Spatial Heterogeneity and Scale: Implications for Carbon Sequestration Policies for Agriculture

1
Spatial Heterogeneity and Scale Implications
for Carbon Sequestration Policies for Agriculture
Susan M. Capalbo Dept. of Ag. Econ. and
Econ. Montana State University November 19, 2002
Prepared for the USDA Symposium on Natural
Resource Management to Offset Greenhouse Gas
Emissions
With the collaboration of
John Antle, Montana State University Sian
Mooney, University of Wyoming Keith Paustian,
Colorado State University
2
Acknowledgements
This research was funded by the USDA special
grants, the USDA National Research Initiative
Competitive Grants Program, the NSF Methods and
Models for Integrated Assessment Program, and the
EPA STAR Program.
3
Objectives

• In this paper, we develop methods to investigate
• the efficiency of alternative types of policies
  or
• contracts for C sequestration in cropland soils,
• taking into account
• 1. The spatial heterogeneity of agricultural
  production systems, and
• Sensitivity of the marginal cost of supplying
  carbon to
• a. carbon rates
• b. scale
• c. yield variations

4
Objectives (2)
We describe per-hectare and per-tonne contracts
for soil C, and use a model of farmers decisions
to participate in soil C contracts to derive the
on-farm opportunity costs for each type of
contract.
5
Objectives (3)
We present an integrated assessment modeling
framework, based on coupled site-specific
biophysical simulation models and site-specific
economic data and models, that can be used to
simulate farmers decisions to participate in
both per-hectare and per-tonne contracts.
6
Objectives (4)

• Using this coupled modeling framework in a
• case study of the dryland grain production
• system of the Northern Plains region of the
• United States, we
• Quantify the inefficiency of per-hectare policies
  vis-a-vis per-tonne policies
• Test the sensitivity of the model to show how the
  costs vary depending upon scale of analysis and
  uncertainty of input parameters.

7
Related Papers available at www.climate.montana.ed
u
Antle, J.M., and S.M. Capalbo, Econometric-Proces
s Models for Integrated Assessment of
Agricultural Production Systems. American
Journal of Agricultural Economics 83 (May 2001)
389-401. Antle, J.M., S.M. Capalbo, S. Mooney, E.
Elliott and K. Paustian, Economics of
Agricultural Soil Carbon Sequestration An
Integrated Assessment Approach. Journal of
Agricultural and Resource Economics 26 (December
2001) 344-367. Mooney, S., J.M. Antle, S.M.
Capalbo, and K.H. Paustian, Contracting for
Carbon Credits Design and Costs of Measurement
and Monitoring. Staff Paper 2002-01, Dept. of
Ag. Econ. Econ., Montana State University.
(forthcoming JEEM) Antle, J.M., S.M. Capalbo, S.
Mooney, E.T. Elliott, and K.H. Paustian.
Sensitivity of Carbon Sequestration Costs to
Soil Carbon Rates. Environmental Pollution 116
(March 2002) 413422.
8
How to Sequester Soil C?

• Command and control
• Legislate best management practices
• Incentive mechanisms let markets work

9
Designing Contracts for Soil C

• Per-hectare contract payment for use of BMP
• Payment independent of quantity of C
• Must monitor practices for compliance with
  contract
• Farmers enter contract if g gt ?ji ?js
• Per-tonne contract pays farmer P/tonne/yr for
  duration of contract
• Payment independent of practice
• Must quantify amount of C
• Establish baseline
• Measure accumulation of C
• Farmers enter contract if P gt (?ji - ?js)/?cjis ,
  I.e. if price per tonne is greater than
  opportunity cost per tonne

10

• Result from earlier papers
• For each quantity of C sequestered, the
  marginal opportunity cost of the per-hectare
  payment mechanism (MCH) is greater than or equal
  to the marginal opportunity cost of the per-tonne
  mechanism (MCT), i.e., MCH ? MCT, and MCT /MCH is
  decreasing with spatial heterogeneity.

11
Marginal Cost Functions for per-hectare and
per-tonne Payments
12
Integrated Assessment Paradigm

• Economic data ? economic production models
• Soils climate data ? crop ecosystem models
• Output of crop ecosystem models ?
• economic models and environmental process models
• Output of economic models ?
• environmental process models

13
(No Transcript)
14
Design of EconometricProcess Simulation Model

• Estimate econometric production models (system of
  supply and input demand fcns) for each activity.
• Simulate econometric models with site-specific
  data to obtain expected returns.
• Use structure of decision making process to make
  land use and management decisions.

15
Simulation of Land Use Using Econometric-Process
Model of Montana Dryland Grain Production

• 1995 MT Cropping Practices Survey
• Statistically representative sample of MLRAs in
  grain producing regions of MT
• Useable data from 425 commercial grain farms

16
Montana Dryland Grain Study Sub-MLRAs
17
Soil Carbon Simulations Performed with the
Century Model for each Sub-MLRA

• Model parameterized for each sub-MLRA using
  various sources of data for soils, climate, and
  cropping practices
• Model executed over 50 years for each cropping
  system for each sub-MLRA to achieve new
  equilibrium soil C levels

18

• Link economic simulation model to Century
  ecosystem model
• Assess the costs of inducing changes in levels of
  soil C (opportunity costs)
• Alternative policies
• per hectare payments
• per tonne payments

19
Land Allocation in Montana Dryland Grain
Production Systems
20
Soil C Levels Predicted by Century Model for
Cropping Systems in Montana
21
(No Transcript)
22
Marginal Cost of per-hectare and per-tonne
Payment Mechanisms
23
Marginal Cost of per-hectare and per-tonne
Payment Mechanisms
24
Marginal Cost of per-hectare and per-tonne
Payment Mechanisms
25

• Sensitivity of the Results to
• Soil carbon rates
• Scale for measuring soil carbon rates
• Yield uncertainites
• Output price uncertainites

26

• Changes in Soil Carbon Rates
• Keep spatial heterogeneity
• Adjust by 50 increase in soil C rates
• Adjust by 50 decrease in soil C rates
• Show results for per tonne and
• per hectare contracts

27
Sensitivity to Carbon Rates Sub-MLRA 52-high
per-tonne contract
per-hectare contract
28
Sensitivity to Carbon Rates Sub-MLRA 58a-low
per-tonne contract
per-hectare contract
29

• Changes in Soil C rates change the quantity of
  soil C sequestered at various prices (shifts the
  MC curve)
• Under per-hectare policy, as soil C rates
  increase, the impact on soil C sequestered
  increases in proportion to the square of the
  increase in soil C rates
• Under per-tonne policy, we have a linear mapping
  of changes in soil C rate and changes in MC curve

30

• Sensitivity to Scale
• Use average rates of soil C across
• all Sub MLRAs
• Use representative rates from
• Sub-MLRA 52-high
• Sub-MLRA 58a-low

31
Carbon Scale Comparison Sub-MLRA 52-high
per-tonne contract
per-hectare contract
32
Carbon Scale Comparison Sub-MLRA 58a-low
per-tonne contract
per-hectare contract
33

• Impacts are specific to Sub-MLRA
• Using mean rates of soil C underestimates the
  MC for Sub-MLRA 52-high
• Using mean rates of soil C overestimates the MC
  for Sub-MLRA 58a-low
• Not sensitive to policy design

34

• Yield Sensitivity
• 30 increase in yields over period
• t 5, 15
• 30 decrease in yields over period
• t 5, 15

35
Sensitivity to Yields Sub-MLRA 52-high
per-tonne contract
per-hectare contract
36
Sensitivity to Yields Sub-MLRA 58a-low
per-tonne contract
per-hectare contract
37
Conclusions

• Contracts based on BMPs (per hectare contracts)
  are as much as 5 times more costly than efficient
  contracts that pay per tonne of C, a degree of
  inefficiency similar to that found in studies of
  industrial regulation.
• The case study confirms that the relative
  inefficiency of per-hectare contracts varies
  spatially and increases with spatial
  heterogeneity.

38
Conclusions (continued)

• The estimates of MC are sensitive to three key
  parameters (variable) in the model
• Soil C rates
• Scale of analysis (biophysical scale only)
• Yields
• Uncertainty in an integrated biophysical/economic
  model affects both biophysical and economic
  measures
• Not always a linear mapping
• Policy design plays a key role in assessing
  impacts of uncertainty and scale