Soil carbon in dynamic Soil carbon in dynamic land use optimization land use optimization

models models

Uwe A. SchneiderUwe A. Schneider

Research Unit Sustainability and Global Research Unit Sustainability and Global ChangeChange

Hamburg UniversityHamburg University

TopicsTopics

I Land use modelsI Land use models

II Linking biophysical and economic modelsII Linking biophysical and economic models

III Soil carbon in economic modelsIII Soil carbon in economic models

IV FASOMIV FASOM

I Land use modelsI Land use models

Research QuestionsResearch Questions

Economic Economic SustainabilitySustainability– FoodFood– EnergyEnergy– CommoditiesCommodities

Environmental Environmental SustainabilitySustainability– AirAir– WaterWater– SoilSoil– Climate Climate – EcosystemsEcosystems

Policy analysisPolicy analysis– Economic Economic

potentialpotential– ImpactsImpacts

Focus of land use modelsFocus of land use models

Technologies (Species, Tillage, Technologies (Species, Tillage, Planting, Fertilizing, Protection, Planting, Fertilizing, Protection, Harvesting)Harvesting)

Economics (Market Prices, Trade, Economics (Market Prices, Trade, Income)Income)

Environment (Resources, Emissions, Environment (Resources, Emissions, Sinks, Wildlife, Climate)Sinks, Wildlife, Climate)

Land use estimationLand use estimation

StorylinesStorylines

StatisticsStatistics

OptimizationOptimization

OptimizationOptimization

Constrained welfare/profit Constrained welfare/profit maximizationmaximization

Normative economics (positive Normative economics (positive economics via calibration)economics via calibration)

Application to structural change Application to structural change beyond historical observationsbeyond historical observations

Land use optimizationLand use optimization

Find welfare maximizing land Find welfare maximizing land managementmanagement

s.t.s.t. resourcesresources

technologiestechnologies

marketsmarkets

policiespolicies

Linear ProgramLinear Program

jallfor0X

iallforbXa..

Xc

j

ij

jij

jjj

ts

Max

II Linking biophysical II Linking biophysical and economic modelsand economic models

Why linkage?Why linkage?

Standalone biophysical models Standalone biophysical models simulate environmental impacts of land simulate environmental impacts of land management but don’t explain why a management but don’t explain why a certain land management is chosencertain land management is chosen

Standalone economic models explain Standalone economic models explain land management adoption but cannot land management adoption but cannot internalize environmental impactsinternalize environmental impacts

ChallengesChallenges

Spatial resolution (field vs. globe)Spatial resolution (field vs. globe)

Temporal resolution (days vs. Temporal resolution (days vs. decades)decades)

Technological resolutionTechnological resolution

Environmental resolutionEnvironmental resolution

Types of Linkage (Problems)Types of Linkage (Problems)

A.A. Economic model Economic model Biophysical Biophysical model (no adaptation, no feedback)model (no adaptation, no feedback)

B.B. Biophysical model Biophysical model Economic Economic model (curse of dimensionality)model (curse of dimensionality)

C.C. Iterative link (costly, ITR)Iterative link (costly, ITR)

D.D. Fully integrated model Fully integrated model (computational limits)(computational limits)

Economic model Economic model Biophysical Biophysical modelmodel

Determine land use trajectory with Determine land use trajectory with economic model for different economic model for different scenariosscenarios

Simulate environmental impacts for Simulate environmental impacts for each scenarioeach scenario

Adaptation of land management to Adaptation of land management to environmental policies ignoredenvironmental policies ignored

Feedback of changing environment Feedback of changing environment on adaptation ignored as wellon adaptation ignored as well

Biophysical model Biophysical model Economic Economic modelmodel

Simulate environmental impacts for Simulate environmental impacts for all possible land use choicesall possible land use choices

Enter environmental impacts in Enter environmental impacts in economic modeleconomic model

Set values for environmental impacts Set values for environmental impacts (environmental policies)(environmental policies)

Find welfare maximizing levelsFind welfare maximizing levels

Curse of Dimensionality?Curse of Dimensionality?

20 Crops20 Crops 5 Management options per crop5 Management options per crop 100 Regions 100 Regions 5 Soil Types per region5 Soil Types per region

50,000 Land use alternatives

Curse of Dimensionality?Curse of Dimensionality?

20 Crops20 Crops 5 Management options per crop5 Management options per crop 100 Regions 100 Regions 5 Soil Types per region5 Soil Types per region 20 Periods20 Periods

5*E42 Trajectories (independent sites)

1*E94 Trajectories (dependent sites)

III Soil carbon in III Soil carbon in economic modelseconomic models

Soil carbon and economicsSoil carbon and economics

Productivity impact of soil carbon Productivity impact of soil carbon (yields, suitability)(yields, suitability)

Economic potential of carbon sinks Economic potential of carbon sinks for climate change mitigationfor climate change mitigation

Carbon sinks vs. bioenergy vs. Carbon sinks vs. bioenergy vs. biodiversity vs. traditional marketsbiodiversity vs. traditional markets

Soil Carbon DeterminantsSoil Carbon Determinants

Crop ChoiceCrop Choice TillageTillage IrrigationIrrigation FertilizationFertilization Residue MgtResidue Mgt Soil CarbonSoil Carbon

Soil Soil

Carbon Carbon

ChangeChange

Soi

l Org

anic

Car

bon

(tC

/ha/

20cm

)

5

10

15

20

25

30

35

40

45

0 10 20 30 40 50Time (years)

Wheat-Lucerne 3/3

Wheat-Lucerne 6/3

No-till wheat-fallow

Tilled wheat-fallow

Simple Multi-Period Land Use Simple Multi-Period Land Use ModelModel

Market Environmentt t ,r,i,u t ,r,i,u t ,r,i,u

t ,

t ,r,i,u t ,r,

r,i,u

iu

s.

Max v v X

t. X l

Indexes: t = time, r = region, i = soil type, u = managementData: = interest rate, v = net benefit, l=land endowment Variables: X = land use

Explicit Land Use Explicit Land Use TrajectoriesTrajectories

d d d

d

d d

d

Market Environment

r,i,u r,i,u r,i,

r,i,u r,i,

ur,i,

u

u

u

Max v v X

s.t. X L

Indexes: r = region, i = soil type, ud = management path

Implicit Land Use Implicit Land Use Trajectories Trajectories

Assume that management history is Assume that management history is manifest in current soil carbon levelsmanifest in current soil carbon levels

Divide soil carbon rangeDivide soil carbon range

Implement Markov ChainImplement Markov Chain

Markov ProcessMarkov Process

Market Environmentt t ,r,i,u,o t ,r,i,u,o t ,r,i,u

t ,r,i,u,

t ,r,i,u,o r,i,u,o,o t 1,r,i,u,ou u,

,ot ,r,i,u,o

o t,r,io

o

u,

s.

Max v v X

X

t X l

X

.

Indexes: t = time, r = region, i = soil type, u = management

o,ố = soil carbon state = transition probability from old state ố to new state o

Soil Carbon Transition Soil Carbon Transition ProbabilitiesProbabilities

SOC1 SOC2 SOC3 SOC4 SOC5 SOC6 SOC7 SOC8

SOC1 0.81 0.19

SOC2 1

SOC3 0.09 0.91

SOC4 0.31 0.69

SOC5 0.5 0.5

SOC6 0.74 0.26

SOC7 1

SOC8 0.04 0.96

No-till wheat-Fallow

5

10

15

20

25

30

35

40

45

0 10 20 30 40 50Time (years)

Wheat-Lucerne 3/3

Wheat-Lucerne 6/3

No-till wheat-fallow

Tilled wheat-fallowSoi

l Org

anic

Car

bon

(tC

/ha/

20cm

)

Curse of Dimensionality?Curse of Dimensionality?

20 Crops20 Crops 5 Management options per crop5 Management options per crop 100 Regions 100 Regions 5 Soil Types per region5 Soil Types per region 20 Periods20 Periods

5E42 Trajectories (independent sites)

1E94 Trajectories (dependent sites)

Curse of Dimensionality?Curse of Dimensionality?

20 Crops20 Crops 5 Management options per crop5 Management options per crop 100 Regions 100 Regions 5 Soil Types per region5 Soil Types per region 20 Periods20 Periods

1E6 Variables (No Soil Carbon)

1E7 Variables (Markov process with 10 states)

5E42..1E94 Variables (Explicit Path)

Extensions?Extensions?

Markov chains are applicable to Markov chains are applicable to relatively independent environmental relatively independent environmental qualities (humus, salt, qualities (humus, salt, contamination)contamination)

Method not suitable for complex Method not suitable for complex environmental properties (climate)environmental properties (climate)

IV Forest and IV Forest and Agricultural Sector Agricultural Sector Optimization Model Optimization Model

FASOMFASOM

Overall ObjectiveOverall Objective

Portray agricultural and forest Portray agricultural and forest commodity markets and internalize all commodity markets and internalize all

land use externalitiesland use externalities

Analyze PoliciesIntegrate Synergies, Trade-offs

Markets

Soil

ClimateWildlife

Land use decisions Water

Farmers

Model StructureModel Structure

Resources Land Use

Technologies

Processing Technologies

Products Markets

Inputs

Limits

Supply Functions

Limits

Demand Functions,Trade

Limits

Environmental Impacts

Economic Surplus Economic Surplus MaximizationMaximization

Implicit Supply and Demand

Forest InventoryLand Supply

Water Supply

Labor Supply

Animal Supply

National Inputs Import Supply

Processing Demand

Feed Demand

Domestic Demand

Export Demand

CS

PS

Spatial ResolutionSpatial Resolution

Soil textureSoil texture Stone contentStone content Altitude levelsAltitude levels SlopesSlopes Soil stateSoil state

Political regionsPolitical regions Ownership Ownership

(forests)(forests) Farm typesFarm types Farm sizeFarm size

Many crop and tree Many crop and tree speciesspecies

Tillage, planting Tillage, planting irrigation, fertilization irrigation, fertilization harvest regimeharvest regime

Altitude:1. < 300 m2. 300-600 m3. 600-1100 m4. >1100 m

Texture:1. Coarse2. Medium3. Medium-fine4. Fine 5. Very fine

Soil Depth:1. shallow2. medium3. deep

Stoniness:1. Low content2. Medium content3. High content

Slope Class:1. 0-3%2. 3-6%3. 6-10%4. 10-15%5. …

Homogeneous Response Units

DE13

DE12

DE11

DE14

Climate Change MitigationClimate Change Mitigation

0

100

200

300

400

500

0 20 40 60 80 100 120 140 160 180 200

Car

bon

pri

ce (

$/tc

e)

Emission reduction (mmtce)

CH4N2O

Ag-Soil sequestration

Afforestation

Biofuel offsets

Soil Carbon PotentialsSoil Carbon Potentials

0

100

200

300

400

500

0 20 40 60 80 100 120 140 160

Car

bon

pri

ce (

$/tc

e)

Soil carbon sequestration (mmtce)

Technical Potential

Economic Potential

Competitive EconomicPotential

Biofuel PotentialsBiofuel Potentials

0

100

200

300

400

500

0 50 100 150 200 250 300 350

Car

bon

pri

ce (

$/tc

e)

Emission reduction (mmtce)

Technical Potential

Economic Potential

Competitive EconomicPotential

Afforestation PotentialsAfforestation Potentials

0

100

200

300

400

500

0 50 100 150 200 250 300

Car

bon

pri

ce (

Eu

ro/t

ce)

Emission reduction (mmtce)

Technical Potential

Economic PotentialCompetitive

Economic Potential