CROPPING SYSTEMS Evaluation of GPFARM for Dryland Cropping Systems in Eastern Colorado

than 90% were interested in using farm management decision support software (Frasier et al., 1997). The GPFARM is an ARS decision support system for strategic (longsame survey also showed that 57% of 219 producer term) planning. This study evaluated its performance for comparing alternative dryland no-till cropping systems and established limits of respondents were interested in a farm management deaccuracy for eastern Colorado, using data collected in 1987 through cision support product. Central to meeting this chal1999 from an ongoing long-term experiment at three locations along lenge, the USDA-ARS Great Plains Systems Research a gradient of potential evapotranspiration (PET) (Sterling, low PET; Unit (GPSR), in a collaborative effort with Colorado Stratton, medium PET; and Walsh, high PET). The crop rotations, State University, developed the GPFARM (Great Plains which included winter wheat (Triticum aestivum L.), corn (Zea mays Framework for Agricultural Resource Management) L.), sorghum [Sorghum bicolor (L.) Moench], proso millet (Panicum decision support system (DSS) (Ascough et al., 2002; miliaceum L.), and varying fallow periods, were wheat–fallow, wheat– Shaffer et al., 2000). The general purpose of GPFARM corn–fallow, and wheat–corn–millet–fallow at Sterling and Stratton is to serve as a whole farm/ranch DSS in strategic planand wheat–fallow, wheat–sorghum–fallow, and wheat–sorghum–millet– ning across the Great Plains, for production, economic fallow at Walsh. The ranges of relative error (RE) of simulated mean and root mean square error (RMSE) were total soil profile water and environmental impact analysis, and site-specific dacontent (RE: 0 to 23%; RMSE: 38 to 76 mm water), dry mass grain tabase generation, from which alternative agricultural yield (RE: 27 to 84%; RMSE: 419 to 2567 kg ha 1), dry mass crop management systems can be tested and compared. The residue (RE: 5 to 42%; RMSE: 859 to 1845 kg ha 1), and total soil term strategic planning is defined here as long-term profile residual nitrate N (RE: 42 to 32%; RMSE: 26 to 78 kg ha 1). planning (e.g., choice of sustainable crop rotation, choice GPFARM simulations agreed with observed trends and showed that of tillage/residue management system, etc.) as opposed productivity and water use efficiency increased with cropping intensifito tactical planning (e.g., scheduling of irrigation, chemication and that Stratton was the most productive and Walsh the cal application, harvesting, etc.), which is done in real least. GPFARM (v. 2.01) was less suited for year-to-year grain yield time. Agricultural consultants and progressive farmers prediction under dryland conditions but has potential as a tool for or ranchers are targeted as the primary users of studying long-term interactions between environment and crop management system. Future development and applications of GPFARM GPFARM. The user requirements for GPFARM were must account for crop-specific responses to stress, detailed hydrology, identified by ARS customer focus groups in the Great better understanding of root uptake processes, and spatial variability Plains, comprised of farmers, ranchers, agricultural conto give more accurate grain yield predictions in water-stressed envisultants, and NRCS and extension professionals. The ronments. major requirements were that (i) the DSS be simple to understand and easy to use and (ii) have minimum input data and parameter requirements. S agriculture demands consideration of The GPFARM model is an aggregate of modules many interrelated factors, processes, resources, and taken from existing agricultural water quality models institutions. In the Great Plains, there has been a recogand new modules specifically developed for GPFARM. nized need for a systems approach for agricultural reFor example, the crop growth module is based on the search and development for attaining sustainability (AsEPIC generic crop growth model that has been widely cough et al., 2002). Peterson et al. (1993) proposed that tested for various crops (e.g., Steiner et al., 1987; Wila systems approach to the study of soil and crop manageliams et al., 1989; Martin et al., 1993; Moulin and Beckie, ment problems is useful for testing present research 1993; Kiniry et al., 1995; Jara and Stockle, 1999) while knowledge to answer practical agricultural problems the water balance module, which is a simplification of and simultaneously identify gaps in basic research the RZWQM (Ahuja et al., 2000) water balance rouknowledge. Likewise, there has been a recognized need tines, has not been extensively tested. for system-level decision support tools for agricultural Most of the modules have been independently tested advisors and producers. In a 1995 Great Plains survey to varying degrees, but there is a need to evaluate of 121 county extension directors, 173 NRCS district GPFARM at the system level to see how well the modconservationists, and 95 agricultural consultants, more Abbreviations: CCC, Colorado Climate Center; CV, coefficient of variation; DSS, decision support system; ET, evapotranspiration; HI, A.A. Andales and L.R. Ahuja, USDA-ARS Great Plains Syst. Res. harvest index; LAImax , maximum leaf area index; PET, potential Unit, P.O. Box E, Fort Collins, CO 80522; and G.A. Peterson, Dep. evapotranspiration; RE, relative error of mean; RMSE, root mean of Soil and Crop Sci., Colorado State Univ., Fort Collins, CO 80523. square error; WCF, wheat–corn–fallow (rotation); WCMF, wheat– Received 29 Aug. 2002. *Corresponding author ([email protected]. corn–millet–fallow (rotation); WC(S)F, wheat–corn (or sorghum)– gov). fallow (rotation); WC(S)MF, wheat–corn (or sorghum)–millet–fallow (rotation); WF, wheat–fallow (rotation); WSF, wheat–sorghum–fallow Published in Agron. J. 95:1510–1524 (2003).  American Society of Agronomy (rotation); WSMF, wheat–sorghum–millet–fallow (rotation); WUE, water use efficiency. 677 S. Segoe Rd., Madison, WI 53711 USA