Estimation of Alfalfa Hay Crop Yield in Northern Nevada using two different regression methods

Abstract

<b><sc>Abstract.</sc></b> Increasing pressure over water resources in the Western U.S. is forcing alfalfa (Medicago sativa L.) producers to irrigate this crop without meeting its full water demands. Crop yield models capable of simulating the complex crop-water-atmosphere relationship can be used for the development of smart water management strategies. In this work a linear model along with a random forest model were used to predict the yield of irrigated alfalfa crop in Northern Nevada. It was found that water (rain + irrigation), the occurrence of extreme temperatures and wind have a greater effect on the crop yield. Other variables that accounted for the photoperiod and the dormant period were also included in the model and are also important. The linear model had the best performance with a R² of 0.854. On the other hand, the R² for the random forest was 0.793. The linear model showed a good response to the water variability. Due to its simplicity, it is a good model to use as a benchmark to evaluate other, more complex and data intensive, alfalfa hay yield models for Northern Nevada. The random forest model can capture non-linear relationships and can be enhanced by including more data for its training.