Outline

27
Least Cost Control of Agricultural Nutrient Agricultural Nutrient Contributions to Local Contributions to Local Waters and the Gulf of Waters and the Gulf of Mexico Dead Zone Mexico Dead Zone Sergey Rabotyagov, Todd Campbell, Manoj Jha, Philip Sergey Rabotyagov, Todd Campbell, Manoj Jha, Philip W. Gassman, Jeffrey Arnold, Lyubov Kurkalova, W. Gassman, Jeffrey Arnold, Lyubov Kurkalova, Silvia Secchi, Hongli Feng, and Catherine L. Kling. Silvia Secchi, Hongli Feng, and Catherine L. Kling. Center for Agricultural and Rural Center for Agricultural and Rural Development, Development, Department of Economics Department of Economics

description

- PowerPoint PPT Presentation

Transcript of Outline

Page 1: Outline

Least Cost Control of Agricultural Least Cost Control of Agricultural

Nutrient Contributions to Local Nutrient Contributions to Local

Waters and the Gulf of Mexico Waters and the Gulf of Mexico

Dead ZoneDead Zone

Sergey Rabotyagov, Todd Campbell, Manoj Jha, Philip W. Sergey Rabotyagov, Todd Campbell, Manoj Jha, Philip W.

Gassman, Jeffrey Arnold, Lyubov Kurkalova, Silvia Secchi, Gassman, Jeffrey Arnold, Lyubov Kurkalova, Silvia Secchi,

Hongli Feng, and Catherine L. Kling.Hongli Feng, and Catherine L. Kling.

Center for Agricultural and Rural Center for Agricultural and Rural

Development, Development,

Department of EconomicsDepartment of Economics

Iowa State UniversityIowa State University

February, 2008February, 2008

Page 2: Outline

3

OutlineOutline

Intro to water quality issues in the U.S. Intro to water quality issues in the U.S. midwestmidwest Local water qualityLocal water quality Gulf of Mexico hypoxic zoneGulf of Mexico hypoxic zone

Agricultural conservation practices and Agricultural conservation practices and nonpoint source pollutionnonpoint source pollution

Economic problem of finding least cost Economic problem of finding least cost solution to this nonpoint problemsolution to this nonpoint problem

Empirical results of applying evolutionary Empirical results of applying evolutionary algorithm with hydrologic simulation modelalgorithm with hydrologic simulation model

Page 3: Outline

4

189,000 square miles in seven states,189,000 square miles in seven states, dominated by agriculture: 67% of total dominated by agriculture: 67% of total

area,area, > 1200 stream segments and lakes on > 1200 stream segments and lakes on

EPAs impaired waters list, EPAs impaired waters list, highest concentrations of phosphorous highest concentrations of phosphorous

found in the worldfound in the world

The UMRB: local water quality

Page 4: Outline

5

Local Water QualityLocal Water Quality

• Nutrients (esp. phosphorous) and sediment primary source

• Agriculture accounts for over 50% of impairments (EPA)

• Multiple conservation practices can ameliorate (Land retirement, conservation tillage, grassed waterways, Land retirement, conservation tillage, grassed waterways, contours, terraces)contours, terraces)

Page 5: Outline

6

Hypoxia = Dead Hypoxia = Dead ZoneZone

• 165 Hypoxic zones worldwide

• Naturally occurring, but far larger due to anthropogenic sources of nutrients

• Depleted oxygen creates zones incapable of supporting most life

• Potential evils: commercial fishing, recreation, ecosystem effects, habitat, etc.

• Effects quite poorly documented, information from other hypoxic zones relevant? Ecological “tipping point?”

Page 6: Outline

7

Summertime satellite observations of ocean color from MODIS/Aqua show highly turbid waters which may include large blooms of phytoplankton extending from the mouth of the Mississippi River all the way to the Texas coast. When these blooms die and sink to the bottom, bacterial decomposition strips oxygen from the surrounding water, creating an environment very difficult for marine life to survive in. Reds and oranges represent high concentrations of phytoplankton and river sediment. Image taken by NASA and provided courtesy of the NASA Mississippi Dead Zone web site.

Hypoxic zone in the Gulf of Mexico.

Page 7: Outline

10

Gulf of Mexico HypoxiaGulf of Mexico Hypoxia• Now, second largest worldwide

•4000 – 22,000 sq km

Source: Dr. Nancy RabalaisUniversities Marine Consortium

Page 8: Outline

11

Gulf Hypoxia: Emission Gulf Hypoxia: Emission SourcesSources

• Causes: nutrients from Mississippi river, nitrates and phosphorous,

• Limiting nutrient may now be P (scientific debate continues)

• Major Contributors:

• UMRB (1+2) = 43%N, 41%P

• Ohio-Tennessee (6+7) = 41%N, 59%P

http://toxics.usgs.gov/pubs/of-2007-1080/methods_large_subbasins.html

Page 9: Outline

12

Examples of conservation Examples of conservation practicespractices

Terraces: an earth embankment, or a combination ridge and Terraces: an earth embankment, or a combination ridge and channel, constructed across the field slope (USDA-NRCS)channel, constructed across the field slope (USDA-NRCS)

Grassed waterways: natural or constructed channel with suitable Grassed waterways: natural or constructed channel with suitable vegetationvegetation

Contour farming: tillage, planting, and other farming operations Contour farming: tillage, planting, and other farming operations performed on or near the contour of the field slopeperformed on or near the contour of the field slope

No-till: managing the amount, orientation and distribution of crop No-till: managing the amount, orientation and distribution of crop and other plant residue on the soil surface year round while and other plant residue on the soil surface year round while limiting soil-disturbing activities to only those necessary to place limiting soil-disturbing activities to only those necessary to place nutrients, condition residue and plant crops nutrients, condition residue and plant crops

Land retirement (CRP): remove land from working production, Land retirement (CRP): remove land from working production, plant with perennial grasses or other appropriate vegetationplant with perennial grasses or other appropriate vegetation

Nutrient management: reduced fertilization, NNutrient management: reduced fertilization, N

Page 10: Outline

13

Conservation practicesConservation practices

Photos courtesy of USDA NRCSPhotos courtesy of USDA NRCS

Page 11: Outline

15

Integrated Economic, Land Integrated Economic, Land use, and Water Quality use, and Water Quality

Model for the UMRBModel for the UMRB Couple large-scale, spatially-detailed watershed Couple large-scale, spatially-detailed watershed

model with economic model to study costs and model with economic model to study costs and water quality changes of conservation policywater quality changes of conservation policy

Focus on agricultural land use decisions – Focus on agricultural land use decisions – croplandcropland

Use NRI as basis for both economics and Use NRI as basis for both economics and watershed model (watershed model (110,000 total “points” and expansion factors, 110,000 total “points” and expansion factors, 37,500 cropland observations)37,500 cropland observations)

Purpose of modeling system is to provide policy Purpose of modeling system is to provide policy level informationlevel information

Consider both upstream water quality (within the Consider both upstream water quality (within the UMRB), and downstream effects (Gulf of Mexico) UMRB), and downstream effects (Gulf of Mexico)

Page 12: Outline

17

Key Watershed FeaturesKey Watershed Features

7070

7120

7090

7030

71007080

7140

7040

7130

7010

70207050

7110

7060

Features of the 4 Digit HUCs

7010 8954 1.2 18 52

7020 7797 0.92 69 91

7030 4113 0.46 10 35

7040 6495 0.65 33 78

7050 3847 0.55 11 40

7060 5930 0.55 42 122

7070 5141 0.66 14 73

7080 14965 1.46 67 128

7090 7167 0.66 56 121

7100 8375 0.9 64 116

7110 5883 0.59 44 69

7120 7661 0.63 55 116

7130 9745 1.13 72 129

7140 7776 0.79 44 79

Average CRP rental rates

4 Digit HUC

Total NRI points

Total area millions of

acres

% total area cropped

Page 13: Outline

19

Least Cost ProblemLeast Cost Problem One field’s pollution damage affected by choices One field’s pollution damage affected by choices

on other fieldson other fields no exogenous “delivery coefficients”no exogenous “delivery coefficients” non-mutually exclusive CP’s can be implemented on any non-mutually exclusive CP’s can be implemented on any

field, different effectiveness and costsfield, different effectiveness and costs precludes simple spatial optimization schemesprecludes simple spatial optimization schemes

Brute forceBrute force using hydrologic model, analyze all the feasible using hydrologic model, analyze all the feasible

scenarios, picking cost-efficient solutions scenarios, picking cost-efficient solutions

But, if there are N conservation practices possible for But, if there are N conservation practices possible for adoption on each field and there are F fields, this adoption on each field and there are F fields, this implies a total of possible Nimplies a total of possible NF F configurations to compareconfigurations to compare

30 fields, 2 options 30 fields, 2 options over 1 billion possible scenarios over 1 billion possible scenarios

Page 14: Outline

24

Multiobjective optimization Multiobjective optimization evolutionary algorithms evolutionary algorithms

provide a way to search for Pareto-optimal setsprovide a way to search for Pareto-optimal sets

SPEA2 (Strength Pareto Evolutionary Algorithm 2) SPEA2 (Strength Pareto Evolutionary Algorithm 2) (Ziztler and Thiele, 2001) is used(Ziztler and Thiele, 2001) is used

Of the three objectives (N, P, Cost), only cost can Of the three objectives (N, P, Cost), only cost can be easily computed for a particular scenario, be easily computed for a particular scenario, nutrient loadings need to be simulated nutrient loadings need to be simulated

Combine:Combine: An evolutionary algorithm, SPEA2An evolutionary algorithm, SPEA2 Hydrologic model, Soil and Water Assessment Tool Hydrologic model, Soil and Water Assessment Tool

(SWAT)(SWAT) Sometimes referred to as Sometimes referred to as simulation-optimization simulation-optimization

frameworkframework

Page 15: Outline

25

One possible watershed One possible watershed configurationconfiguration

a

d b

a b

c

a

d

a

b

aa

a

13 Fields 13 Fields 4 conservation practices4 conservation practices131344=28561 possible configurations=28561 possible configurations

Genetic Algorithm lingoField = genePractice options =allele setwatershed configuration = individual (described by set of genes)Population = set of configurations

Page 16: Outline

26

Algorithm flow diagramAlgorithm flow diagram

Individual = watershed configuration = specific assignment of practices to fields

Population = set of watershed configurations

Page 17: Outline

29

Fitness assignment Fitness assignment exampleexample

Strength SStrength S(i(i))== # of individuals # of individuals ii dominates dominates Raw fitness RRaw fitness R(i(i))== sum of strengths of individuals that sum of strengths of individuals that

dominate dominate ii

Page 18: Outline

32

Soil and Water Assessment Soil and Water Assessment Tool (SWAT)Tool (SWAT)

A hydrologic and water quality model A hydrologic and water quality model developed by the U.S. Department of developed by the U.S. Department of Agriculture’s Agricultural Research Service Agriculture’s Agricultural Research Service (USDA-ARS)(USDA-ARS)

A long-term continuous watershed-scale A long-term continuous watershed-scale simulation model that operates on a daily time simulation model that operates on a daily time step and is designed to assess the impact of step and is designed to assess the impact of different management practices on water, different management practices on water, sediment, and agricultural chemical yieldssediment, and agricultural chemical yields

Gassman et al. (2007) identify over 250 Gassman et al. (2007) identify over 250 publications using SWATpublications using SWAT

Page 19: Outline

33

The conservation The conservation practices setpractices set

For each hydrologic unit in the watershed, For each hydrologic unit in the watershed, consider 1 of 33 mutually exclusive optionsconsider 1 of 33 mutually exclusive options

One is land retirementOne is land retirement Obtain the rest of options by interacting 4 tillage Obtain the rest of options by interacting 4 tillage

types (CT, RT, MT, NT) with types (CT, RT, MT, NT) with Practices: Practices:

TerracesTerraces ContouringContouring Grassed WaterwaysGrassed Waterways

20% N fertilizer reduction20% N fertilizer reduction Baseline conservation practices impose a set of Baseline conservation practices impose a set of

constraintsconstraints In this application, algorithm only allowed to add In this application, algorithm only allowed to add

practicespractices

Page 20: Outline

34

Costs of practicesCosts of practices

Sources of information on the cost of CP’s:Sources of information on the cost of CP’s: Terraces, Grassed Waterways, Contouring, No-Till:Terraces, Grassed Waterways, Contouring, No-Till:

EQIP: Federal Conservation ProgramEQIP: Federal Conservation Program IFIP: State Conservation ProgramIFIP: State Conservation Program CRP: Federal Conservation ProgramCRP: Federal Conservation Program

Land retirement: 2007 Iowa Cash Rental Rates Land retirement: 2007 Iowa Cash Rental Rates SurveySurvey

Costs of CP’s vary by subbasin in each Costs of CP’s vary by subbasin in each watershedwatershed

Develop a cost estimate for fertilizer reductionDevelop a cost estimate for fertilizer reduction Approximate cost=Yield reduction*Corn priceApproximate cost=Yield reduction*Corn price

Page 21: Outline

36

Pareto frontier: UMRBPareto frontier: UMRB

Page 22: Outline

38

Tradeoffs of NPS control costs Tradeoffs of NPS control costs and and

water quality benefitswater quality benefits Nitrate loadings at the outlet vs. costsNitrate loadings at the outlet vs. costs

Thus, if a policy seeks to reduce N by more than 20%, Thus, if a policy seeks to reduce N by more than 20%, more than 30% reductions in P follow more than 30% reductions in P follow

Empirical cost curve for nitrate reductions

0

200

400

600

800

1000

1200

1400

1600

1800

2000

0 20 40 60 80 100 120

Nitrates, % of baseline

Co

st,

% o

f b

asel

ine

Unconstrained cost curve

Binding constraint on 30% Preduction

Page 23: Outline

39

Tradeoffs of NPS control costs Tradeoffs of NPS control costs and and

water quality benefitswater quality benefits Phosphorus loadings at the outlet vs. costsPhosphorus loadings at the outlet vs. costs

In this case, nitrate reduction constraint is binding up to a 50% In this case, nitrate reduction constraint is binding up to a 50% reduction in Preduction in P

This suggests an asymmetry in the control of the two nutrients in This suggests an asymmetry in the control of the two nutrients in the UMRBthe UMRB

Empirical cost curve for phosphorus reductions

0

200

400

600

800

1000

1200

1400

1600

1800

0 20 40 60 80 100 120

P loadings, % of baseline

Co

st,

% o

f b

asel

ine

Unconstrained cost curve

Binding constraint on 30% Nreduction

Page 24: Outline

40

Tradeoffs between pollutants Tradeoffs between pollutants (Nitrates and P)(Nitrates and P)

The frontier also highlights what nutrient reductions are The frontier also highlights what nutrient reductions are feasible for a given budgetfeasible for a given budget

As cost increases (to 10X the baseline), the scope of the As cost increases (to 10X the baseline), the scope of the tradeoff fallstradeoff falls reflects complementarities embedded in land retirementreflects complementarities embedded in land retirement

12

14

16

18

20

22

24

26

28

220 270 320 370 420

Mill

ion

s

Millions

Nitrate loadings, kg

P lo

ad

ing

s, k

g 1000 % baseline cost

500 % baseline cost

400 % baseline cost

300 % baseline cost

200 % baseline cost

Page 25: Outline

44

A closer look at the frontierA closer look at the frontier

The frontier is also useful in answering how one can The frontier is also useful in answering how one can achieve specific pollution reduction targetsachieve specific pollution reduction targets

For example, For example, Requiring a 30% reduction in outlet NORequiring a 30% reduction in outlet NO3 3 automatically leads automatically leads

to a 35% reduction in outlet Pto a 35% reduction in outlet P But: requiring a 30% reduction in outlet P leads only to a But: requiring a 30% reduction in outlet P leads only to a

9% reduction in outlet NO9% reduction in outlet NO33

The annual additional cost is estimated to be:The annual additional cost is estimated to be: $ 1.4 billion for reducing NO$ 1.4 billion for reducing NO3 3 by 30% (more than by 30% (more than

quadrupling baseline cost)quadrupling baseline cost) $ 370 million for reducing P by 30% (less than doubling the $ 370 million for reducing P by 30% (less than doubling the

baseline cost)baseline cost)

Page 26: Outline

46

What practices are What practices are selected?selected?

4715: 30% N, 36% P4715: 30% N, 36% P 3821: 9% N, 30% P3821: 9% N, 30% P

Grassed Waterways Grassed Waterways 87%87%

Reduced Fertilizer Reduced Fertilizer 89%89%

Terraces 2%Terraces 2%

Reduced Tillage 58%Reduced Tillage 58%

Land Retirement 9%Land Retirement 9%

Cost = $1.4 Cost = $1.4 billion/yearbillion/year

Grassed Waterways Grassed Waterways 91%91%

Reduced Fertilizer Reduced Fertilizer 7%7%

Terraces 1%Terraces 1%

Reduced Tillage 66%Reduced Tillage 66%

Land Retirement 0%Land Retirement 0%

Cost = $ 370 Cost = $ 370 million/yearmillion/year

Page 27: Outline

49

ConclusionsConclusions The approachThe approach

Tries to fully account for the complexity Tries to fully account for the complexity of NPS pollution processesof NPS pollution processes

Searches for efficient solutions to Searches for efficient solutions to quantify tradeoffsquantify tradeoffs Between cost and pollution reductionsBetween cost and pollution reductions Between different pollutantsBetween different pollutants

As always, multiple caveats are in As always, multiple caveats are in orderorder But the method is flexible and amenable But the method is flexible and amenable

to improvementto improvement