Soil moisture index improves models of forage production in central Iran

This study is focused on the relative importance of bioclimatic variables and soil moisture index (SMI) for predicting forage production in Central Iran rangelands. We used a dataset of 93 forage production collected from 16 rangeland sites during 8 years in Isfahan province. Climatic and topographic variables were obtained using a digital elevation model and geographical information system. SMI is derived using a triangle space concept between the land surface temperature and normalized difference vegetation index from MODIS satellite data over Isfahan province. Using Nonparametric Multiplicative Regression model, we produced two distinct models to predict forage production, the first model with climatic and topographic variables (reduced model) and the second model with climatic, topographic and SMI variables (complete model). The inclusion of SMI significantly improved the quality of forage production predictions within the study area. The forage production is predicted to vary from 41.4 to 474 kg ha^?1 in first model and 20.5 to 485 kg ha^?1 in second model. We anticipated that using SMI would improve standing crop estimates relative to using bioclimatic variables. The cross-validated R² (×R²) increased from 57% to 63% when SMI is added to NPMR model.