A Comparison of Techniques for Estimating NDVI for Agricultural Intervention Impact Assessment

Food security is a major issue in Nepal in the face of environmental changes. Agricultural interventions have been performed in Nepal with a goal of increasing crop productivity to combat food insecurity. In this study, we compare three techniques for estimating NDVI, serving as a proxy for agricultural productivity, over fields in western Nepal - a Multi-Linear Regression approach, a Random Forest approach, and an Artificial Neural Network approach. Assessment of agricultural productivity is important for understanding the effectiveness of agricultural interventions. The methodology heavily utilizes space-borne remote sensing datasets and the research addresses three core questions - whether we can accurately predict NDVI at a single time using remote sensing, whether we can accurately predict the temporal response of NDVI time series in a given year using remote sensing, and if we can develop a framework for assessing the impact of interventions using remote sensing and statistical techniques. The results of this work will help to build capacity in agriculture intervention assessment by giving decision makers and the scientific community a better idea of the effectiveness of Multi-Linear Regression, Random Forest, and Artificial Neural Networks for the purpose of estimating crop production, a well-established need in the region.