Preliminary Evaluation of a DSSAT Model for Forage Forecasting in Southern Great Plains Grasslands.

Rangelands are an important part of grassland ecosystems, characterized by high inter-annual variability in precipitation amounts leading to inter-annual variation in forage production that ultimately determines appropriate stocking rates in grazing livestock production. Accurate and timely forecast of forage production several months in advance could be helpful for grazing management, allowing farmers, ranchers, and livestock managers to maintain the sustainability of their operations against adverse conditions such as droughts. Rainfall, soil moisture, and standing crop (measured biomass) have shown potential as predictors of future levels of above-ground biomass in prior studies. Yet, none of the prior studies have resulted in a proven method for in season forage forecasts informed by soil moisture observations. The CROPGRO perennial forage model has been successfully used to simulate growth of tropical guinea grass and palisade grasses in South America but the accuracy of this model in North American tallgrass prairie is unknown. To evaluate the potential use of this model for in season forage forecasting in the US Great Plains, it was necessary to first quantify the model’s accuracy using plant parameters taken or adapted from the existing literature. During the growing season of 2012-2013, live mass and soil moisture in form of volumetric water content was measured in tallgrass prairie near Stillwater, OK. The study spanned nine patches located within three pastures under patch burn management. The uncalibrated forage model was not effective in estimating the average RMSE of live mass but it was reasonably accurate in estimating the average RMSE of soil moisture. In light of this finding, the next step would be using calibrating the model’s plant parameters from the observed live mass data to allow improved prediction accuracy from the model.