Multi-Crop Plant Growth Modeling for Agricultural Models and Decision Support Systems

Agricultural models and decision support systems are becoming increasingly available for a wide audience of users. The Great Plains Framework for Agricultural Resource Management (GPFARM) DSS is a strategic planning tool for farmers, ranchers, and agricultural consultants that incorporates a science simulation model with an economic analysis package and multi-criteria decision aid for evaluating individual fields or aggregating to the entire enterprise. The GPFARM DSS is currently being expanded to include 1) better strategic planning by simulating a greater range of crops over a wider geographic range and management systems, 2) incorporating a tactical planning component, and 3) adding a production, environmental, and economic risk component. The plant growth component within the science simulation model is subdivided into separate submodels for crops and rangeland forage. User requirements have determined that the DSS must be easy to use in terms of setup, and therefore little parameterization or calibration for a specific site can be required. Based on evaluation of both the crop and rangeland forage growth module of GPFARM, improvements are needed to more accurately simulate plant responses to varying levels of water availability. This paper presents our approach and some preliminary results for improving the plant growth models. Our approach is based on using the stand-alone plant growth model derived from the Wind Erosion Prediction System (WEPS), which is based on the EPIC plant growth model. Steps that should improve the plant growth models include 1) incorporate modifications from work done to other models that are based on the EPIC plant growth model (e.g., GPFARM; Water Erosion Prediction Project, WEPP; ALMANAC; and Soil and Water Assessment Tool, SWAT), and 2) thoroughly evaluate how the plant processes are represented in these models. Deficiencies in adequately simulating plant growth responses to water availability can fall under two general categories: inadequate quantification of the process or omission of a needed process in the model. High priority needs identified to date include: 1) seedling emergence, 2) phenology, 3) biomass generation, 4) biomass partitioning, 5) root growth, and 6) plant stress factors. Initial work has created stand- alone submodels for predicting seedling emergence (as a function of soil water and thermal time) and phenology (by predicting specific growth stages and responses to different levels of soil water availability). Evaluation of alternative approaches for generating biomass (e.g., radiation use efficiency, transpiration use efficiency, plant growth analysis), biomass partitioning (e.g., modifications to generating LAI and partitioning coefficients partly based on better phenology prediction), and stress factors (e.g., single-most limiting, additive, multiplicative) is underway. We envision that these modifications and enhancements should improve model responses to varying levels of soil water availability.