Regression Models in Forecasting Crop Yield under Climate Change Scenarios

Abstract

This paper reviews documented regression models applied in simulating the future crop yields under uncertain climatic futures across different parts of the world. The paper also evaluates the appropriateness of regression models in forecasting wheat yields under different climate change scenarios in Kenya. Crop Yield in this case is considered as the final recorded harvest while climate change scenarios as Global Circulation Model (GCM) climate change scenarios based on Intergovernmental Panel on Climate Change (IPCC) approved modeled data. Even though linear regression models appear to dominate literature relating to climate-crop simulations, other forms of regression too have been used in different parts of the world and are appropriate for use in wheat yield prediction in Kenya given different projected future climates. The choices of regression model depend on the nature and number of climatic variables factored in the model for extrapolation of the grain yield. Further, for reliable results, regression models require evaluation in-terms of time, applicability under varied surface physiognomies of an area, and statistical strength to affirm their efficacy in specific studies.