows patchy deposition of nutrient-rich organic material tocover a greater spatial extent of the stream bed If wood is always an-chored in place then organic material deposition is limited to thoseareas near anchored wood

Although rehabilitationwith anchoredwoodmay be appropriate forreaches inwhichwood is naturally lessmobile anchoredwood is inher-ently flawed for reaches inwhich natural processes facilitate the regularand episodic transfer of wood over time periods shorter than those ofdesired beneficial ecological outcomes Wood naturally moves down-stream or laterally onto floodplains so the best and most cost effectivemanagement practice is to ensure that the stream has a supply of newwood preferably including some large pieces via upstream recruitmentfrom riparian forests and then allow the river to redistribute and re-peatedly mobilize the wood (Kail et al 2007)

If management strategies are adopted to increase unanchored woodloads there are concerns that the mobile wood will endanger instreamstructures and increase flood damage from clogging These concernshave merit because higher wood loads can facilitate more recruitmentduring floods (Johnson et al 2000) However field evidence indicatesthat increased amount of wood in storage along a reach does not trans-late to increased wood against bridges as presumed but instead maydecrease the hazard of wood clogging by increasing the likelihood thatthe pieceswill be trapped on log jams and complex channelmargins be-fore accumulating against structures (Lassettre and Kondolf 2003 Maoet al 2013) Most wood comes from pieces recruited from upslopelandslides and bank failures during a flood not from previouslytransported fluvial wood (Lassettre and Kondolf 2003 Luciacutea et al2015) Thus the common management practice of removing instreamwood may actually facilitate the downstream mobility of these newlyrecruited pieces rapidly delivering them to the first available obstruc-tion often bridges weirs or other built structures Rather than remov-ing wood Lassettre and Kondolf (2003) recommended that the mosteconomical option would be to replace culverts and bridges with struc-tures designed to pass wood typical for a stream

411 Storage patterns and transport processesStored wood characteristics have been used to develop wood bud-

gets to infer wood flux and transport distance (eg Martin and Benda2001) because the pattern of wood stored in river networks is assumedto reflect input and transport processes (Keller and Swanson 1979Davidson et al 2015) As we have shown and others have noted (egvan der Nat et al 2003) flows under bankfull generally transportb30 of the stored wood roughly corresponding to available loosewood positioned under bankfull stage (Fig 3 Table 7) However thelinkages between wood transport and wood storage are poorly definedand complex and thus the interpretation of wood transport onlythrough patterns in storage is limited Video monitoring of wood in ac-tive transport on the Ain River a large meandering river in France hasshown that estimates of wood export derived solely from storage andrecruitment characteristics may be underestimating actual wood fluxby two to ten times (MacVicar and Pieacutegay 2012)

The main shortcoming is that storage characteristics only provide asnapshot of conditions based on the time interval monitored limitinginferences regarding short- or long-term temporalfluctuations in trans-port from seasonal resurveys of storage Another problem is substantialvariability on the reach scale caused by increases or decreases in chan-nel retentiveness and trapping sites both longitudinally along a river



and with changing flow stage (Bertoldi et al 2013) For exampleIroumeacute et al (2010) found that for low order mountain headwaterchannels in Chile wood movement ( pieces mobilized) only occurredin a few reaches Also a large discontinuity in wood storage and trans-port processes exists between flows that access floodplains and thosethat do not (MacVicar and Pieacutegay 2012 Dixon and Sear 2014) Thuswhen estimating wood flux from field measurements results can behighly dependent on design and spatial and temporal extent ofsampling

The largest advances in linking transport processes to wood exportand storage characteristics have been made in flumes with simplifiedwood and mobile banks These flume experiments have expanded theunderstanding of linkages between input rates and storage regimes(Braudrick et al 1997 Bocchiola et al 2008 Bertoldi et al 2014) en-trapment processes related to the interaction of channel form withwood and flow characteristics (Braudrick and Grant 2001 Bocchiolaet al 2006a Welber et al 2013) and how live wood moderates andcontrols storage and export (Bertoldi et al 2015) Although these phys-ical models are useful they are simplifications of the complex interac-tions found in the field The combination of infrequent complex fieldobservations coupled with frequent yet simplified scaled measure-ments in flumes and models yields the biggest advances as in Ruiz-Villanueva et al (2016b)

42 Regular and episodic transfer of wood

Wood transfer is both regular and episodic Jochner et al (2015) de-scribed a four end-member conceptual model linking combinations ofcontinuous recruitment and export with episodic recruitment and ex-port Even in the Pacific Northwest which is known formore stable log-jams there are high rates of movement especially for the smallerfractions of large wood such that wood flux is a constant process(Keim et al 2000) Juxtaposed on this constant flux is large-scale epi-sodic flux of the biggest jams and largest pieces during floods withlong recurrence intervals

Wehave conceptualized regular versus episodicfluxofwood in stor-age in Fig 7 The stepped profile of wood storage in the diagram relatesto small episodic delivery and export of wood and the large steps repre-sent rarer episodic wood fluxes Small channels have large episodicwood flux but minimal to no yearly regular flux whereas larger riverhavemore frequent regular flux but smaller scale episodic flux attribut-able to the increased sites of deposition on floodplains during largeflows Plots such as Fig 7 for different timescales could illuminatelinks between wood storage and flux through time and help to charac-terize variability in wood storage as a function of time as done byKramer et al (in press) Developing such plots would involve continuedmonitoring of known sites of wood retention via cameras (minutes todays) field revisits (months to years) and satellite or remote imageryanalysis (years to decades) Especially useful would be measuring theflow stage or discharge so that sudden changes in wood storage couldbe more easily linked to hydrology

Time

ET

Amount

wood in

storage

(S)

+∆S

RT RT

+∆S

Fig 7 Conceptual model of dynamic equilibrium of wood in storage (S) through time Steppetransfer (ET) of wood


Most studies are yearly resurveys that include multiple high flowsrather than investigating the impact from one flood which is problem-atic for relating wood to flow because wood movement is an episodicflood-driven process Thus there is a disconnection between howmobi-lization is measured and the process that drives mobilization Somestudies have used cameras or GPS trackers tomonitor change in storageat finer timescales (Bertoldi et al 2014 Ravazzolo et al 2015b) andothers have made efforts to return to field sites during (Schenk et al2014) or immediately after floods (Dixon and Sear 2014) We recom-mend that more investigators attempt to differentiate wood mobilityfor each wood-transporting flow rather than simply finding the yearlyaveraged change in storage

Our quantitative analysis shows that when flows are sufficientlyhigh large amounts of wood are episodically recruited to or exportedfrom reaches with potential for about 80ndash90 turnover for remobiliza-tion of fluvial wood (Fig 3) The highest turnover rates are for new re-cruits and in areas with large morphological changes (see Section 3)However for a typical year there also appears to be a background fluxof wood for low flows that is at maximum ~30 of the total amount ofwood in storage (Fig 3) Additional data collection focused on the goalof separately measuring smaller-scale yearly flux versus flux frommore rare episodic large flows with potential for high flux would beuseful as would constraining thresholds (for volume of export and forrecurrence interval) between the two types ofwood transfer Consistentlow-flow flux rates may be related to elevations of loose wood belowbankfull If this relationship holds true then wood stored within thelow flow channel could potentially be used universally to obtain roughestimates of background wood exported from basins On top of thisbackground rate episodic flux could be estimated or modelled basedon characteristics of flow wood and the channel reach

Despite having high interannual variability most systems appear tobe in dynamic equilibrium with regard to wood storage and export(conceptualized in Fig 7) Commonly wood mobilization studies notethat wood mobilized out of a reach or buried is replaced with newwood so that the total storage volumes remain nearly the samewith lit-tle net change from year to year (Benke and Wallace 1990 Marcus etal 2002 van der Nat et al 2003 Wohl and Goode 2008 Schenk etal 2014) Thus remote sensing studies that only record changes intotal volumes within a reach may report low mobility because of littlenet change in volume when in fact there was high mobility in piece topiece exchange not identifiable from greater observation distances(eg Curran 2010)

43 Wood transport capacity

The phrase transport capacity is used liberally in wood research torefer to reaches that do not store large amounts of wood Themost gen-eral definition of capacity is the original given by Gilbert and Murphy(1914) lsquothe maximum load a stream can carry (35) With regard towood transport capacity is determined by the effectiveness of a reachat retaining wood for a particular flow (Marcus et al 2002) At low

ET

ET= episodic

transfer

ETRT

-∆S +∆S ∆S=0

RT=regular

transfer

equilibrium

d profile reflects alternating smaller scale regular transfer (RT) and large-scale episodic


moving in

contact with

channel deflecting

against channel

boundaries

0

of time

in contact

with

channel

100

unimpeded

floating

0

of

time fully

floating

100

Transport Stage Ratio (HtH

i)

1

immobile

partially

mobile

fully mobile

QtQ

bf

0

proportion

of stored

wood

1

341

Fig 8 Conceptual transport regimes and thresholds as functions of flow Models inspiredby data supported conceptualizations of bedload particles by Haschenburger (2013) TopMovement regimes of single logs as a function of transport stage (stage height at time t(Ht)stage height at initial mobilization (Hi)) At the threshold for movement (1) thereis a sharp decrease in the time that wood spends in contact with the bed and the timethat it spends floating starts to increase As stage increases wood spends more timefloating and less time deflecting off of channel features In this figure logs do not floatfor 100 of the time because sometimes logs become beached for shorter amounts oftime before continuing downstream Bottom Mobility regimes as a function ofdischarge at time t (Qt) relative to bankfull (Qbf) Position of thresholds depicted couldchange depending on patterns of antecedent floods recruitment events flow stage andchanging channel characteristics Mapping how thresholds change for changingconditions may be a fruitful endeavor


flows many reaches are transport limited and cannot pass wood espe-cially wood of large sizes At high flows when wood transport occurs inalmost all cases rivers can pass most of the wood supplied so that allbut small natural rivers are commonly supply limited at peak woodflux As flood waters recede transport capacity decreases

The rate and timing of decreased transport capacity on the fallinglimb depend not only on the steepness of the falling limb but also onreach andwood characteristics Blanket statements that refer to reachesas having low transport capacity are not particularly helpful becausetransport capacity is not a fixed quantity (Lisle and Smith 2003) but isdependent on wood supply water levels and channel retentivenessMuch research on sediment dynamics has focused on better under-standing the transition between a particle moving as bedload orsuspended load or immobile to mobile during discharge pulses influmes and natural settings Similarly developing relationships be-tween water levels and channel retentiveness for wood pieces of vary-ing sizes as rivers transition from supply limited to transport limitedon the falling limbs of floods would be useful Basically when andwhere does wood transition from immobile to mobile and back toimmobile

Borrowing frombedload research we have constructed two concep-tual models showing transport regimes of wood related to water stageIn Fig 8 (top) we diagram theoretical thresholds for wood movementregimes for a single log as a function of transport stage (ratio of waterstage over stage at incipient motion (HtHi)) on the rising limb of ahydrograph In Fig 8 (bottom) we depict hypotheticalmobility regimesrelated to discharge as a fraction of bankfull (QtQbf)

Transport regimes in Fig 8 (top) include (i) moving in contact withthe channel (ii) deflecting against channel boundaries and (iii) unim-peded floating Mobility regimes in Fig 8 (bottom) include immobilepartially mobile and fully mobile wood loads Again borrowing fromsediment research we define immobile partially mobile and fully mo-bile as b10 10ndash90 and N90mobilization ofwood in storage respec-tively These categories can also be conceptualized based ondownstreammovement as not moved locally repositioned and exporteddownstream These conceptualmodels could easily be tested developedand refined They can also be used for visual display of differences inthreshold position as piece sizes change or for different hydrologic orchannel conditions

44 Downstream wood transport

The downstream transport distance of a sediment particle is com-monly described as path length which is the total lifetime streamwisedisplacement of particles composed of multiple step lengths separatedby rest periods (Haschenburger 2013) Analogous to this is the spiral-ling metaphor for wood movement introduced by Latterell andNaiman (2007) and revitalized in the recent review by Wohl (2016this issue) The spiral metaphor describes the lifetime transport ofwood as a series of spirals along a path The spirals represent rest pe-riods and the width of the spirals depict residence time The distancebetween spirals is the step length between resting locations We haveredrawn the spiral metaphor in the context of a fluvial system in Fig 9to show hypothetical differences in rest periods from small to great riv-ers and downstreammobility on floodplains Quantifying and contrast-ing the density functions of step lengths and rest periods for differentrivers and through basin networks could be a useful manner in whichto identify longitudinal and regional patterns in wood mobility

The length of wood is arguably the most important control on howwood behaves (where it is stored and when it is mobilized) Diameter(or flotation depth) is also an important variable when flow depthsare similar to flotations depths (see Section 22) In our quantitativeanalysis we used L (length of woodchannel width) to place channelsinto woody dynamic size categories of small medium large and great(Fig 5) The L proved to be a useful dimensionless metric that allowedfor easy comparison of wood mobility metrics across stream type and


size However as discussed in Section 3 more work needs to be doneto determine which length of wood should be used in the L equationto maximize predictability of wood mobility Should this be Lmax or asmaller size fraction like L90 or L85 Also MacVicar and Pieacutegay (2012)suggested using ϕ = log2 (L) to define wood size classes as done byCadol andWohl (2010) and Iroumeacute et al (2010) and as used to definesediment size categories Although this idea has not thus far garneredmuch support from field scientists the base 2 log transform of woodlengths has proven useful when modelling and explaining woodmobility

Despite the importance of rootwads for limiting mobility and trans-port distances (Wohl andGoode 2008Davidson et al 2015) almost noinformation exists relating rootwad size type and shape characteristicsto mobility beyond noting presence or absence One exception is a re-cent flume study that suggests that shorter bole lengths on pieceswith rootwads actually increases stability over longer lengths(Davidson et al 2015) but this is yet to be corroborated in the field

Investigators commonly assume that the transport of wood in-creases as the widths of channels increase relative to the length ofwood and therefore larger rivers have greater transport capacity thansmaller rivers Although potential transport distance appears to betwo orders of magnitude greater for channels wider than maximum


Small Rivers

Medium

Rivers

Large Rivers

Great Rivers

Residence Time (Tr)

Travel

Distance

(Xw)

lateral

export

basin

export

Fig 9 Conceptual model of the spiralling movement of large wood through a fluvialsystem Size of loops represent residence times at a single location before reenteringtransport Connected loops to the side of the main vector represent downstreamtransport on a floodplain The lifetime for a piece of wood from its starting location inthe channel to its final resting place is the sum of all the residence times and the traveltime (most likely very small compared to rest periods) The distance travelled is thesum of all the step lengths Throughout the lifetime of the wood piece from source tosea the piece undergoes vertical lateral and downstream exchanges Along any spiral apiece of wood may decay in place ending its path Diagram after Latterell and Naiman(2007) and Schumm (1977)


log lengths (Fig 5) potential mobilization of wood in storage betweenchannel sizes is nuanced Median values of mobilization increase withincreasing channel size which suggests that larger channels do trans-port higher proportions of stored wood more regularly than smallerchannels (Fig 6) However event-based turnover of stored wood ismaximized in medium channels (Fig 5) Lower maximummobilizationrates on large rivers compared to medium rivers seem reasonable be-cause the stored wood in larger rivers is more likely to be partially bur-ied or scattered on floodplains at farther distances from the mainchannel Medium-sized channels are typically more confined thanlarge or great rivers with smaller floodplains and coarser substrateConsequently most of the stored wood in medium rivers is closer tothe thalweg where flow velocities are the greatest and less opportunityfor anchoring via burial or by instream live-wood We suggest that thelive-wood growing on islands bars and frequently inundated flood-plains likely counterbalances the increased conveyance of wood in larg-er rivers

Unfortunately general ignorance of the relative importance of fac-tors influencing downstream transport of mobilized wood exists dueto the scarcity of studies that actually track transport of wood duringfloodsWe found 17 studies that trackedwood via tagging RFID or teth-ered GPS boxes Of these only two actually tracked all wood regardless


of how far each piece travelled (Dixon and Sear 2014 Ravazzolo et al2015b) Two other studies had 100 recovery for a year because theywere able to recover the one piece that moved during lower-than-aver-age flows (Pecorari 2008 Schenk et al 2014) Because maximumtransport distances were commonly bounded by study reach lengthsreported mean transport distances do not reflect the entire transportedpopulation

45 Functional classification from wood dynamics

In this paper we have used functional guidelines to differentiatesmall medium large and great rivers based on wood dynamics (Table1 Fig 2) Originally introduced by researchers in the Pacific Northwest(Keller and Swanson 1979 Nakamura and Swanson 1993) and suc-cinctly summarized by Church (1992) this functional classification ofstreams has been utilized in Europe (Pieacutegay and Gurnell 1997Gurnell et al 2002) and more recently revitalized in the context of rel-ative importance of inputs and outputs in wood budgets (Wohl 2016this issue) Categorizing rivers based onwood characteristics is justifiedfor forested channels because the geomorphology of these channels iswood-driven (Le Lay et al 2013) and as channel size increases the be-haviour of wood also changes (Church 1992 Gurnell et al 2002Wohl2016) (Table 1)

Because channels are not subdivided into these categories basedsolely on their size but on the patterns and functions of wood withinthem Church (1992) makes the point that a small river could behavelike a large river if all the wood being transported is less than thewidth of the channel However referring to headwater reaches aslarge when they do not carry larger pieces of wood is confusing andnot visually intuitive Because the names of these functional categoriza-tions are small medium large and great therewill always be a propen-sity to classify them into these categories strictly by channel size ratherthan by incorporating wood dynamic process domains as originallyintended

Although the functional wood dynamic classification (Table 1) con-nects certain wood process regimes to sizes of channels and thus en-ables testing of general network trends it fails to capture spatial andtemporal network heterogeneity For example a small creek may havesome reaches that are like a small channel (recruitment-dominated)and others that behave like a medium channel (jam-dominated) In ad-dition to referring to rivers by size class to convey scale we suggest clas-sifying rivers by process domains related to definable regimes of wooddynamics Universal succinct process domain categories are useful tofacilitate comparison among studies to explore the temporal-spatialdistribution and heterogeneitywithin channel networks and to identifysets of predictive equations that performbetter under different regimes

One option is to designate process domain names that refer to thedominant storage transport and recruitment regimes Davidson et al(2015) described two storage regimes a randomly distributed newlyrecruited state and a self-organized jam stable state Braudrick et al(1997) described three transport regimes congested semi-congestedand uncongested flow We suggest that primary reach-scale wood pro-cess domains are recruitment-dominated regimes jam-dominated re-gimes flow-dominated regimes and burialexhumation-dominatedregimes Additional process domains are also possible For exampleflash-flood-dominated regime might be the best category for desertephemeral channels Jochner et al (2015) present a compelling fourend-member conceptual model of wood regimes as event-driven ex-port (continuous recruitment episodic export) event-driven delivery(episodic recruitment continuous export) fully episodic (episodic re-cruitment and export) or fully continuous (continuous recruitmentand export)

Reach-scale process domains likely transition through time fromone regime to another because of fluctuations in flow and changes inmorphology disturbance or systemwide trajectories caused by region-al drivers such as climate change or alterations of land use Once a



consistent set of process domains is defined this can be used as a tool toexplore and map temporal and spatial transitions between processdomains

46 A discontinuous network - the traffic metaphor

Wooddynamics are different for differently sized streams and deliv-ery ofwood fromonepart of a streamnetwork to another is a discontin-uous or episodic process as depicted in Figs 7 and 9 Thus we proposethat an aptmetaphor forwood transport is vehicle traffic Just as hydrol-ogy wood characteristics and biogeomorphic reach characteristics con-trol the movement of wood through stream networks (Gurnell et al2015) motivation to travel type of vehicle and road conditions governhow people travel through road networks

If hydrologic conditions do not meet base thresholds little to nowood flux occurs which is analogous to the underlying motivationsthat govern when and how far people will travel (ie wood transportmobilization and travel distance) andwhen andwhere they are station-ary (ie wood residence time) Sometimes special events cause extracongested traffic This is similar to high wood congestion caused bylarge disturbancesMore commonly daily routines andwork commutesgovern traffic conditions This is similar to regular background woodflux from normal yearly floods Travelling at night is less common andonly under special circumstances will drivers be on the road duringthis time This is similar to lack of wood flux during low flows unlesswood is newly delivered from a localized bank failure

Wood characteristics can be thought of as the type of vehicle inwhich one is travelling The vehicle governs which paths or road canbe taken and the speed of travel Someone on a motorized scooter willtake different paths than other vehicles just as a small piece without arootwad may take a different path than larger pieces with rootwadsHowever in some conditions such as a traffic jam everyone travelsthe same speed which equates to fully congested wood transport

Traffic conditions and movement are not only governed by the mo-tivation (hydrology) and the vehicle (wood characteristics) but thestate of the roads which controls how fast or how slow a destination

Rising

Limb

gt~34 bf

le bf

lt1

lt1

Wood Flow and C

Governing Transp

hydrograph position

Falling

Limb

stage height

gt34 bf

gt1

antecedent peak divide peak

Hydrologic Thresholds

for Transport

asymp1

gt1

wood length

divide

flow width fl

Deposi

Domina

the more conditions in gre

stop go

Fig 10Conceptualmodel ofwoodmovement throughdrainagenetworks as a traffic stop lightmfigure the reader is referred to the web version of this article)


is reached Road conditions are similar to how biogeomorphic charac-teristics of reaches such as presence of live-wood degree of confine-ment channel planform complexity density of logjams access tofloodplains and other factors that can limit or increase the distanceand rate of movement of wood downstream

We have conceptualized this metaphor in Fig 10 as a flow chart ofstoplights Whether a piece of wood will be moved at any given mo-ment can be predicted by whether the piece meets minimum mobilitythresholds that allow it to go forward (green) possibly move or prog-ress slowly (yellow) or stop (red)We first assess hydrologic conditionsfor transport to determine whether wood meets minimum thresholdsfor mobility prior to assessing wood or reach characteristics If a pieceof wood has potential to be transported based on hydrology then theunique interactions between its characteristics and the channel areassessed Thus for anymoment of time a snapshot can be used to assesswhere individualwood pieces are located andwhether they are likely tomove based on the hydrologic wood and channel traffic conditions (seeFig 10) Assessments made over the course of a flood can be used toprovide estimates of the overall efficiency of wood movement for spe-cific floods We recognize that this model is a simplification of the com-plex interactions involved However we consider the model a usefulframework for modelling and exploring temporal variations in woodflux

47 Moving forward

The main limitations to describing wood transport are inadequateobservation timescales and lack of sufficient mobility data from diverserivers and regions that also capture variability between reaches Theseare similar hurdles to quantifying bedload Bedload transport equationsdo not perform well for coarse-grained substrate with poor sorting be-cause they fail to integrate spatial and temporal complexities that influ-ence grain entrainment (Haschenburger 2013)

In order to improvemodels ofwood flux on local and regional scaleswe need better characterization of average step lengths within the life-time travel path of a piece of wood (see Fig 9) and we need a better

lowlimited noneFew

High

hannel Boundary Interactions

ort Variability

transport

congestion

high

degree of

flow

obstruction

Complete

partial

living

vegetation

in flow

lots

some

flow depths

lt

otation depth

most

partial

Transport Distances and Wood Flux

ModLow

semi

tion

ted

Transport

Dominated

en and the greater amount of time in green = more transport

stop

go

etaphor Fully described in Section 46 (For interpretation of the references to color in this



understanding of how changes in storage through time are related tovariability in wood flux and hydrology Efforts to better define entrain-ment and entrapment conditions and thresholds will be extremely use-ful especially focused on individual flows or patterns over decades Awider range of flow observations for diverse rivers will help to linkreach-scale processes with network-scale processes

A general approach that will lead to efficient quantification of woodtransport is to design studies to constrain thresholds between transportregimes for different channel sizes channel morphologies hydrologyand wood characteristics Until recently most field studies that includedwood transport data treated wood transport as a secondary study goalafter description ofwood storagewood recruitment or ecological impactAs more researchers make wood movement their primary focus thesethresholds will be rapidly identified We have presented several concep-tual frameworks that may prove useful to guide such work (Figs 8ndash10)

Below we summarize some specific suggested approaches for ac-quiring data to achieve broader spatial and temporal coverage ofwood dynamics

bull Monitor wood in action This can be done at a station by monitoringpassage via automatic video monitoring (eg MacVicar and Pieacutegay2012) coarse interval timelapse cameras (eg Kramer and Wohl2014 Kramer et al in press) or radio tags (eg Schenk et al 2014)The downstream movement of logs can by tracked with GPS (egRavazzolo et al 2015b) or by actively following radio tags duringflooding via boat or aircraft (eg Schenk et al 2014)

bull Monitor change in storage at known retention sites at varying timescales(eg Wohl and Goode 2008 Moulin et al 2011 Bertoldi et al 2013Schenk et al 2014 Boivin et al 2015)

bull Use wood characteristics to fingerprint wood source This can help iden-tify wood-contributing subbasins and travel distances Moulin andPieacutegay (2004) were very successful in making inferences aboutwood flux at basin scales based on the characteristics of wood trappedin a reservoir

bull Quantify the amount of buried wood Buried wood has been identifiedsuccessfully using acoustic bathymetry (White and Hodges 2003)and ground penetrating radar (Kramer et al 2012 Valdebenito etal 2016) but values of wood buried in stream beds are largelyunquantified Buried wood is an important component of wood fluxbecause in rivers with sediment loads capable of easily buryingwood at least the same amount or more that is exported may be bur-ied For example three-quarters of thewood exported from the LowerRoanoke River to the ocean in North Carolina was buried ordecomposed en route (Schenk et al 2014)

bull Use remote sensing techniques to assess change on larger spatial andlonger temporal scales (eg Bertoldi et al 2013 Atha 2014 Ulloaet al 2015 Kramer et al in press)

bull Conduct stratigraphic andor other analysis of wood deposited in basinsand floodplains to obtain long-term (decade to millennial scale) re-cords of wood flux from watersheds (Guyette et al 2008 Seo et al2008 Seo and Nakamura 2009 Fremier et al 2010 Boivin et al2015 Kramer and Wohl 2015)

bull Use already existing data from unconventional sources Hidden datawithin government agencies individual scientists or private compa-nies that have never been published or otherwise made easily avail-able but that can be acquired if requested is often quite usefulHeidorn (2008) calls these dark data For example Moulin andPieacutegay (2004) Seo et al (2008) and Fremier et al (2010) successfullyused reservoir debris extraction records to indirectly estimate basinwood flux In addition to finding and using dark data a vast amountof unconventionally collected data is freely available on the internetWood researchers have yet to take advantage of this Numerousposted photos and videos of rivers that includewood could be utilizedto expand the geographic extent of studies Videos of wood transportespecially wood transport from ice-jam flooding flash flooding in de-serts and catastrophic flooding could be analyzed to estimate wood


flux during rare events Some internet contributors have YouTube(wwwYouTubecom) channels devoted to chasing flash floods Alsowith the increased use of small waterproof action cameras outdoorrecreation enthusiasts are posting to the internet continuous footageof their excursions down sections of rivers in remote regions Reportsof changes in wood are commonly posted to online whitewater kay-aking forum boards on sections of rivers that are regularly navigatedFinding ways to automatically and easily curate and utilize these datais a worthy research endeavor

bull Participate and use Web 20 and create and utilize citizen science initia-tives Web 20 refers to the part of the internet that is interactivesuch as social media and citizen science platforms (eg wwwcitsciorg wwwcrowcraftingorg) Citizen science refers to the use of non-scientists to help collect process or analyze data Although citizen sci-ence initiatives have been utilized in ecology medicine andastronomy (eg bird surveys gene mapping star classification)they have been underutilized by large instream wood researchers(we came across none) In a short review of the use of citizen scienceSilvertown (2009) made the point that Almost any project that seeksto collect large volumes offield data over awide geographical area canonly succeed with the help of citizen scientists (469) Citizen sciencecould be used not only to collect data from diverse regions but alsoto validate and train automatic image processing routinesWeb 20 not only opens up real-time interaction between scientistsand non-scientists but can facilitate data collection and curationfrom diverse individual scientists globally to reduce the amount oflsquodark data and facilitate synthesis between studies This has alreadybeen done in medical fields to advance treatment for particular dis-eases by synthesizing and collecting information on case studiesfrom doctors practicing independently (eg Butzkueven et al2006) A large but highly rewarding project would be to develop anonline interactive river wood data platform where field scientistsand managers could add and contribute their data while interactingwith each other This would facilitate better curation of metadatacommon reporting of metrics access to dark data and internationalcollaborations across disciplines

bull Compile quantitative reviews that integrate case study information forbasin- and reach-scale wood flux wood recruitment rates residencetimes and storage patterns

5 Conclusion

Wepropose that the stop and go the jamming and unjamming or thediscontinuity of wood flux is the most important aspect of the wood re-gime for river morphology dynamics and biota rather than wood stabil-ity Therefore wood transport dynamics need to be incorporated intoconceptual and quantitativemodels of river systems riparian ecosystemsand nutrient routing This requires substantial effort to obtain an equiva-lent working knowledge of wood transport as currently exists for sedi-ment transport (eg Haschenburger 2013 Kuhnle 2013) This papercontributes to this effort by summarizing existing transport premisesand ideas from prior studies synthesizing quantitative results on woodtransport from field studies and presenting knowledge gaps conceptualmodels and hypotheses that can be used to design future field flumeand modelling studies In an era in which new remote technologies andnew sources of data are increasingly accessible and applicable to researchonwood in river corridorswe anticipate that studies ofwood transport inrivers are poised to yield significant insights on wood dynamics and riverecosystem management

Acknowledgements

This research was funded by the Warner College of Natural Re-sources Colorado State University and ColoradoWater Institute Student



Grant (National Institutes for Water Resources (NIWR) Fund5328011) We would like to thank Dan Scott for review and feedbackas well as the rest of the CSU Fluvial Geomorphology graduate researchlab Wewould also like to thank three anonymous reviewers and editorLorenzo Picco who all greatly improved the manuscript through theirthoughtful comments

Appendix A Equations of wood mobility

A1 Transport distance based on exponential scaling by reach retentionrates (Millington and Sear 2007)

PdP0eminuskd

where Pd = pieces that moved distance d in meters P0 = initial ofpieces introduced k = per meter retention rate 1k represents themean travel distance in m Equation used in experimental addition ofsmall dowels (le106 m in length and le 035 m in diameter) released inHighland Water a small meandering natural channel in England

A2 Transport distance as an exponential function of water depth at peakflow (Haga et al 2002)

y1 frac14 052e1275xR2 frac14 058 n frac14 60eth THORNy2 frac14 045e1301xR2 frac14 065 n frac14 37eth THORNy3 frac14 026e1397xR2 frac14 071 n frac14 9eth THORNy4 frac14 384e832xR2 frac14 061 n frac14 15eth THORN

where yi is travel distance for a series of flow events and x is an estimateof water depth at peak flow Equations derived from experimental re-lease of logs (Length = 17 mplusmn 04 Diam= 14 cmplusmn 4) stripped of ir-regularities in a 5500-m-long section of the gravel-cobble beddedOyabu Creek in Japanwith no boulders or instreamwood to block trav-el bankfull width = 9 m gradient = 40 size of released logs not rep-resentative of maximum riparian heights (20ndash30 m) riparian treespecies were beech (Fagus crenata) oak (Quercus mongolica) Japanesecherry birch (Betura grossa) fir (Abies firma) and hemlock (Tsugasieboldii)

A3 Transport distance as a function of jam spacing (Martin and Benda2001)

ξ frac14 L jTp

T jβminus1

where ξ is the transport distance over a lifetime of a piece of wood Lj isthedistance between transport obstructing jams TpTj is the longevity ofwood over longevity of jams and β is the transport-obstructing effec-tiveness of jams Theoretical quantitative equation based on interjamspacing and degree of channel obstruction for 28 reaches ranging from33 to 23 m in width within the Game Creek watershed in southeastAlaska assumed that transport distance is limited to interjam spacingdid not directly measure distance tree species are western hemlock(Tsuga heterophylla) Sitka spruce (Picea sitchensis) Sitka alder (Alnussinuate) and salmonberry (Rubus spectabilis)

A4 Stochastic models of travel distance and mobility of individual pieces(Lassettre and Kondolf 2003)

Td frac14 L jMF ln RIeth THORNfrac12 Pm frac14 m1 thornm2MF thornm3 ln RIeth THORNfrac12 MF frac14 1

1thorn em1thornm2

Lwbkf

thornm3DthornC1thornC2thornC3thornC4thornC5


whereMF is a mobility factor with values between 0 and 1 Pm is prob-ability of movement and Td is travel distance RI is recurrence intervalm1 m2 and m3 are constants L is the length of a piece of wood D is av-erage diameter of a piece of woodwbkf is the channel width at bankfulland Ci are categorical variables of decay class by species species stabil-ity by type rootwad presence and cut Empirical equations developedfor individual pieces of wood in central California in meandering sec-tions of Amaya Creek (wbkf=6 m) and East Branch Soquel Creek(wbkf=12 m) characterized by channel-spanning log jams streamwood included big leaf maple (A macrophyllum) red alder (A rubra)tanoak (Lithocarpus densiflora) coast redwood (S sempervirens) andDouglas-fir (Psuedotsugamenziesii)maximum log lengthwas 60m as-sumed wood moved only the average jam spacing (Lj) for yearly occur-ring floods travel distance was adjusted upwards for more mobilepieces and to account for the fact that wood could pass jams by amulti-ple of MF and the natural log of the return period (RI)

A5 Empirical logisticmodel of key piecemobility based onwood character-istics (Wohl and Goode 2008)

Pm frac14 1= 1thorn eminusxeth THORNx frac14 14thorn 052Llog thorn 005D

logminus002C1minus013C2 thorn 020C3R2 frac14 047

where Pm is probability of key piece mobility L is the piece length di-vided by average reference channel width and D is the dimensionlessannual peakflowdepth divided by piece diameter Categorical variablesC1 C2 and C3 arewhether a key piece is a bridge unattached or rampedrespectively Developed using five high elevation streams of the RockyMountains in Colorado after 10 years of repeat surveys channel widthsranged from 43 to 65 m maximum wood length was 18 m and in-stream wood was mostly conifers Engelmann spruce (Piceaengelmannii) subalpine fir (Abies lasiocarpa) and lodgepole pine(Pinus contorta)

A6 Empirical linear relationships for percent mobility based on flow char-acteristics (Iroumeacute et al 2015)

M frac14 minus42thorn 00065ω Hmaxfrac12 R frac14 062 n frac14 17eth THORNM frac14 minus14514thorn 20 Hmax=HBketh THORNR frac14 060 n frac14 17eth THORN

where M is percent mobility ω[Hmax] is unit stream power for maxi-mum stage height in Nm3 and HmaxHBk is ratio of maximum stageheight over stage height at bankfull Fitted for forested mountainousheadwater rivers in Chile channel widths range from 5 to 13 m maxi-mum length of instream wood was 25 m and wood type is dominatedby native coihue (Nathofagus dombeyi and nervosa) and tepa(Laureliopsis philippiana) as well as plantations of eucalyptus (Eucalyp-tus globulus) and pine (Pinus sp) results showed wide scatter with in-creasing variance among higher values of the explanatory variable

A7 Empirical exponential relationship between wood volume and woodvelocity (Ravazzolo et al 2015b)

vw frac14 071V022w R2 frac14 087 n frac14 5eth THORN

where vw is wood velocity in ms and Vw in m3 is wood volume Equa-tion was developed using data from five logs with GPS tags and trackedduring a flood in the large 800 mwide bar-braided Rio Tagliamento innortheastern Italy in-stream wood is at maximum 25 m in length andprimarily alder (Alnus incana) poplar (Popoulas nigra) and willow(Salix alba)


7


A8 Logistic model of mobilization based on wood characteristics (Mertenet al 2010)

Pmob frac14 ex= 1thorn exeth THORNx frac14 039minus264β1 thorn 086β2minus152β3minus077β4minus080β5minus009β6minus159βn frac14 865 Pb0001NagelkerkesR2 frac14 039

where Pmob is the probability ofmobilizationβ1= burial β2= effectivedepth β3 = length ratio β4 = bracing β5 = rootwad presence β6 =downstream force ratio β7 = draft ratio Developed using data frominstream large wood within the channel from nine forested streamsdraining into the north shore of Lake Superior Minnesota piece lengths38 plusmn 3 m and diameters 018 plusmn 13 m no tree species specified flowdepths ranged from 053 to 248 m velocity from 086 to192 msstream power from 15 to 252 Nm s slopes from 0001 to 002 bankfullwidths from 34 to 24 m and peak flow from 21 to 547 m3s data col-lected during year of extreme drought

A9 Transport ratio as a function of wood characteristics and discharge forsingle thread (TrS) versus multithread (TrM) reach (Ruiz-Villanueva et al2015c)

as a function of wood Volume (Vw)

TrS frac14 031Vminus029w R2 frac14 056

TrM frac14 003Vminus125w R2 frac14 033

as a function of wood diameter (Dw) and mean water depth (Wdepth)

TrS frac14 minus018 Dw=Wdepth thorn 032R2 frac14 093

TrM frac14 minus049 Dw=Wdepth thorn 058R2 frac14 094

as a function of wood length (Lw) and channel width (w)

TrS frac14 minus219 Lw=weth THORN thorn 092R2 frac14 091TrM frac14 240 Lw=weth THORN thorn 012R2 frac14 082forLw=wb012TrM frac14 minus291 Lw=weth THORN thorn 077R2 frac14 073forLw=wN012

as a function of wood density (ρw)

TrS frac14 3278eminus389ρwR2 frac14 098n frac14 4

TrM frac14 1036minus183ρwR2 frac14 084n frac14 5

as a function of discharge (Q)

TrS frac14 minus012thorn 0004Q R2 frac14 091 forQb100RI frac14 10TrS frac14 minus025thorn 0001Q R2 frac14 044 forQN100RIN10

OR

TrS frac14 minus196thorn 35Q011R2 frac14 091TrM frac14 minus017thorn 0005Q R2 frac14 097 forQb110RIb15TrM frac14 03thorn 0001Q R2 frac14 055 forQN110RIN15

OR

TrM frac14 minus212thorn 132Q013R2 frac14 095

The transport ratio T is the amount exported divided by the totalamount imported to the reach These series of equations developedfrom numerical simulation using Iber Wood computer model and simu-lating wood and channel characteristics from field data collected fromthe Czarny Dunajec River in the Polish Carpathians simulated multi-and single-thread reaches tree species included large alders (Alnusincana) large willows (Salix fragilis and S alba) and young willows (Spurpurea and S eleagnos) wood lengths ranged from 1 to18 mmean= 125m widths from 005 to 08 mmean= 023m and density


from 04 to 095 g cmminus1 mean= 056 g cmminus1 results are from simula-tion of 10-year flood (Q = 105 m3s) and used mean values except forthe variable for which the relationship was modelled

Appendix B Supplementary data

Supplementary data to this article can be found online at httpdxdoiorg101016jgeomorph201608026

References

Abbe T Brooks A 2011 Geomorphic engineering and ecological considerations whenusing wood in river restoration Stream Restor Dyn Fluvial Syst pp 419ndash451httpdxdoiorg1010292010GM001004

Anderson NH Sedell JR 1979 Detritus processing by macroinvertebrates in streamecosystems Annu Rev Entomol 24 351ndash377

Angradi TR Schweiger EW Bolgrien DW Ismert P Selle T 2004 Bank stabilizationriparian land use and the distribution of large woody debris in a regulated reach ofthe upper Missouri River North Dakota USA River Res Appl 20 829ndash846 httpdxdoiorg101002rra797

Angradi TR Taylor DL Jicha TM Bolgrien DW Pearson MS Hill BH 2010 Littoraland shoreline wood inmid-continent great rivers (usa) River Res Appl 26 261ndash278httpdxdoiorg101002rra1257

Atha JB 2014 Identification of fluvial wood using Google Earth River Res Appl 30857ndash864 httpdxdoiorg101002rra2683

Beckman ND Wohl E 2014 Effects of forest stand age on the characteristics of logjamsin mountainous forest streams Earth Surf Process Landf 39 1421ndash1431 httpdxdoiorg101002esp3531

Benda LE Sias JC 2003 A quantitative framework for evaluating the mass balance ofin-stream organic debris For Ecol Manag 172 1ndash16 httpdxdoiorg101016S0378-1127(01)00576-X

Benke AC Wallace JB 1990 Wood dynamics in coastal plain blackwater streams CanJ Fish Aquat Sci 47 92ndash99

Berg N Carlson A Azuma D 1998 Function and dynamics of woody debris in streamreaches in the central Sierra Nevada California Can J Fish Aquat Sci 551807ndash1820 httpdxdoiorg101139f98-064

Bertoldi W Gurnell A Welber M 2013 Wood recruitment and retention the fate oferoded trees on a braided river explored using a combination of field and remote-ly-sensed data sources Geomorphology 180181 146ndash155 httpdxdoiorg101016jgeomorph201210003

Bertoldi WWelber M Mao L Zanella S Comiti F 2014 A flume experiment onwoodstorage and remobilization in braided river systems Earth Surf Process Landf 39804ndash813 httpdxdoiorg101002esp3537

Bertoldi WWelber M Gurnell A Mao L Comiti F Tal M 2015 Physical modelling ofthe combined effect of vegetation and wood on river morphology Geomorphology246 178-187 httpdxdoiorg101016jgeomorph201505038

Bevan A 1948 Floods and Forestry University of Washington Forest Club Quarterly USForest Service

Bilby RE 1984 Removal of woody debris may affect stream channel stability J For 82(10) Society of American Foresters pp 609ndash613

Bilby RE 1981 Role of organic debris dams in regulating the export of dissolved andparticulate matter from a forested watershed Ecology 62 1234ndash1243 httpdxdoiorg1023071937288

Bilby RE Ward JW 1989 Changes in characteristics and function of woody debris withincreasing size of streams inwesternWashington Trans Am Fish Soc 118 368ndash378httpdxdoiorg1015771548-8659(1989)118b0368CICAFON23CO2

Bocchiola D Rulli M Rosso R 2006a Flume experiments on wood entrainment in riv-ers Adv Water Resour 29 1182ndash1195 httpdxdoiorg101016jadvwatres200509006

Bocchiola D Rulli M Rosso R 2006b Transport of large woody debris in the presenceof obstacles Geomorphology 76 166ndash178 httpdxdoiorg101016jgeomorph200508016

Bocchiola D Rulli MC Rosso R 2008 A flume experiment on the formation of woodjams in rivers Water Resour Res 44 httpdxdoiorg1010292006WR005846

Boivin M Buffin-Beacutelanger T Pieacutegay H 2015 The raft of the Saint-Jean River Gaspeacute(Queacutebec Canada) a dynamic feature trapping most of the wood transported fromthe catchment Geomorphology 231 270ndash280 httpdxdoiorg101016jgeomorph201412015

Bragg DC 2000 Simulating catastrophic and individualistic large woody debris recruit-ment for a small riparian system Ecology 81 1383ndash1394 httpdxdoiorg1018900012-9658(2000)081[1383SCAILW]20CO2

Braudrick CA Grant GE 2000 When do logs move in rivers Water Resour Res 36571ndash583 httpdxdoiorg1010291999WR900290

Braudrick CA Grant GE 2001 Transport and deposition of large woody debris instreams a flume experiment Geomorphology 41 263ndash283 httpdxdoiorg101016S0169-555X(01)00058-7

Braudrick CA Grant GE Ishikawa Y Ikeda H 1997 Dynamics of wood transport instreams a flume experiment Earth Surf Process Landf 22 669ndash683 httpdxdoiorg101002(SICI)1096-9837(199707)227b669AID-ESP740N30CO2-L

Brooks AP Brierley GJ 2002 Mediated equilibrium the influence of riparian vegeta-tion and wood on the long-term evolution and behaviour of a near-pristine riverEarth Surf Process Landf 27 343ndash367



Butzkueven H Chapman J Cristiano E GrandMaison F Hoffmann M Izquierdo GJolley D Kappos L Leist T Poumlhlau D et al 2006 MSBase an international onlineregistry and platform for collaborative outcomes research in multiple sclerosis MultScler 12 769ndash774

Cadol D Wohl E 2010 Wood retention and transport in tropical headwater streams LaSelva Biological Station Costa Rica Geomorphology 123 61ndash73 httpdxdoiorg101016jgeomorph201006015

Cenderelli D Wohl E 2003 Flow hydraulics and geomorphic effects of glacial-lake out-burst floods in the Mount Everest region Nepal Earth Surf Process Landf 28385ndash407 httpdxdoiorg101002esp448

Choneacute G Biron PM 2015 Assessing the relationship between river mobility and habi-tat River Res Appl httpdxdoiorg101002rra2896

Church M 1992 Channel morphology and typology The Rivers Handbook 1pp 126ndash143

Collins BD Montgomery DR Fetherston KL Abbe TB 2012 The floodplain large-wood cycle hypothesis a mechanism for the physical and biotic structuring of tem-perate forested alluvial valleys in the North Pacific coastal ecoregion Geomorphology139140 460ndash470 httpdxdoiorg101016jgeomorph201111011

Curran JC 2010 Mobility of large woody debris (LWD) jams in a low gradient channelGeomorphology 116 320ndash329 httpdxdoiorg101016jgeomorph200911027

Daniels MD 2006 Distribution and dynamics of large woody debris and organic matterin a low-energy meandering stream Geomorphology 77 286ndash298 httpdxdoiorg101016jgeomorph200601011

Davidson SL Eaton BC 2015 Simulating riparian disturbance reach scale impacts onaquatic habitat in gravel bed streams Water Resour Res httpdxdoiorg1010022015WR017124

Davidson S MacKenzie L Eaton B 2015 Large wood transport and jam formation in aseries of flume experiments Water Resour Res httpdxdoiorg1010022015WR017446

Desloges J Church M 1992 Geomorphic implications of glacier outburst floodingNoeick River valley British Columbia Can J Earth Sci 29 551ndash564 httpdxdoiorg101139e92ndash048

Diez JR Elosegi A Pozo J 2001 Woody debris in North Iberian streams influence ofgeomorphology vegetation and management Environ Manag 28 (5) 687ndash698httpdxdoiorg101007s002670010253 (Springer)

Dixon SJ Sear DA 2014 The influence of geomorphology on large wood dynamics in alow gradient headwater streamWater Resour Res 50 9194ndash9210 httpdxdoiorg1010022014WR015947

Eaton B Hassan M Davidson S 2012 Modeling wood dynamics jam formation andsediment storage in a gravel-bed stream J Geophys Res Earth Surf 117 httpdxdoiorg1010292012JF002385

Elosegi A Diacuteez J Pozo J 2007 Contribution of dead wood to the carbon flux in forestedstreams Earth Surf Process Landf 32 1219ndash1228 httpdxdoiorg101002esp1549

Elosegi A Diez JR Pozo J 1999 Abundance characteristics and movement of woodydebris in four Basque streams Arch Hydrobiol 144 (4) 455ndash471 Schweizerbart

Fremier AK Seo JI Nakamura F 2010Watershed controls on the export of large woodfrom stream corridors Geomorphology 117 33ndash43 httpdxdoiorg101016jgeomorph200911003

Gallisdorfer MS Bennett SJ Atkinson JF Ghaneeizad SM Brooks AP Simon ALangendoen EJ 2014 Physical-scalemodel designs for engineered log jams in riversJ Hydro Environ Res 8 115ndash128 httpdxdoiorg101016jjher201310002

Gibling MR Davies NS 2012 Palaeozoic landscapes shaped by plant evolution NatGeosci 5 99ndash105 httpdxdoiorg101038ngeo1376

Gilbert GK Murphy EC 1914 The transportation of debris by running water TechnicalReport US Geological Survey Professional Paper 86

Gippel CJ 1995 Environmental hydraulics of large woody debris in streams and riversJ Environ Eng 121 388ndash395

Gippel CJ Finlayson BL ONeill IC 1996 Distribution and hydraulic significance oflarge woody debris in a lowland Australian river Hydrobiologia 318 179ndash194httpdxdoiorg101007BF00016679

Grant GE OConnor JE WolmanMG 2013 A river runs through it conceptual modelsin fluvial geomorphology In Wohl E (Ed) Treatise on Fluvial Geomorphology Ac-ademic Press San Diego pp 6ndash21 httpdxdoiorg101016B978-0-12-374739-600227-X

Grette GB 1985 The role of large organic debris in juvenile salmonid rearing habitat insmall streams MS thesis University of Washington Seattle Wash

Gurnell AM 2003 Wood storage and mobility In Gregory S Boyer K AM G (Eds)The Ecology and Management of Wood in World Rivers American Fisheries SocietySymposium pp 75ndash91

Gurnell A Piegay H Swanson F Gregory S 2002 Large wood and fluvial processesFreshw Biol 47 601ndash619 httpdxdoiorg101046j1365-2427200200916x

Gurnell AM Corenblit D Garciacutea de Jaloacuten D Gonzaacutelez del Taacutenago M Grabowski RCOHare MT Szewczyk M 2015 A conceptual model of vegetation-hydrogeomorphology interactions within river corridors River Res Appl httpdxdoiorg101002rra2928

Guyette RP Dey DC Stambaugh MC 2008 The temporal distribution and carbonstorage of large oak wood in streams and floodplain deposits Ecosystems 11643ndash653 httpdxdoiorg101007s10021-008-9149-9

Haga H Kumagai T Otsuki K Ogawa S 2002 Transport and retention of coarse woodydebris in mountain streams an in situ field experiment of log transport and a fieldsurvey of coarse woody debris distribution Water Resour Res 38 httpdxdoiorg1010292001WR001123

Harmon ME Franklin JF Swanson FJ Sollins P Gregory S Lattin J Anderson NCline S Aumen N Sedell J et al 1986 Ecology of coarse woody debris in temper-ate ecosystems Adv Ecol Res 15 302


Harvey A 1984 Geomorphological response to an extreme flood a case from southeastSpain Earth Surf Process Landf 9 267ndash279 httpdxdoiorg101002esp3290090306

Haschenburger J 2013 Bedload kinematics and fluxes In Wohl E (Ed) Treatise on Flu-vial Geomorphology Academic Press San Diego pp 103ndash123

Heidorn PB 2008 Shedding light on the dark data in the long tail of science Libr Trends57 280ndash299 httpdxdoiorg101353lib00036

Hickin EJ 1984 Vegetation and river channel dynamics Can Geogr 28 111ndash126Hilton RG Galy A Hovius N 2008 Riverine particulate organic carbon from an active

mountain belt importance of landslides Glob Biogeochem Cycles 22 httpdxdoiorg1010292006GB002905

Hogan D 1984 The influence of large organic debris on channel morphology in QueenCharlotte island streams Proceedings of the Western Division of the American Fish-eries Society Victoria British Columbia pp 263ndash273

Iroumeacute A Andreoli A Comiti F Ulloa H Huber A 2010 Large wood abundance dis-tribution and mobilization in a third order coastal mountain range river systemsouthern Chile For Ecol Manag 260 480ndash490 httpdxdoiorg101016jforeco201005004

Iroumeacute A Mao L Andreoli A Ulloa H Paz Ardiles M 2015 Large wood mobility pro-cesses in low-order Chilean river channels Geomorphology 228 681ndash693 httpdxdoiorg101016jgeomorph201410025

Jackson K Wohl E 2015 Instreamwood loads in montane forest streams of the Colora-do Front Range USA Geomorphology 234 161ndash170 httpdxdoiorg101016jgeomorph201501022

Jacobson PJ Jacobson KM Angermeier PL Cherry DS 1999 Transport retentionand ecological significance of woody debris within a large ephemeral river J NAm Benthol Soc 18 429ndash444 httpdxdoiorg1023071468376

Jochner M Turowski J Badoux A Stoffel M Rickli C 2015 The role of log jams andexceptional flood events in mobilizing coarse particulate organic matter in a steepheadwater stream Earth Surf Dyn 3 311 httpdxdoiorg105194esurf-3-311-2015

Johnson SL Swanson FJ Grant GE Wondzell SM 2000 Riparian forest disturbancesby a mountain flood - the influence of floated wood Hydrol Process 14 3031ndash3050

Kail J Hering D Muhar S Gerhard M Preis S 2007 The use of large wood in streamrestoration experiences from 50 projects in Germany and Austria J Appl Ecol 441145ndash1155 httpdxdoiorg101111j1365-2664200701401x

Keim F Skaugset E Bateman S 2000 Dynamics of coarse woody debris placed in threeOregon streams For Sci 46 13ndash22

Keller EA Swanson FJ 1979 Effects of large organic material on channel form and flu-vial processes Earth Surf Process Landf 4 361ndash380 httpdxdoiorg101002esp3290040406

Kindle E 1919 Notes on sedimenation in the Mackenzie River basin J Geol 26341ndash360

Kindle E 1921 Mackenzie River driftwood Geogr Rev 11 50ndash53King L Hassan MA Wei X Burge L Chen X 2013Wood dynamics in upland streams

under different disturbance regimes Earth Surf Process Landf 38 1197ndash1209httpdxdoiorg101002esp3356

Knighton D 1998 Fluvial Forms and Processes A New Perspective Hodder ArnoldLondon

Kramer N Wohl E 2014 Estimating fluvial wood discharge using time-lapse photogra-phy with varying sampling intervals Earth Surf Process Landf 39 844ndash852 httpdxdoiorg101002esp3540

Kramer N Wohl E 2015 Driftcretions the legacy impacts of driftwood on shorelinemorphology Geophys Res Lett 42 5855ndash5864 httpdxdoiorg1010022015GL064441

Kramer N Wohl EE Harry DL 2012 Using ground penetrating radar to unearth bur-ied beaver dams Geology 40 43ndash46 httpdxdoiorg101130G326821

Kramer N Wohl E Hess-Homeier B Leisz S 2016 The pulse of driftwood at multipletimescales in a great northern river Water Resour Res (in press)

Kuhnle R 2013 Suspended load In Wohl E (Ed) Treatise on Fluvial GeomorphologyAcademic Press San Diego pp 124ndash136

Lane E 1955 The importance of fluvial morphology in hydraulic engineering J HydraulAm Soc Civ Eng 81 1ndash17

Larrantildeaga S Diacuteez JR Elosegi A Pozo J 2003 Leaf retention in streams of the Aguumlerabasin (northern Spain) Aquat Sci 65 158ndash166 httpdxdoiorg101007s00027-003-0623-3

Lassettre NS Kondolf GM 2003 Process based management of large woody debris atthe basin scale Soquel Creek California Technical Report Department of LandscapeArchitecture and Environmental Planning Berkeley California Prepared forCalifornia Department of Forestry and Fire Protection and Soquel DemonstrationState Forest

Latterell JJ Naiman RJ 2007 Sources and dynamics of large logs in a temperate flood-plain river Ecol Appl 17 1127ndash1141 httpdxdoiorg10189006-0963

Le Lay YF Moulin B Pieacutegay H 2013Wood entrance deposition transfer and effects onfluvial forms and processes problem statements and challenging issues In ShroderJ (Ed) Treatise on GeomorphologyEcogeomorphology vol 12 Academic Presspp 20ndash36

Leopold LB Maddock Jr T 1953 The hydraulic geometry of stream channels and somephysiographic implications Technical Report US Geological Survey ProfessionalPaper 252

Lieacutebault F Pieacutegay H 2002 Causes of 20th century channel narrowing in mountain andpiedmont rivers of southeastern France Earth Surf Process Landf 27 425ndash444httpdxdoiorg101002esp328

Lienkaemper GW Swanson FJ 1987 Dynamics of large woody debris in streams inold-growth douglas-fir forests Can J For Res 17 150ndash156 httpdxdoiorg101139x87-027



Lisle TE Smith B 2003 Dynamic transport capacity in gravel-bed river systems InAraya T Kuroki M Marutani T (Eds) International Workshop for Source to SinkSedimentary Dynamics in Catchment Scale Organizing Committee of the Internation-al Workshop for Sedimentary Dynamics Sapporo Japan pp 16ndash20 (httpwwwtreesearchfsfeduspubs7831)

Luciacutea A Comiti F Borga M Cavalli M Marchi L 2015 Dynamics of large wood duringa flash flood in two mountain catchments Nat Hazards Earth Syst Sci 151741ndash1755 httpdxdoiorg105194nhess-15-1741-2015

Luppi L Rinaldi M Teruggi LB Darby SE Nardi L 2009 Monitoring and numericalmodelling of riverbank erosion processes a case study along the Cecina River (centralItaly) Earth Surf Process Landf 34 530ndash546 httpdxdoiorg101002esp1754

Mackin JH 1948 Concept of the graded river Geol Soc Am Bull 59 463ndash512MacVicar B Pieacutegay H 2012 Implementation and validation of video monitoring for

wood budgeting in a wandering piedmont river the Ain River (France) Earth SurfProcess Landf 37 1272ndash1289 httpdxdoiorg101002esp3240

MacVicar B Pieacutegay H Henderson A Comiti F Oberlin C Pecorari E 2009 Quantify-ing the temporal dynamics of wood in large rivers field trials of wood surveying dat-ing tracking and monitoring techniques Earth Surf Process Landf 34 2031ndash2046httpdxdoiorg101002esp1888

Mao L Andreoli A Iroumeacute A Comiti F Lenzi M 2013 Dynamics andmanagement al-ternatives of in-channel large wood in mountain basins of the southern AndesBosque (Valdivia) 34 319ndash330

Marcus WA Marston RA Colvard CRJ Gray RD 2002 Mapping the spatial and tem-poral distributions of woody debris in streams of the Greater Yellowstone EcosystemUSA Geomorphology 44 323ndash335 httpdxdoiorg101016S0169-555X(01)00181-7 (geomorphology on large rivers)

Martin DJ Benda LE 2001 Patterns of instreamwood recruitment and transport at thewatershed scale Trans Am Fish Soc 130 940ndash958 httpdxdoiorg1015771548-8659(2001)130b0940POIWRAN20CO2

Massong TM Montgomery DR 2000 Influence of sediment supply lithology andwood debris on the distribution of bedrock and alluvial channels Geol Soc AmBull 112 591ndash599

Mazzorana B Comiti F Scherer C Fuchs S 2012 Developing consistent scenarios toassess flood hazards in mountain streams J Environ Manag 94 112ndash124 httpdxdoiorg101016jjenvman201106030

Merten E Finlay J Johnson L Newman R Stefan H Vondracek B 2010 Factorsinfluencing wood mobilization in streams Water Resour Res 46

Merten EC Finlay J Johnson L Newman R Stefan H Vondracek B 2011 Environ-mental controls of wood entrapment in upper Midwestern streams Hydrol Process25 593ndash602 httpdxdoiorg101002hyp7846

Merten EC Vaz PG Decker-Fritz JA Finlay JC Stefan HG 2013 Relative importanceof breakage and decay as processes depleting large wood from streams Geomorphol-ogy 190 40ndash47 httpdxdoiorg101016jgeomorph201302006

Millington CE Sear DA 2007 Impacts of river restoration on small-wood dynamics in alow-gradient headwater stream Earth Surf Process Landf 32 1204ndash1218 httpdxdoiorg101002esp1552

Montgomery DR Abbe TB 2006 Influence of logjam-formed hard points on the formationof valley-bottom landforms in an old-growth forest valley Queets river WashingtonUSA Quat Res 65 147ndash155 httpdxdoiorg101016jyqres200510003

Montgomery DR Pieacutegay H 2003 Wood in rivers interactions with channel morphol-ogy and processes Geomorphology 51 1ndash5 httpdxdoiorg101016S0169-555X(02)00322-7 (interactions between Wood and Channel Forms and Processes)

Moulin B Pieacutegay H 2004 Characteristics and temporal variability of large woodydebris trapped in a reservoir on the River Rhone (Rhone) implications for riverbasin management River Res Appl 20 79ndash97 httpdxdoiorg101002rra724

Moulin B Schenk ER Hupp CR 2011 Distribution and characterization of in-channellarge wood in relation to geomorphic patterns on a low-gradient river Earth SurfProcess Landf 36 1137ndash1151 httpdxdoiorg101002esp2135

Murphy ML Koski KV 1989 Input and depletion of woody debris in alaska streamsand implications for streamside management N Am J Fish Manag 9 427ndash436

Naiman RJ Balian EV Bartz KK Bilby RE Latterell JJ 2002 Dead wood dynamics instream ecosystems Proceedings of the Symposium on the Ecology and Managementof Dead Wood in Western Forests USDA Forest Service Gen Tech Rep PSW-GTR-181 pp 23ndash48

Nakamura F Swanson FJ 1993 Effects of coarse woody debris on morphology and sed-iment storage of a mountain stream system in western Oregon Earth Surf ProcessLandf 18 43ndash61 httpdxdoiorg101002esp3290180104

OConnor J Jones M Haluska T 2003 Flood plain and channel dynamics of the Qui-nault and Queets Rivers Washington USA Geomorphology 31ndash59 httpdxdoiorg101016S0169-555X(02)00324-0

Opperman JJ Meleason M Francis RA Davies-Colley R 2008 ldquoLivewoodrdquo geomor-phic and ecological functions of living trees in river channels Bioscience 581069ndash1078 httpdxdoiorg101641B581110

Oswald EB Wohl E 2008 Wood-mediated geomorphic effects of a joumlkulhlaup in theWind River Mountains Wyoming Geomorphology 100 549ndash562 httpdxdoiorg101016jgeomorph200802002

Parker G Toro-Escobar CM 2002 Equalmobility of gravel in streams the remains of theday Water Resour Res 38 httpdxdoiorg1010292001WR000669 46-1ndash46-8

Pecorari E 2008 Il materiale legnoso in corsi dacqua a canali intrecciati volumimobilitaacute degradazione ed influenza morfologica (PhD thesis) Dept Land andAgroforest Environments University of Padova Italy

Pettit NE Latterell JJ Naiman RJ 2006 Formation distribution and ecological conse-quences of flood-related wood debris piles in a bedrock confined river in semi-aridSouth Africa River Res Appl 22 1097ndash1110 httpdxdoiorg101002rra959


Pettit NE Warfe DM Kennard MJ Pusey BJ Davies PM Douglas MM 2013 Dy-namics of in-streamwood and its importance as fish habitat in a large tropical flood-plain river River Res Appl 29 (7) 864ndash875 (Issn 1535ndash1467)

Phillips JD 2012 Log-jams and avulsions in the San Antonio River Delta Texas EarthSurf Process Landf 37 936ndash950 httpdxdoiorg101002esp3209

Phillips JD Park L 2009 Forest blowdown impacts of Hurricane Rita on fluvial systemsEarth Surf Process Landf 34 1069ndash1081 httpdxdoiorg101002esp1793

Pieacutegay H Gurnell A 1997 Large woody debris and river geomorphological pattern ex-amples from SE France and S England Geomorphology 19 99ndash116 httpdxdoiorg101016S0169-555X(96)00045-1

Pieacutegay H Le Lay YF Moulin B 2005 Les risques lieacutes aux embacirccles de bois dans lescours deau eacutetat des connaissances et principes de gestion In Vallauri D AndreacuteJ Dodelin B Eynard-Machet R Rambaud D (Eds) Bois mort et agrave caviteacute une cleacutepour des forecircts vivantes Lavoisier et Editions Tec amp Doc pp 193ndash202

Pizzuto J 2009 An empirical model of event scale cohesive bank profile evolution EarthSurf Process Landf 34 1234ndash1244 httpdxdoiorg101002esp1808

Polvi LE Wohl E 2013 Biotic drivers of stream planform implications for understand-ing the past and restoring the future Bioscience 63 439ndash452 httpdxdoiorg101525bio20136366

Ravazzolo D Mao L Garniga B Picco L Lenzi AM 2015a Volume and travel distanceof wood pieces in the Tagliamento River (Northeastern Italy) Engineering Geologyfor Society and Territory 3 Springer pp 135ndash138

Ravazzolo D Mao L Picco L Lenzi M 2015b Tracking log displacement during floodsin the Tagliamento River using RFID and GPS tracker devices Geomorphology 228226ndash233 httpdxdoiorg101016jgeomorph201409012

Rinaldi M Mengoni B Luppi L Darby SE Mosselman E 2008 Numerical simulationof hydrodynamics and bank erosion in a river bend Water Resour Res 44 httpdxdoiorg1010292008WR007008

Ruiz-Villanueva V Bodoque J Diacuteez-Herrero A Eguibar M Pardo-Iguacutezquiza E 2013Reconstruction of a flash flood with large wood transport and its influence on hazardpatterns in an ungauged mountain basin Hydrol Process 27 3424ndash3437 httpdxdoiorg101002hyp9433

Ruiz-Villanueva V Bodoque J Dez-Herrero A Bladeacute E 2014a b Large wood transportas significant influence on flood risk in a mountain village Nat Hazards 74 967ndash987

Ruiz-Villanueva V Bladeacute Castellet E Diacuteez-Herrero A Bodoque JM Saacutenchez-Juny M2014b Two-dimensional modelling of large wood transport during flash floodsEarth Surf Process Landf 39 438ndash449 httpdxdoiorg101002esp3546

Ruiz-Villanueva V Piegay H Stoffel M Gaertner V Perret F 2015a Analysis of wooddensity to improve understanding of wood buoyancy in rivers Engineering Geologyfor Society and Territory 3 Springer pp 163ndash166

Ruiz-Villanueva V Wyżga B Hajdukiewicz H Stoffel M 2015b Exploring large woodretention and deposition in contrasting river morphologies linking numerical model-ling and field observations Earth Surf Process Landf httpdxdoiorg101002esp3832

Ruiz-Villanueva V Wyżga B Zawiejska J Hajdukiewicz M Stoffel M 2015c Factorscontrolling large-wood transport in a mountain river Geomorphology httpdxdoiorg101016jgeomorph201504004

Ruiz-Villanueva V Pieacutegay H Gaertner V Perret F Stoffel M 2016aWood density andmoisture sorption and its influence on large wood mobility in rivers Catena 140182ndash194 httpdxdoiorg101016jcatena2016020010341-8162

Ruiz-Villanueva V Wyżga B Mikuś P Hajdukiewicz H Stoffel M 2016b The role offlood hydrograph in the remobilization of large wood in a wide mountain riverJ Hydrol httpdxdoiorg101016jjhydrol201602060

Schenk ER Moulin B Hupp CR Richter JM 2014 Large wood budget and transportdynamics on a large river using radio telemetry Earth Surf Process Landf 39487ndash498 httpdxdoiorg101002esp3463

Schmocker L Hager WH 2011 Probability of drift blockage at bridge decks J HydraulEng 137 470ndash479 httpdxdoiorg101061(ASCE)HY1943-79000000319

Schumm SA 1977 The Fluvial System vol 338 Wiley New YorkSear DA Millington CE R KD Jeffries R 2010 Logjam controls on channel flood-

plain interactions in wooded catchments and their role in the formation of multi-channel patterns Geomorphology 116 305ndash319 httpdxdoiorg101016jgeomorph200911022

Sedell JR Bisson PA Swanson FJ Gregory SV 1988 What we know about large treesthat fall into streams and rivers In Maser C Tarrant R Trappe J Franklin J (Eds)From the Forest to the Sea a Story of Fallen Trees United States Department of Agri-culture PNW Research Station Portland OR pp 47ndash81 (For Ser Gen Tech RepGTR-PNW-229)

Seo JI Nakamura F 2009 Scale-dependent controls upon the fluvial export of largewood from river catchments Earth Surf Process Landf 34 786ndash800 httpdxdoiorg101002esp1765

Seo JI Nakamura F Nakano D Ichiyanagi H Chun KW 2008 Factors controlling thefluvial export of large woody debris and its contribution to organic carbon budgets atwatershed scales Water Resour Res 44 httpdxdoiorg1010292007WR006453

Shields FD Morin N Cooper CM 2004 Large woody debris structures for sand-bedchannels J Hydraul Eng 130 208ndash217 httpdxdoiorg101061(ASCE)0733-9429(2004)1303(208)

Shields FD Knight SS Stofleth JM 2006 Large wood addition for aquatic habitat re-habilitation in an incised sand-bed stream Little Topashaw Creek MississippiRiver Res Appl 22 803ndash817 httpdxdoiorg101002rra937

Shields FD Pezeshki SR Wilson GV Wu W Dabney SM 2008 Rehabilitation of anincised stream using plant materials the dominance of geomorphic processes EcolSoc 13 (3) 54 httpwwwecologyandsocietyorgvol13iss2art54

Silvertown J 2009 A new dawn for citizen science Trends Ecol Evol 24 467ndash471httpdxdoiorg101016jtree200903017



Triska F 1984 Role of wood debris in modifying channel geomorphology and riparianareas of a large lowland river under pristine conditions a historical case studyVerh Int Ver Limnol 22 1876ndash1892

Turowski J Badoux A Bunte K Rickli C Federspiel N Jochner M 2013 Themass dis-tribution of coarse particulate organic matter exported from an Alpine headwaterstream Earth Surf Dyn 1 1 httpdxdoiorg105194esurf-1-1-2013

Ulloa H Iroumeacute A Lenzi MA Andreoli A Aacutelvarez C Barrera V 2011 Material lentildeosode gran tamantildeo en dos cuencas de la cordillera de la costa de chile con diferentehistoria de uso del suelo Bosque (Valdivia) 32 235ndash245

Ulloa H Iroumeacute A Mao L Andreoli A Diez S Lara LE 2015 Use of remote imageryto analyse changes in morphology and longitudinal large wood distribution in theBlanco River after the 2008 Chaiteacuten volcanic eruption southern Chile Geogr AnnSer A 97 523ndash541 httpdxdoiorg101111geoa12091

Valdebenito G Tonon A Iroume A Alvarado D Fuentes C Picco L Mario L 2016 3_d modeling using tls and gpr techniques to characterize above and below-groundwood distribution in pyroclastic deposits along the Blanco River (Chilean Patagonia)EGU General Assembly

van der Nat D Tockner K Edwards PJ Ward J 2003 Largewood dynamics of complexalpine river floodplains J N Am Benthol Soc 22 35ndash50 httpdxdoiorg1023071467976

Warren DR Kraft CE 2008 Dynamics of large wood in an eastern US mountainstream For Ecol Manag 256 808ndash814 httpdxdoiorg101016jforeco200805038

Welber M Bertoldi W Tubino M 2013 Wood dispersal in braided streams resultsfrom physical modeling Water Resour Res 49 7388ndash7400 httpdxdoiorg1010022013WR014046

West A Lin C Lin T Hilton R Liu S Chang C Lin K Galy A Sparkes R Hovius N2011 Mobilization and transport of coarse woody debris to the oceans triggered byan extreme tropical storm Limnol Oceanogr 56 77ndash85

White L Hodges BR 2003 Identification of large woody debris in acoustic bathymetrydata Technical report Center for Research inWater Resources University of Texas atAustin Final Report to Texas Water Development Board under Contract No 2003-469 CRWR Online Report pp 03ndash07


Wohl E 2011 Seeing the forest and the trees wood in stream restoration in the Colora-do Front Range Stream Restoration in Dynamic Fluvial Systems Scientific Ap-proaches Analyses and Tools American Geophysical Union Press Washington DCpp 399ndash418

Wohl E 2014a A legacy of absence wood removal in US rivers Prog Phys Geogr 38637ndash663 httpdxdoiorg1011770309133314548091

Wohl E 2014b Time and the rivers flowing fluvial geomorphology since 1960 Geomor-phology 216 263ndash282

Wohl E 2016 Bridging the gaps wood across time and space in diverse riversGeomorphology

Wohl E Beckman N 2014 Controls on the longitudinal distribution of channel-span-ning logjams in the Colorado Front Range USA River Res Appl 30 112ndash131httpdxdoiorg101002rra2624

Wohl E Cadol D 2011 Neighborhood matters Patterns and controls on wood distribu-tion in old-growth forest streams of the Colorado Front Range USA Geomorphology125 132ndash146 httpdxdoiorg101016jgeomorph201009008

Wohl E Goode JR 2008 Wood dynamics in headwater streams of the Colorado RockyMountains Water Resour Res 44 httpdxdoiorg1010292007WR006522

Wohl E Polvi LE Cadol D 2011 Wood distribution along streams draining old-growthfloodplain forests in Congaree National Park South Carolina USA Geomorphology126 108ndash120 httpdxdoiorg101016jgeomorph201010035

Wohl E Dwire K Sutfin N Polvi L Bazan R 2012 Mechanisms of carbon storage inmountainous headwater rivers Nat Commun 3 1263 (doi 101038 ncomms2274)

Wohl E Lane SNWilcox AC 2015 The science and practice of river restorationWaterResour Res 51 5974ndash5997

Wyżga B Zawiejska J 2005 Wood storage in a wide mountain river case study of theCzarny Dunajec Polish Carpathians Earth Surf Process Landf 30 1475ndash1494 httpdxdoiorg101002esp1204

Young MK 1994 Movement and characteristics of stream-borne coarse woody debris inadjacent burned and undisturbed watersheds in Wyoming Can J For Res 241933ndash1938 httpdxdoiorg101139x94-248


Rules of the road A qualitative and quantitative synthesis of large wood transport through drainage networks

1 Motivation






31 Methods

32 Results

4 Discussion


411 Storage patterns and transport processes






47 Moving forward

5 Conclusion

Acknowledgements


A1 Transport distance based on exponential scaling by reach retention rates (Millington and Sear 2007)

A2 Transport distance as an exponential function of water depth at peak flow (Haga et al 2002)

A3 Transport distance as a function of jam spacing (Martin and Benda 2001)

A4 Stochastic models of travel distance and mobility of individual pieces (Lassettre and Kondolf 2003)

A5 Empirical logistic model of key piece mobility based on wood characteristics (Wohl and Goode 2008)

A6 Empirical linear relationships for percent mobility based on flow characteristics (Iroumeacute et al 2015)

A7 Empirical exponential relationship between wood volume and wood velocity (Ravazzolo et al 2015b)

A8 Logistic model of mobilization based on wood characteristics (Merten et al 2010)

A9 Transport ratio as a function of wood characteristics and discharge for single thread (TrS) versus multithread (TrM) r


References

Table 1Functional classification of river size based on wood dynamics Categorization based oncharacteristic dimensions of wood pieces relative to channel size and patterns of recruit-ment storage and transport of wood after Keller and Swanson (1979) Church (1992)Nakamura and Swanson (1993) Pieacutegay and Gurnell (1997) Gurnell et al (2002) andWohl (2016 this issue)

Small rivers

Size 1st to 2nd order key piece wood length N channel width diameterof logs N flow depth

Recruitment Characteristics of riparian woodland of overriding importancehillslope instability blowdown snow avalanches individualriparian tree mortality

Storage Individual wood pieces form stable features that control sedimentdeposition channel morphology and gradient pieces typically spanand are suspended above the active channel jams may be presentbut most wood stored as individual pieces distribution of woodclose to random and governed by sites of input wood causes localwidening where channel boundaries are erodible wood conditiondepends on input mechanisms and forest disturbance history lowmobility can result in high levels of wood decay

Transport Wood not reorganized or transported except during unusual floodsor debris flows with recurrence intervals N decadal large piecesmostly immobile long residence times regulated by decay andphysical breakdown rather than fluvial transport

Medium rivers

Size 2nd to 4th order log diameter ~ flow depth length of key logs N or ~channel width

Recruitment Hillslope riparian and bank source areas tributary inputstransport from upstream

Storage Wood commonly stored in non-random spatially discontinuousjams that partly or completely span the channel jams comprised ofmixed sizes of wood pieces with larger key pieces trapping smallermore mobile pieces jams form at roughness elements such asboulders and planform irregularities jams force bank erosion andavulsion can create multithread planform high inter-reachvariation in wood loads wood regulates sediment transport

Transport Hydrologic regime is dominant control drives periodic transport ofstored and newly recruited pieces during high flows key pieces andjams remain in place during smaller floods accumulating smallerpieces larger more infrequent floods break up and rearrange jams

Large rivers

Size ge5th order log diameter ≪ depth of flooded channel all woodlengths b channel width transition between medium and largerivers at channel widths ~20ndash50 m wide because longest woodpiece lengths are in this range

Recruitment Transport of wood from upstream lateral channel erosionexhumation of previously buried wood on floodplains

Storage River morphology dictates wood storage sites wood generally hasless decay because of greater mobility wood influences bar andchannel sedimentation and regulates formation of secondarychannels channel planform and widening of valley floor reaches


1948) and Lanes balance (Lane 1955) and quantitative understandingof river systems - such as hydraulic geometry relations (Leopold andMaddock 1953) The large-scale removal of wood from rivers duringthe 20th century (Montgomery and Pieacutegay 2003 Wohl 2014a) likelyled to a neglect of the geomorphic effects of wood on river processand form during this formative period of fluvial geomorphology

As long as rivers are wood-depleted models of river form adjustingprimarily to sediment and water are adequate for managing riversHowever in order to improve valued ecosystem services such as de-ni-trification sustainable fisheries improved water quality and enhancedphysical and mental health of human communities (eg Wohl et al2015)managers now focus on reintroducingwood as engineered struc-tures (Abbe and Brooks 2011 Gallisdorfer et al 2014) leaving mobileun-engineered wood in place on the floodplain (Pieacutegay et al 2005Wohl et al 2015) andmanaging riparian forests to increase the amountof wood that can be recruited (Kail et al 2007 Wohl et al 2015)

In Europe afforestation of river corridors (Lieacutebault and Pieacutegay 2002)has led to increased wood in transport resulting in wood impound-ments against bridge piers during floods (eg Luciacutea et al 2015) in-creasing flood levels and forcing large-scale channel changes beyondwhat can be predicted through existing models (Pieacutegay et al 2005Ruiz-Villanueva et al 2014b) Inability to plan for the impacts and haz-ards of large wood on channels is a management concern and stemslargely from the fact that current models of river form and processused by managers do not include wood dynamics

In the past few years substantial progress in numerical simulation ofwood transport has come from flume studies (Bertoldi et al 2014Davidson et al 2015) and incorporation of woodmodules into comput-er simulations such as the reach scale channel simulator for habitatmodelling (Eaton et al 2012 Davidson and Eaton 2015) and iberwood for hydrodynamic modelling of wood transport (Ruiz-Villanuevaet al 2014a 2015c) Simulations from these models have led to betterunderstanding of wood-related flooding hazards channel change andaquatic habitat Howevermodels are simplified versions of the complexinteractions found in the field and they require information from fieldstudies to constrain variables and assess reliability of scenario responses(Ruiz-Villanueva et al 2015b) Recent rapid growth in numerical andmodelling publications onwood transport is notmatched by equivalentgrowth in field studies (Fig 1)

In this paper we first qualitatively and then quantitatively summa-rize analyze and synthesize literature on wood transport within theframework of a functional classification of small medium large andgreat rivers (Table 1 Fig 2) We focus on presenting a thorough reviewof case studies that have directly measured wood transport We exam-ine existing transport premises present new conceptual frameworks

Flume or Modelling Studies

Field Studies

1

2

7

5

1

2

4

Fig 1 Comparison of field and modelling publications that report direct measurement ofwood transport by year Bar heights equal total number of articles found

with greater sinuosity more bars and lower gradients typicallyhave higher wood loads wood accumulates within active channelat sites such as apex of bars outer upstream facing margins ofchannel bends backwaters bank benches and as individual piecesalong channel margins during flood recession large proportion ofwood may be buried in floodplains and channel bars woodaccumulates on upstream side of living vegetation and influencesmorphologic evolution of stable vegetated islands and formation ofmultithread planform wood does not form channel spanning jamsbut can form channel spanning dynamic wood rafts that persist fordecades and cause channel avulsions

Transport Wood exported downstream laterally onto floodplains or buriedwood exported regularly during high flow amount of wood transferhighly variable and largely dependent on pattern of antecedentpeak flows high variability in water levels during flooding createsmany opportunities for wood sequestration in long-term storage onthe floodplain causing large variability in wood residence time

Great riversa

Size ~106 km2 or larger mean annual discharge N 103 m3s perennialflow commonly have vast seasonally inundated floodplains greathydraulic diversity large fine sediment loads and deep channels

Recruitment Mostly via inputs from upstream local inputs include lateralexchange with floodplain exhumation of buried wood inputs from


Table 1 (continued)

Small rivers
















Magnitude









Duration









Flow history

















Anchoring






Length












Orientation










Table 3 (continued)












Channel morphology













Live-wood






















































































NorthDakota

















31 Methods
















32 Results








~80 high flow

max mobility

threshold

Bertoldi et

al(2013)

Van der Nat

et al(2007)MacVicar

and Pieacutegay

(2009)

mobility envelope

remobilized

redeposited

repositioned

Fraction bankfull

M

ob

ile













-100 -50 0 50

6080

4020

0

100

~30 low flow

maximum mobility

threshold


(2009)

mobility

envelope

threshold

T

ota

l m

ob

ility


~80 high flow

maximum mobility



Haga et al

(2002)

Millington and

Sear (2007)

MacVicar

et al (2009)

Ulloa et al (2015)


MacVicar

et al (2012)

Van der Nat

et al (2003)

Small

Medium

Large Large

amp Great

Bertoldi et

al(2013)

Van der Nat

et al (2003)

Bertoldi

et al(2013)

Ulloa et al

(2015)

MacVicar et al

(2009)

high mobility from

bull extreme events



changes



stable channels

mobility envelope

remobilized

redeposited

repositioned

M

ob

ile

M

ob

ile

Stream width (m)

Stream width (m)

Stream width (m)

Tra

ve

l d

ista

nce (m

)

L

L

L









Lgt5 1leL le 5

Small Medium

Llt1

Large

Large

Braided

Med=9

Med=14

Re-m

obiliz

ation

ra

te (

)

Med=32

Med=80

Min=25

Max=95

Max=83Max=80

Max=33

n=45 n=5n=23 n=136

020

40

60

80

100






4 Discussion


















Time

ET

Amount

wood in

storage

(S)

+∆S

RT RT

+∆S








ET

ET= episodic

transfer

ETRT

-∆S +∆S ∆S=0

RT=regular

transfer

equilibrium



moving in

contact with

channel deflecting

against channel

boundaries

0

of time

in contact

with

channel

100

unimpeded

floating

0

of

time fully

floating

100


i)

1

immobile

partially

mobile

fully mobile

QtQ

bf

0

proportion

of stored

wood

1

341















Small Rivers

Medium

Rivers

Large Rivers

Great Rivers

Residence Time (Tr)

Travel

Distance

(Xw)

lateral

export

basin

export





















Rising

Limb

gt~34 bf

le bf

lt1

lt1

Wood Flow and C

Governing Transp

hydrograph position

Falling

Limb

stage height

gt34 bf

gt1



for Transport

asymp1

gt1

wood length

divide

flow width fl

Deposi

Domina


stop go





47 Moving forward



lowlimited noneFew

High


ort Variability

transport

congestion

high

degree of

flow

obstruction

Complete

partial

living

vegetation

in flow

lots

some

flow depths

lt

otation depth

most

partial


ModLow

semi

tion

ted

Transport

Dominated


stop

go


















5 Conclusion


Acknowledgements







PdP0eminuskd






ξ frac14 L jTp

T jβminus1




1thorn em1thornm2

Lwbkf















7



















OR


OR







References


















































































































































































1 Motivation






31 Methods

32 Results

4 Discussion








47 Moving forward

5 Conclusion

Acknowledgements












References

Table 1 (continued)

Small rivers
















Magnitude









Duration









Flow history

















Anchoring






Length












Orientation










Table 3 (continued)












Channel morphology













Live-wood






















































































NorthDakota

















31 Methods
















32 Results








~80 high flow

max mobility

threshold

Bertoldi et

al(2013)

Van der Nat

et al(2007)MacVicar

and Pieacutegay

(2009)

mobility envelope

remobilized

redeposited

repositioned

Fraction bankfull

M

ob

ile













-100 -50 0 50

6080

4020

0

100

~30 low flow

maximum mobility

threshold


(2009)

mobility

envelope

threshold

T

ota

l m

ob

ility


~80 high flow

maximum mobility



Haga et al

(2002)

Millington and

Sear (2007)

MacVicar

et al (2009)

Ulloa et al (2015)


MacVicar

et al (2012)

Van der Nat

et al (2003)

Small

Medium

Large Large

amp Great

Bertoldi et

al(2013)

Van der Nat

et al (2003)

Bertoldi

et al(2013)

Ulloa et al

(2015)

MacVicar et al

(2009)

high mobility from

bull extreme events



changes



stable channels

mobility envelope

remobilized

redeposited

repositioned

M

ob

ile

M

ob

ile

Stream width (m)

Stream width (m)

Stream width (m)

Tra

ve

l d

ista

nce (m

)

L

L

L









Lgt5 1leL le 5

Small Medium

Llt1

Large

Large

Braided

Med=9

Med=14

Re-m

obiliz

ation

ra

te (

)

Med=32

Med=80

Min=25

Max=95

Max=83Max=80

Max=33

n=45 n=5n=23 n=136

020

40

60

80

100






4 Discussion


















Time

ET

Amount

wood in

storage

(S)

+∆S

RT RT

+∆S








ET

ET= episodic

transfer

ETRT

-∆S +∆S ∆S=0

RT=regular

transfer

equilibrium



moving in

contact with

channel deflecting

against channel

boundaries

0

of time

in contact

with

channel

100

unimpeded

floating

0

of

time fully

floating

100


i)

1

immobile

partially

mobile

fully mobile

QtQ

bf

0

proportion

of stored

wood

1

341















Small Rivers

Medium

Rivers

Large Rivers

Great Rivers

Residence Time (Tr)

Travel

Distance

(Xw)

lateral

export

basin

export





















Rising

Limb

gt~34 bf

le bf

lt1

lt1

Wood Flow and C

Governing Transp

hydrograph position

Falling

Limb

stage height

gt34 bf

gt1



for Transport

asymp1

gt1

wood length

divide

flow width fl

Deposi

Domina


stop go





47 Moving forward



lowlimited noneFew

High


ort Variability

transport

congestion

high

degree of

flow

obstruction

Complete

partial

living

vegetation

in flow

lots

some

flow depths

lt

otation depth

most

partial


ModLow

semi

tion

ted

Transport

Dominated


stop

go


















5 Conclusion


Acknowledgements







PdP0eminuskd






ξ frac14 L jTp

T jβminus1




1thorn em1thornm2

Lwbkf















7



















OR


OR







References


















































































































































































1 Motivation






31 Methods

32 Results

4 Discussion








47 Moving forward

5 Conclusion

Acknowledgements












References



Magnitude









Duration









Flow history

















Anchoring






Length












Orientation










Table 3 (continued)












Channel morphology













Live-wood






















































































NorthDakota

















31 Methods
















32 Results








~80 high flow

max mobility

threshold

Bertoldi et

al(2013)

Van der Nat

et al(2007)MacVicar

and Pieacutegay

(2009)

mobility envelope

remobilized

redeposited

repositioned

Fraction bankfull

M

ob

ile













-100 -50 0 50

6080

4020

0

100

~30 low flow

maximum mobility

threshold


(2009)

mobility

envelope

threshold

T

ota

l m

ob

ility


~80 high flow

maximum mobility



Haga et al

(2002)

Millington and

Sear (2007)

MacVicar

et al (2009)

Ulloa et al (2015)


MacVicar

et al (2012)

Van der Nat

et al (2003)

Small

Medium

Large Large

amp Great

Bertoldi et

al(2013)

Van der Nat

et al (2003)

Bertoldi

et al(2013)

Ulloa et al

(2015)

MacVicar et al

(2009)

high mobility from

bull extreme events



changes



stable channels

mobility envelope

remobilized

redeposited

repositioned

M

ob

ile

M

ob

ile

Stream width (m)

Stream width (m)

Stream width (m)

Tra

ve

l d

ista

nce (m

)

L

L

L









Lgt5 1leL le 5

Small Medium

Llt1

Large

Large

Braided

Med=9

Med=14

Re-m

obiliz

ation

ra

te (

)

Med=32

Med=80

Min=25

Max=95

Max=83Max=80

Max=33

n=45 n=5n=23 n=136

020

40

60

80

100






4 Discussion


















Time

ET

Amount

wood in

storage

(S)

+∆S

RT RT

+∆S








ET

ET= episodic

transfer

ETRT

-∆S +∆S ∆S=0

RT=regular

transfer

equilibrium



moving in

contact with

channel deflecting

against channel

boundaries

0

of time

in contact

with

channel

100

unimpeded

floating

0

of

time fully

floating

100


i)

1

immobile

partially

mobile

fully mobile

QtQ

bf

0

proportion

of stored

wood

1

341















Small Rivers

Medium

Rivers

Large Rivers

Great Rivers

Residence Time (Tr)

Travel

Distance

(Xw)

lateral

export

basin

export





















Rising

Limb

gt~34 bf

le bf

lt1

lt1

Wood Flow and C

Governing Transp

hydrograph position

Falling

Limb

stage height

gt34 bf

gt1



for Transport

asymp1

gt1

wood length

divide

flow width fl

Deposi

Domina


stop go





47 Moving forward



lowlimited noneFew

High


ort Variability

transport

congestion

high

degree of

flow

obstruction

Complete

partial

living

vegetation

in flow

lots

some

flow depths

lt

otation depth

most

partial


ModLow

semi

tion

ted

Transport

Dominated


stop

go


















5 Conclusion


Acknowledgements







PdP0eminuskd






ξ frac14 L jTp

T jβminus1




1thorn em1thornm2

Lwbkf















7



















OR


OR







References


















































































































































































1 Motivation






31 Methods

32 Results

4 Discussion








47 Moving forward

5 Conclusion

Acknowledgements












References



Anchoring






Length












Orientation










Table 3 (continued)












Channel morphology













Live-wood






















































































NorthDakota

















31 Methods
















32 Results








~80 high flow

max mobility

threshold

Bertoldi et

al(2013)

Van der Nat

et al(2007)MacVicar

and Pieacutegay

(2009)

mobility envelope

remobilized

redeposited

repositioned

Fraction bankfull

M

ob

ile













-100 -50 0 50

6080

4020

0

100

~30 low flow

maximum mobility

threshold


(2009)

mobility

envelope

threshold

T

ota

l m

ob

ility


~80 high flow

maximum mobility



Haga et al

(2002)

Millington and

Sear (2007)

MacVicar

et al (2009)

Ulloa et al (2015)


MacVicar

et al (2012)

Van der Nat

et al (2003)

Small

Medium

Large Large

amp Great

Bertoldi et

al(2013)

Van der Nat

et al (2003)

Bertoldi

et al(2013)

Ulloa et al

(2015)

MacVicar et al

(2009)

high mobility from

bull extreme events



changes



stable channels

mobility envelope

remobilized

redeposited

repositioned

M

ob

ile

M

ob

ile

Stream width (m)

Stream width (m)

Stream width (m)

Tra

ve

l d

ista

nce (m

)

L

L

L









Lgt5 1leL le 5

Small Medium

Llt1

Large

Large

Braided

Med=9

Med=14

Re-m

obiliz

ation

ra

te (

)

Med=32

Med=80

Min=25

Max=95

Max=83Max=80

Max=33

n=45 n=5n=23 n=136

020

40

60

80

100






4 Discussion


















Time

ET

Amount

wood in

storage

(S)

+∆S

RT RT

+∆S








ET

ET= episodic

transfer

ETRT

-∆S +∆S ∆S=0

RT=regular

transfer

equilibrium



moving in

contact with

channel deflecting

against channel

boundaries

0

of time

in contact

with

channel

100

unimpeded

floating

0

of

time fully

floating

100


i)

1

immobile

partially

mobile

fully mobile

QtQ

bf

0

proportion

of stored

wood

1

341















Small Rivers

Medium

Rivers

Large Rivers

Great Rivers

Residence Time (Tr)

Travel

Distance

(Xw)

lateral

export

basin

export





















Rising

Limb

gt~34 bf

le bf

lt1

lt1

Wood Flow and C

Governing Transp

hydrograph position

Falling

Limb

stage height

gt34 bf

gt1



for Transport

asymp1

gt1

wood length

divide

flow width fl

Deposi

Domina


stop go





47 Moving forward



lowlimited noneFew

High


ort Variability

transport

congestion

high

degree of

flow

obstruction

Complete

partial

living

vegetation

in flow

lots

some

flow depths

lt

otation depth

most

partial


ModLow

semi

tion

ted

Transport

Dominated


stop

go


















5 Conclusion


Acknowledgements







PdP0eminuskd






ξ frac14 L jTp

T jβminus1




1thorn em1thornm2

Lwbkf















7



















OR


OR







References


















































































































































































1 Motivation






31 Methods

32 Results

4 Discussion








47 Moving forward

5 Conclusion

Acknowledgements












References



Channel morphology













Live-wood






















































































NorthDakota

















31 Methods
















32 Results








~80 high flow

max mobility

threshold

Bertoldi et

al(2013)

Van der Nat

et al(2007)MacVicar

and Pieacutegay

(2009)

mobility envelope

remobilized

redeposited

repositioned

Fraction bankfull

M

ob

ile













-100 -50 0 50

6080

4020

0

100

~30 low flow

maximum mobility

threshold


(2009)

mobility

envelope

threshold

T

ota

l m

ob

ility


~80 high flow

maximum mobility



Haga et al

(2002)

Millington and

Sear (2007)

MacVicar

et al (2009)

Ulloa et al (2015)


MacVicar

et al (2012)

Van der Nat

et al (2003)

Small

Medium

Large Large

amp Great

Bertoldi et

al(2013)

Van der Nat

et al (2003)

Bertoldi

et al(2013)

Ulloa et al

(2015)

MacVicar et al

(2009)

high mobility from

bull extreme events



changes



stable channels

mobility envelope

remobilized

redeposited

repositioned

M

ob

ile

M

ob

ile

Stream width (m)

Stream width (m)

Stream width (m)

Tra

ve

l d

ista

nce (m

)

L

L

L









Lgt5 1leL le 5

Small Medium

Llt1

Large

Large

Braided

Med=9

Med=14

Re-m

obiliz

ation

ra

te (

)

Med=32

Med=80

Min=25

Max=95

Max=83Max=80

Max=33

n=45 n=5n=23 n=136

020

40

60

80

100






4 Discussion


















Time

ET

Amount

wood in

storage

(S)

+∆S

RT RT

+∆S








ET

ET= episodic

transfer

ETRT

-∆S +∆S ∆S=0

RT=regular

transfer

equilibrium



moving in

contact with

channel deflecting

against channel

boundaries

0

of time

in contact

with

channel

100

unimpeded

floating

0

of

time fully

floating

100


i)

1

immobile

partially

mobile

fully mobile

QtQ

bf

0

proportion

of stored

wood

1

341















Small Rivers

Medium

Rivers

Large Rivers

Great Rivers

Residence Time (Tr)

Travel

Distance

(Xw)

lateral

export

basin

export





















Rising

Limb

gt~34 bf

le bf

lt1

lt1

Wood Flow and C

Governing Transp

hydrograph position

Falling

Limb

stage height

gt34 bf

gt1



for Transport

asymp1

gt1

wood length

divide

flow width fl

Deposi

Domina


stop go





47 Moving forward



lowlimited noneFew

High


ort Variability

transport

congestion

high

degree of

flow

obstruction

Complete

partial

living

vegetation

in flow

lots

some

flow depths

lt

otation depth

most

partial


ModLow

semi

tion

ted

Transport

Dominated


stop

go


















5 Conclusion


Acknowledgements







PdP0eminuskd






ξ frac14 L jTp

T jβminus1




1thorn em1thornm2

Lwbkf















7



















OR


OR







References


















































































































































































1 Motivation






31 Methods

32 Results

4 Discussion








47 Moving forward

5 Conclusion

Acknowledgements












References





























































NorthDakota

















31 Methods
















32 Results








~80 high flow

max mobility

threshold

Bertoldi et

al(2013)

Van der Nat

et al(2007)MacVicar

and Pieacutegay

(2009)

mobility envelope

remobilized

redeposited

repositioned

Fraction bankfull

M

ob

ile













-100 -50 0 50

6080

4020

0

100

~30 low flow

maximum mobility

threshold


(2009)

mobility

envelope

threshold

T

ota

l m

ob

ility


~80 high flow

maximum mobility



Haga et al

(2002)

Millington and

Sear (2007)

MacVicar

et al (2009)

Ulloa et al (2015)


MacVicar

et al (2012)

Van der Nat

et al (2003)

Small

Medium

Large Large

amp Great

Bertoldi et

al(2013)

Van der Nat

et al (2003)

Bertoldi

et al(2013)

Ulloa et al

(2015)

MacVicar et al

(2009)

high mobility from

bull extreme events



changes



stable channels

mobility envelope

remobilized

redeposited

repositioned

M

ob

ile

M

ob

ile

Stream width (m)

Stream width (m)

Stream width (m)

Tra

ve

l d

ista

nce (m

)

L

L

L









Lgt5 1leL le 5

Small Medium

Llt1

Large

Large

Braided

Med=9

Med=14

Re-m

obiliz

ation

ra

te (

)

Med=32

Med=80

Min=25

Max=95

Max=83Max=80

Max=33

n=45 n=5n=23 n=136

020

40

60

80

100






4 Discussion


















Time

ET

Amount

wood in

storage

(S)

+∆S

RT RT

+∆S








ET

ET= episodic

transfer

ETRT

-∆S +∆S ∆S=0

RT=regular

transfer

equilibrium



moving in

contact with

channel deflecting

against channel

boundaries

0

of time

in contact

with

channel

100

unimpeded

floating

0

of

time fully

floating

100


i)

1

immobile

partially

mobile

fully mobile

QtQ

bf

0

proportion

of stored

wood

1

341















Small Rivers

Medium

Rivers

Large Rivers

Great Rivers

Residence Time (Tr)

Travel

Distance

(Xw)

lateral

export

basin

export





















Rising

Limb

gt~34 bf

le bf

lt1

lt1

Wood Flow and C

Governing Transp

hydrograph position

Falling

Limb

stage height

gt34 bf

gt1



for Transport

asymp1

gt1

wood length

divide

flow width fl

Deposi

Domina


stop go





47 Moving forward



lowlimited noneFew

High


ort Variability

transport

congestion

high

degree of

flow

obstruction

Complete

partial

living

vegetation

in flow

lots

some

flow depths

lt

otation depth

most

partial


ModLow

semi

tion

ted

Transport

Dominated


stop

go


















5 Conclusion


Acknowledgements







PdP0eminuskd






ξ frac14 L jTp

T jβminus1




1thorn em1thornm2

Lwbkf















7



















OR


OR







References


















































































































































































1 Motivation






31 Methods

32 Results

4 Discussion








47 Moving forward

5 Conclusion

Acknowledgements












References








32 Results








~80 high flow

max mobility

threshold

Bertoldi et

al(2013)

Van der Nat

et al(2007)MacVicar

and Pieacutegay

(2009)

mobility envelope

remobilized

redeposited

repositioned

Fraction bankfull

M

ob

ile













-100 -50 0 50

6080

4020

0

100

~30 low flow

maximum mobility

threshold


(2009)

mobility

envelope

threshold

T

ota

l m

ob

ility


~80 high flow

maximum mobility



Haga et al

(2002)

Millington and

Sear (2007)

MacVicar

et al (2009)

Ulloa et al (2015)


MacVicar

et al (2012)

Van der Nat

et al (2003)

Small

Medium

Large Large

amp Great

Bertoldi et

al(2013)

Van der Nat

et al (2003)

Bertoldi

et al(2013)

Ulloa et al

(2015)

MacVicar et al

(2009)

high mobility from

bull extreme events



changes



stable channels

mobility envelope

remobilized

redeposited

repositioned

M

ob

ile

M

ob

ile

Stream width (m)

Stream width (m)

Stream width (m)

Tra

ve

l d

ista

nce (m

)

L

L

L









Lgt5 1leL le 5

Small Medium

Llt1

Large

Large

Braided

Med=9

Med=14

Re-m

obiliz

ation

ra

te (

)

Med=32

Med=80

Min=25

Max=95

Max=83Max=80

Max=33

n=45 n=5n=23 n=136

020

40

60

80

100






4 Discussion


















Time

ET

Amount

wood in

storage

(S)

+∆S

RT RT

+∆S








ET

ET= episodic

transfer

ETRT

-∆S +∆S ∆S=0

RT=regular

transfer

equilibrium



moving in

contact with

channel deflecting

against channel

boundaries

0

of time

in contact

with

channel

100

unimpeded

floating

0

of

time fully

floating

100


i)

1

immobile

partially

mobile

fully mobile

QtQ

bf

0

proportion

of stored

wood

1

341















Small Rivers

Medium

Rivers

Large Rivers

Great Rivers

Residence Time (Tr)

Travel

Distance

(Xw)

lateral

export

basin

export





















Rising

Limb

gt~34 bf

le bf

lt1

lt1

Wood Flow and C

Governing Transp

hydrograph position

Falling

Limb

stage height

gt34 bf

gt1



for Transport

asymp1

gt1

wood length

divide

flow width fl

Deposi

Domina


stop go





47 Moving forward



lowlimited noneFew

High


ort Variability

transport

congestion

high

degree of

flow

obstruction

Complete

partial

living

vegetation

in flow

lots

some

flow depths

lt

otation depth

most

partial


ModLow

semi

tion

ted

Transport

Dominated


stop

go


















5 Conclusion


Acknowledgements







PdP0eminuskd






ξ frac14 L jTp

T jβminus1




1thorn em1thornm2

Lwbkf















7



















OR


OR







References


















































































































































































1 Motivation






31 Methods

32 Results

4 Discussion








47 Moving forward

5 Conclusion

Acknowledgements












References



~80 high flow

max mobility

threshold

Bertoldi et

al(2013)

Van der Nat

et al(2007)MacVicar

and Pieacutegay

(2009)

mobility envelope

remobilized

redeposited

repositioned

Fraction bankfull

M

ob

ile













-100 -50 0 50

6080

4020

0

100

~30 low flow

maximum mobility

threshold


(2009)

mobility

envelope

threshold

T

ota

l m

ob

ility


~80 high flow

maximum mobility



Haga et al

(2002)

Millington and

Sear (2007)

MacVicar

et al (2009)

Ulloa et al (2015)


MacVicar

et al (2012)

Van der Nat

et al (2003)

Small

Medium

Large Large

amp Great

Bertoldi et

al(2013)

Van der Nat

et al (2003)

Bertoldi

et al(2013)

Ulloa et al

(2015)

MacVicar et al

(2009)

high mobility from

bull extreme events



changes



stable channels

mobility envelope

remobilized

redeposited

repositioned

M

ob

ile

M

ob

ile

Stream width (m)

Stream width (m)

Stream width (m)

Tra

ve

l d

ista

nce (m

)

L

L

L









Lgt5 1leL le 5

Small Medium

Llt1

Large

Large

Braided

Med=9

Med=14

Re-m

obiliz

ation

ra

te (

)

Med=32

Med=80

Min=25

Max=95

Max=83Max=80

Max=33

n=45 n=5n=23 n=136

020

40

60

80

100






4 Discussion


















Time

ET

Amount

wood in

storage

(S)

+∆S

RT RT

+∆S








ET

ET= episodic

transfer

ETRT

-∆S +∆S ∆S=0

RT=regular

transfer

equilibrium



moving in

contact with

channel deflecting

against channel

boundaries

0

of time

in contact

with

channel

100

unimpeded

floating

0

of

time fully

floating

100


i)

1

immobile

partially

mobile

fully mobile

QtQ

bf

0

proportion

of stored

wood

1

341















Small Rivers

Medium

Rivers

Large Rivers

Great Rivers

Residence Time (Tr)

Travel

Distance

(Xw)

lateral

export

basin

export





















Rising

Limb

gt~34 bf

le bf

lt1

lt1

Wood Flow and C

Governing Transp

hydrograph position

Falling

Limb

stage height

gt34 bf

gt1



for Transport

asymp1

gt1

wood length

divide

flow width fl

Deposi

Domina


stop go





47 Moving forward



lowlimited noneFew

High


ort Variability

transport

congestion

high

degree of

flow

obstruction

Complete

partial

living

vegetation

in flow

lots

some

flow depths

lt

otation depth

most

partial


ModLow

semi

tion

ted

Transport

Dominated


stop

go


















5 Conclusion


Acknowledgements







PdP0eminuskd






ξ frac14 L jTp

T jβminus1




1thorn em1thornm2

Lwbkf















7



















OR


OR







References


















































































































































































1 Motivation






31 Methods

32 Results

4 Discussion








47 Moving forward

5 Conclusion

Acknowledgements












References

Haga et al

(2002)

Millington and

Sear (2007)

MacVicar

et al (2009)

Ulloa et al (2015)


MacVicar

et al (2012)

Van der Nat

et al (2003)

Small

Medium

Large Large

amp Great

Bertoldi et

al(2013)

Van der Nat

et al (2003)

Bertoldi

et al(2013)

Ulloa et al

(2015)

MacVicar et al

(2009)

high mobility from

bull extreme events



changes



stable channels

mobility envelope

remobilized

redeposited

repositioned

M

ob

ile

M

ob

ile

Stream width (m)

Stream width (m)

Stream width (m)

Tra

ve

l d

ista

nce (m

)

L

L

L









Lgt5 1leL le 5

Small Medium

Llt1

Large

Large

Braided

Med=9

Med=14

Re-m

obiliz

ation

ra

te (

)

Med=32

Med=80

Min=25

Max=95

Max=83Max=80

Max=33

n=45 n=5n=23 n=136

020

40

60

80

100






4 Discussion


















Time

ET

Amount

wood in

storage

(S)

+∆S

RT RT

+∆S








ET

ET= episodic

transfer

ETRT

-∆S +∆S ∆S=0

RT=regular

transfer

equilibrium



moving in

contact with

channel deflecting

against channel

boundaries

0

of time

in contact

with

channel

100

unimpeded

floating

0

of

time fully

floating

100


i)

1

immobile

partially

mobile

fully mobile

QtQ

bf

0

proportion

of stored

wood

1

341















Small Rivers

Medium

Rivers

Large Rivers

Great Rivers

Residence Time (Tr)

Travel

Distance

(Xw)

lateral

export

basin

export





















Rising

Limb

gt~34 bf

le bf

lt1

lt1

Wood Flow and C

Governing Transp

hydrograph position

Falling

Limb

stage height

gt34 bf

gt1



for Transport

asymp1

gt1

wood length

divide

flow width fl

Deposi

Domina


stop go





47 Moving forward



lowlimited noneFew

High


ort Variability

transport

congestion

high

degree of

flow

obstruction

Complete

partial

living

vegetation

in flow

lots

some

flow depths

lt

otation depth

most

partial


ModLow

semi

tion

ted

Transport

Dominated


stop

go


















5 Conclusion


Acknowledgements







PdP0eminuskd






ξ frac14 L jTp

T jβminus1




1thorn em1thornm2

Lwbkf















7



















OR


OR







References


















































































































































































1 Motivation






31 Methods

32 Results

4 Discussion








47 Moving forward

5 Conclusion

Acknowledgements












References

Lgt5 1leL le 5

Small Medium

Llt1

Large

Large

Braided

Med=9

Med=14

Re-m

obiliz

ation

ra

te (

)

Med=32

Med=80

Min=25

Max=95

Max=83Max=80

Max=33

n=45 n=5n=23 n=136

020

40

60

80

100






4 Discussion


















Time

ET

Amount

wood in

storage

(S)

+∆S

RT RT

+∆S








ET

ET= episodic

transfer

ETRT

-∆S +∆S ∆S=0

RT=regular

transfer

equilibrium



moving in

contact with

channel deflecting

against channel

boundaries

0

of time

in contact

with

channel

100

unimpeded

floating

0

of

time fully

floating

100


i)

1

immobile

partially

mobile

fully mobile

QtQ

bf

0

proportion

of stored

wood

1

341















Small Rivers

Medium

Rivers

Large Rivers

Great Rivers

Residence Time (Tr)

Travel

Distance

(Xw)

lateral

export

basin

export





















Rising

Limb

gt~34 bf

le bf

lt1

lt1

Wood Flow and C

Governing Transp

hydrograph position

Falling

Limb

stage height

gt34 bf

gt1



for Transport

asymp1

gt1

wood length

divide

flow width fl

Deposi

Domina


stop go





47 Moving forward



lowlimited noneFew

High


ort Variability

transport

congestion

high

degree of

flow

obstruction

Complete

partial

living

vegetation

in flow

lots

some

flow depths

lt

otation depth

most

partial


ModLow

semi

tion

ted

Transport

Dominated


stop

go


















5 Conclusion


Acknowledgements







PdP0eminuskd






ξ frac14 L jTp

T jβminus1




1thorn em1thornm2

Lwbkf















7



















OR


OR







References


















































































































































































1 Motivation






31 Methods

32 Results

4 Discussion








47 Moving forward

5 Conclusion

Acknowledgements












References







Time

ET

Amount

wood in

storage

(S)

+∆S

RT RT

+∆S








ET

ET= episodic

transfer

ETRT

-∆S +∆S ∆S=0

RT=regular

transfer

equilibrium



moving in

contact with

channel deflecting

against channel

boundaries

0

of time

in contact

with

channel

100

unimpeded

floating

0

of

time fully

floating

100


i)

1

immobile

partially

mobile

fully mobile

QtQ

bf

0

proportion

of stored

wood

1

341















Small Rivers

Medium

Rivers

Large Rivers

Great Rivers

Residence Time (Tr)

Travel

Distance

(Xw)

lateral

export

basin

export





















Rising

Limb

gt~34 bf

le bf

lt1

lt1

Wood Flow and C

Governing Transp

hydrograph position

Falling

Limb

stage height

gt34 bf

gt1



for Transport

asymp1

gt1

wood length

divide

flow width fl

Deposi

Domina


stop go





47 Moving forward



lowlimited noneFew

High


ort Variability

transport

congestion

high

degree of

flow

obstruction

Complete

partial

living

vegetation

in flow

lots

some

flow depths

lt

otation depth

most

partial


ModLow

semi

tion

ted

Transport

Dominated


stop

go


















5 Conclusion


Acknowledgements







PdP0eminuskd






ξ frac14 L jTp

T jβminus1




1thorn em1thornm2

Lwbkf















7



















OR


OR







References


















































































































































































1 Motivation






31 Methods

32 Results

4 Discussion








47 Moving forward

5 Conclusion

Acknowledgements












References

moving in

contact with

channel deflecting

against channel

boundaries

0

of time

in contact

with

channel

100

unimpeded

floating

0

of

time fully

floating

100


i)

1

immobile

partially

mobile

fully mobile

QtQ

bf

0

proportion

of stored

wood

1

341















Small Rivers

Medium

Rivers

Large Rivers

Great Rivers

Residence Time (Tr)

Travel

Distance

(Xw)

lateral

export

basin

export





















Rising

Limb

gt~34 bf

le bf

lt1

lt1

Wood Flow and C

Governing Transp

hydrograph position

Falling

Limb

stage height

gt34 bf

gt1



for Transport

asymp1

gt1

wood length

divide

flow width fl

Deposi

Domina


stop go





47 Moving forward



lowlimited noneFew

High


ort Variability

transport

congestion

high

degree of

flow

obstruction

Complete

partial

living

vegetation

in flow

lots

some

flow depths

lt

otation depth

most

partial


ModLow

semi

tion

ted

Transport

Dominated


stop

go


















5 Conclusion


Acknowledgements







PdP0eminuskd






ξ frac14 L jTp

T jβminus1




1thorn em1thornm2

Lwbkf















7



















OR


OR







References


















































































































































































1 Motivation






31 Methods

32 Results

4 Discussion








47 Moving forward

5 Conclusion

Acknowledgements












References

Small Rivers

Medium

Rivers

Large Rivers

Great Rivers

Residence Time (Tr)

Travel

Distance

(Xw)

lateral

export

basin

export





















Rising

Limb

gt~34 bf

le bf

lt1

lt1

Wood Flow and C

Governing Transp

hydrograph position

Falling

Limb

stage height

gt34 bf

gt1



for Transport

asymp1

gt1

wood length

divide

flow width fl

Deposi

Domina


stop go





47 Moving forward



lowlimited noneFew

High


ort Variability

transport

congestion

high

degree of

flow

obstruction

Complete

partial

living

vegetation

in flow

lots

some

flow depths

lt

otation depth

most

partial


ModLow

semi

tion

ted

Transport

Dominated


stop

go


















5 Conclusion


Acknowledgements







PdP0eminuskd






ξ frac14 L jTp

T jβminus1




1thorn em1thornm2

Lwbkf















7



















OR


OR







References


















































































































































































1 Motivation






31 Methods

32 Results

4 Discussion








47 Moving forward

5 Conclusion

Acknowledgements












References








Rising

Limb

gt~34 bf

le bf

lt1

lt1

Wood Flow and C

Governing Transp

hydrograph position

Falling

Limb

stage height

gt34 bf

gt1



for Transport

asymp1

gt1

wood length

divide

flow width fl

Deposi

Domina


stop go





47 Moving forward



lowlimited noneFew

High


ort Variability

transport

congestion

high

degree of

flow

obstruction

Complete

partial

living

vegetation

in flow

lots

some

flow depths

lt

otation depth

most

partial


ModLow

semi

tion

ted

Transport

Dominated


stop

go


















5 Conclusion


Acknowledgements







PdP0eminuskd






ξ frac14 L jTp

T jβminus1




1thorn em1thornm2

Lwbkf















7



















OR


OR







References


















































































































































































1 Motivation






31 Methods

32 Results

4 Discussion








47 Moving forward

5 Conclusion

Acknowledgements












References
















5 Conclusion


Acknowledgements







PdP0eminuskd






ξ frac14 L jTp

T jβminus1




1thorn em1thornm2

Lwbkf















7



















OR


OR







References


















































































































































































1 Motivation






31 Methods

32 Results

4 Discussion








47 Moving forward

5 Conclusion

Acknowledgements












References





PdP0eminuskd






ξ frac14 L jTp

T jβminus1




1thorn em1thornm2

Lwbkf















7



















OR


OR







References


















































































































































































1 Motivation






31 Methods

32 Results

4 Discussion








47 Moving forward

5 Conclusion

Acknowledgements












References

7



















OR


OR







References


















































































































































































1 Motivation






31 Methods

32 Results

4 Discussion








47 Moving forward

5 Conclusion

Acknowledgements












References
























































































































































1 Motivation






31 Methods

32 Results

4 Discussion








47 Moving forward

5 Conclusion

Acknowledgements












References

Rules of the road: A qualitative and quantitative ... · Rules of the road: A qualitative and...

Documents

Transcript of Rules of the road: A qualitative and quantitative ... · Rules of the road: A qualitative and...