Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

SWAP (Statewide Agricultural Production Model)

General Description

The Statewide Agricultural Production Model (SWAP) is a multi-region, multi-input and output model of agricultural production, which self-calibrates using the method of Positive Mathematical Programming (PMP). PMP SWAP simulates the effects of policy changes on cropping patterns at the extensive and intensive margins by integrating observed data into model calibration. The model chooses the optimal amounts of land, water, labor, and other input use subject to these constraints and definitionsresource constraints and other economic parameters including production costs, crop yields and prices. Profit is revenue minus costs, where revenue is price x yield per acre times total acres. The model website is at: http://swap.ucdavis.edu

Model Domain

Agricultural economics of crops in California

Developer

Howitt, Medellin-Azuara, MacEwan, and Lund. Originally at the University of California at Davis (UC Davis)

Hardware computing requirementsNot specified

Runs in MS windows environment.

Code language

Generalized Algebraic Modeling System (GAMS)

Original application

Developed as an agricultural economic component of CALVIN model of California water systemssupply system. http://calvin.ucdavis.edu

Public/proprietary and cost

For research-related applications available through UC Davis Center for Watershed Sciences or UC Merced (public).

Business application through ERA Economics (proprietary).

Physically or empirically based

Empirical

Mathematical methods used

Based on Positive Mathematical Programming (Howitt 1995) calibrated to a base datasets on crop production input use, prices, costs and yields. Constant Elasticity of Substitution (CES) production functions are generated for every crop in every region.

Input data requirements

Input data requirements include (values in parentheses refer to example source datasets, these can vary):

  • Land Use
  • Crop prices (County ag commissioners)
  • Crop Yields (UCCE crop budgets)
  • Interest rates (UCCE crop budgets)
  • Land Use costs (UCCE crop budgets)
  • Supply costs (UCCE crop budgets)
  • Labor costs (UCCE crop budgets)
  • Surface water costs (USBR/DWR/water districts)
  • Irrigation water (DWR)
  • Available water (CVPM/DWR/USBR/water district)
  • Crop Price Supply Elasticities (Russo et al. 2008)

Outputs

Outputs for 27 base regions (21 CVPM regions) in the Central Valley for 20 crop groups following DWR's CalSIMETAW classification.

  • Allocation of acreage per crop type. Input use: land, water, labor and other supplies.
  • Economic value

Pre-processing and post-processing tools

None specified.

Representation of uncertainty

None specified.

Prevalence

Used in policy analysis by both academic and state agencies.

Ease of use for public entities

Requires knowledge of GAMS software and access to program through UC Daviscode through the University of California. Alternatively, ERA Economics provides paid access to model interface and modelling services for SWAP.

Ease of obtaining information and availability of technical support

Mode Model can be obtained through UC Davis the University of California, Center for Watershed Sciences for research purposes.

ERA Economics provides services to run scenarios. Freely available manuals or information on technical model runs is not currently available.

Source code availability

Not freely available.

Status of model development

Model Version 6.1 is available for research and other application through UC Davis Center for Watershed Sciences. Version RTS (decreasing returns to scale) available through ERA economics.

Challenges for integration

Currently integrated with CALVIN, IMPLAN, and CALSIM II models. Detailed inputs. Constrained crop groups may not map well to other models. New crop groups or updated crop valuation requires calibration for economics as well as land use. Model is data driven, can be run without recoding by changing input data. Model output scale may not match with other models.

...