«Working Paper 05-WP 397 June 2005 Center for Agricultural and Rural Development Iowa State University Ames, Iowa 50011-1070 ...»
Historical Development and Applications of the
EPIC and APEX Models
Philip W. Gassman, Jimmy R. Williams, Verel W. Benson,
R. César Izaurralde, Larry M. Hauck, C. Allan Jones, Jay D. Atwood,
James R. Kiniry, and Joan D. Flowers
Working Paper 05-WP 397
Center for Agricultural and Rural Development
Iowa State University
Ames, Iowa 50011-1070
Authors information: Philip W. Gassman, Center for Agricultural and Rural Development, Iowa
State University; Jimmy R. Williams, Blackland Research and Extension Center, Texas A&M University; Verel W. Benson, Food and Agricultural Policy Research Institute, University of Missouri-Columbia; R. César Izaurralde, Joint Global Change Research Institute, University of Maryland; Larry M. Hauck, Texas Institute for Applied Environmental Research, Tarleton State University; C. Allan Jones, Texas Water Resources Institute, Texas A&M University, Jay D.
Atwood, USDA-NRCS, Water Resources Assessment Team, Blackland Research and Extension Center, Texas A&M University; James R. Kiniry, USDA-ARS, Grassland Research Center, Texas A&M University; Joan D. Flowers, Carter & Burgess, Fort Worth, Texas.
This paper was presented at the 2004 ASAE/CSAE annual international meeting in Ottawa, Ontario, Canada, as Meeting Paper No. 042097.
This paper is available online on the CARD Web site: www.card.iastate.edu. Permission is granted to reproduce this information with appropriate attribution to the authors.
For questions or comments about the contents of this paper, please contact Philip Gassman, 560E Heady Hall, Iowa State University, Ames, IA 50011-1070; Ph: 515-294-6313; Fax: 515-294E-mail: email@example.com.
Iowa State University does not discriminate on the basis of race, color, age, religion, national origin, sexual orientation, sex, marital status, disability, or status as a U.S. Vietnam Era Veteran. Any persons having inquiries concerning this may contact the Director of Equal Opportunity and Diversity, 1350 Beardshear Hall, 515-294-7612.
Abstract The development of the field-scale Erosion Productivity Impact Calculator (EPIC) model was initiated in 1981 to support assessments of soil erosion impacts on soil productivity for soil, climate, and cropping conditions representative of a broad spectrum of U.S. agricultural production regions. The first major application of EPIC was a national analysis performed in support of the 1985 Resources Conservation Act (RCA) assessment. The model has continuously evolved since that time and has been applied for a wide range of field, regional, and national studies both in the U.S. and in other countries. The range of EPIC applications has also expanded greatly over that time, including studies of (1) surface runoff and leaching estimates of nitrogen and phosphorus losses from fertilizer and manure applications, (2) leaching and runoff from simulated pesticide applications, (3) soil erosion losses from wind erosion, (4) climate change impacts on crop yield and erosion, and (5) soil carbon sequestration assessments. The EPIC acronym now stands for Erosion Policy Impact Climate, to reflect the greater diversity of problems to which the model is currently applied. The Agricultural Policy EXtender (APEX) model is essentially a multi-field version of EPIC that was developed in the late 1990s to address environmental problems associated with livestock and other agricultural production systems on a whole-farm or small watershed basis. The APEX model also continues to evolve and to be utilized for a wide variety of environmental assessments. The historical development for both models will be presented, as well as example applications on several different scales.
Keywords: APEX, carbon sequestration, climate change, EPIC, modeling, soil erosion, water quality.
HISTORICAL DEVELOPMENT AND APPLICATIONS OF THE
EPIC AND APEX MODELSIntroduction The 1977 Resources Conservation Act (RCA) charged the U.S. Department of Agriculture (USDA) with the responsibility to assess the status of the nation’s soil and water resources on a regular basis. The first RCA appraisal conducted in 1980 revealed a significant need for improved technology for evaluating the impacts of soil erosion on soil productivity (Putnam, Williams, and Sawyer 1988). In response, the Erosion Productivity Impact Calculator (EPIC) model was developed by a USDA modeling team in the early 1980s to address this technology gap (Williams, Jones, and Dyke 1984; Williams 1990;
Sharpley and Williams 1990; Jones et al. 1991). The first major application of EPIC was for the second RCA appraisal in 1985, in which the model was used to evaluate soil erosion impacts for 135 U.S. land resource regions (Putnam, Williams, and Sawyer 1988).
Ongoing evolution of the model, including incorporation of additional functions related to water quality and atmospheric CO2 change, resulted in the model name eventually being changed to Environmental Policy Impact Climate (Williams et al. 1996). The latest version of EPIC features enhanced carbon cycling routines (Izaurralde et al. 2001) that are based on the approach used in the Century model (Parton et al. 1994).
Numerous applications of EPIC have been performed in the United States and in other regions of the world across a broad spectrum of environmental conditions. Example applications include assessment of sediment and nutrient losses as a function of different tillage systems, crop rotations, and fertilizer rates (Phillips et al. 1993; King, Richardson, and Williams 1996); nutrient losses from livestock manure applications (Edwards et al. 1994;
Pierson et al. 2001); nitrate-nitrogen (NO3-N) losses through subsurface tile drainage (Chung et al. 2001; Chung et al. 2002); nutrient cycling as a function of cropping system (Cavero et al. 1999; Bernardos et al. 2001); soil loss due to wind erosion (Potter et al. 1998;
Bernardos et al. 2001); climate change impacts on crop yield and/or soil erosion (FavisGassman et al.
Mortlock et al. 1991; Brown and Rosenberg 1999); losses from field applications of pesticides (Williams, Richardson, and Griggs 1992; Sabbagh et al. 1992); irrigation impacts on crop yields (Cabelguenne, Jones, and Williams 1995; Rinaldi 2001); estimation of soil temperature (Potter and Williams 1994; Roloff, de Jong, and Nolin 1998a); and soil carbon sequestration as a function of cropping and management systems (Lee, Phillips, and Liu 1993; Apezteguía, Izaurralde, and Sereno 2002). The flexibility of EPIC has also led to its adoption within several integrated economic and environmental modeling systems that have been used to evaluate agricultural policies at the farm, watershed, and/or regional scale (e.g., Taylor, Adams, and Miller 1992; Bernardo et al. 1993; Foltz, Lee, and Martin 1993; Babcock et al. 1997).
The Agricultural Policy EXtender (APEX) model (Williams et al. 1995; Williams
2002) was developed in the 1990s to facilitate multiple subarea scenarios and/or manure management strategies, such as automatic land application of liquid manure from waste storage ponds, which cannot be simulated in EPIC. The catalyst for creating APEX was the
U.S. Environmental Protection Agency (USEPA) funded “Livestock and the Environment:
A National Pilot Project (NPP),” which was initiated in 1992 to study livestock environmental problems on a watershed basis. The APEX model was used extensively for a wide range of livestock farm and nutrient management (manure and fertilizer) scenarios within the Comprehensive Economic Environmental Optimization Tool – Livestock and Poultry (CEEOT-LP), an economic-environmental modeling system developed for the NPP (Gassman et al. 2002; Osei et al. 2000; Osei, Gassman, and Saleh 2000; Osei et al. 2003a, b). It has also been applied within a number of other studies.
A brief overview of the structure of both models is presented here, followed by reviews of how the models have been applied up to the present time. Emerging applications of the two models are also discussed.
Overview of EPIC The EPIC model can be subdivided into nine separate components defined as weather, hydrology, erosion, nutrients, soil temperature, plant growth, plant environment control, tillage, and economic budgets (Williams 1990). It is a field-scale model that is designed to simulate drainage areas that are characterized by homogeneous weather, soil, Historical Development and Applications of the EPIC and APEX Models / 3 landscape, crop rotation, and management system parameters. It operates on a continuous basis using a daily time step and can perform long-term simulations for hundreds and even thousands of years. A wide range of crop rotations and other vegetative systems can be simulated with the generic crop growth routine used in EPIC. An extensive array of tillage systems and other management practices can also be simulated with the model.
Seven options are provided for simulating water erosion and five options are available for simulating potential evapotranspiration (PET). Detailed discussions of the EPIC components and functions are given in Williams, Jones, and Dyke 1984; Williams 1990;
Sharply and Williams 1990; and Williams 1995.
EPIC Applications Initial tests of the EPIC model were reported by Williams, Renard, and Dyke (1983), which were performed in support of the 1985 RCA analysis. The model was shown to replicate realistically mean surface runoff and sediment yields that were measured for three small watersheds in Falls County, Texas, and for corn, oat, or soybean yields that were measured in Iowa, Missouri, and Ohio. Nearly 12,000 100-year EPIC simulations were then performed for different crop, tillage, soil, climate, and conservation practice combinations to support the RCA analysis, which included economic assessments conducted with a linear programming model (Putnam, Williams, and Sawyer 1988).
The EPIC model has continued to evolve and to be applied to an ever-increasing range of scenarios since the 1985 RCA analysis. Some applications have focused specifically on testing of different EPIC components, which in some cases resulted in modifications to existing routines and improved model performance. Other enhancements and refinements have been made to the model to facilitate the interest of various users or to meet the needs of specific applications. Table 1 lists examples of modifications that have been made to the EPIC model over roughly the past 15 years. “Spin-off” versions of the model have also been developed by several users for region- or task-specific applications, for example, the AUSCANE model created by Jones et al. (1989) to simulate Australian sugarcane production. Trends in the use of EPIC are highlighted here as a function of example applications for key EPIC indicators such as estimation of crop yields and soil erosion losses by wind and water.
Subcategories are also delineated for notable region- or application-specific uses.
4 / Gassman et al.
Improved nitrogen fixation routine for legume crops that Bouniols et al. (1991) calculates fixation as a function of soil water, soil N, and crop physiological stage
Incorporation of CO2 and vapor pressure effects on radiation Stockle et al. (1992a) use efficiency, leaf resistance, and transpiration of crops Incorporation of functions that allow two or more crops to be Kiniry et al. (1992b) grown simultaneously
Improved crop growth parameters for cereal, oilseed, and Kiniry et al. (1995) forage crops grown in the North American northern Great Plains region
Crop Growth and Yield Studies A key output provided by EPIC is crop yield predictions. Several studies have been performed in the United States and other countries that focused specifically on testing the accuracy of EPIC crop growth and yield predictions. Such tests have also been incorporated as parts of other studies. One of the most comprehensive tests of the crop growth submodel was performed by Williams et al. (1989), who describe the results of testing an updated EPIC crop growth model (Table 1) for simulated barley, corn, rice, soybean, sunflower, and wheat yields at several U.S. locations and for sites in Asia, France, and South America. The predicted yields were compared with measured yields for periods ranging from 1 to 11 years. The average predicted yields were always within 7% of the average measured yields, and there was no significant difference between any of the simulated and measured yields at the 95% confidence level. However, r2 statistics computed between the simulated and measured yields of the six crops ranged from relatively strong values of 0.80 and 0.65 for wheat and corn to only 0.20 for barley and soybean.
EPIC-predicted yields have been shown in other studies to replicate accurately both mean and annual yields for different crops and conditions. Bryant et al. (1992) found that EPIC accurately predicted mean and annual corn yields measured for 38 irrigation stress experiments conducted during the 1975-77 period at Bushland, Texas, after the effects of a hail storm were accounted for in 1976. Gray et al. (1997) reported r2 values of 0.82 and
0.85 for corn yield production functions developed from EPIC simulations that were performed for irrigation timing experiments conducted at Bushland, Texas, during the years 1990-93. Geleta et al. (1994) found that predicted mean and annual yields accurately reflected irrigated corn, sorghum, and winter wheat yields measured near Goodwell in the Oklahoma Panhandle region between 1984 and 1988 using a version of EPIC called EPIC-PST (Sabbagh et al. 1991). Parsons, Pease, and Martens (1995) reported that EPIC predicted mean yields accurately and explained 55% to 89% of the measured yield variance for five of six treatments (excluding 1986 from four of the treatments) for corn grown during the 1978-93 period on three Virginia soil types fertilized with either inorganic fertilizer or a “heavy” manure application. Cavero et al. (1997) and Cavero et al.
(1999) found that EPIC replicated measured annual yields accurately, except when disease problems affected yields, for irrigated tomato, safflower, and winter wheat grown 6 / Gassman et al.