Building Databases to Calibrate Alfalfa Crop Models: Paving the Way for an Advanced Yield Forecasting Tool

Abstract

Escalating pressure on water resources in the western U.S. has led alfalfa (Medicago sativa L.) producers in this region to adopt irrigation practices falling below the crop’s optimal evapotranspiration levels. This has resulted in deficit irrigation and diminished yields. This study is part of a project that aims to develop an alfalfa hay Yield Forecasting Tool (YFT) that can be used to estimate the effects on yield of different irrigation management decisions. The YFT will utilize weather, soil characteristics, crop agronomics, crop management, and crop development indicators for this purpose. In the future, the YFT will be integrated into a Decision Support System that will be designed to recommend irrigation management decisions that can minimize yield losses caused by insufficient irrigation. This study aims to build comprehensive databases with varying levels of data completeness that can be used to train and test alfalfa crop growth models and machine learning algorithms that will be embedded in the YFT for robust alfalfa yield forecasting. The databases were created using 210 crop years of data from historical and ongoing field experiments in the Texas High Plains and Northern Nevada. Managing extensive information from diverse experimental domains with varying data completeness necessitates comprehensive databases for training different crop growth models and machine learning algorithms. This study will describe the generation of such databases.