Master of Environmental Data Science
(2023)

Measuring Agricultural Adaptation to Climate Change in Zambia Using Satellite Imagery and Machine Learning

Deliverables:

Proposal

Description

Food security in sub-Saharan Africa depends heavily on local agricultural productivity and is increasingly under threat from climate change. Current data limitations imply that little is known about how farmers are adapting to climate change in these resource-poor contexts, which substantially hinders local and global policy efforts to ensure food security. This project will generate an assessment of farmer adaptation to drought across Zambia using Dr. Tamma Carleton and the CropMOSAIKS team’s previous work. In 2022, the CropMOSAIKS team demonstrated that satellite imagery and machine learning could be used to predict spatial and temporal variation in maize yields across Zambia. This was a critical first step toward understanding farmer adaptation to climate change. In this project, we will scale up this initial proof of concept, using hundreds of thousands of newly accessed household-level training data observations. These observations have a richer set of ground truth variables that will allow us to directly quantitatively assess farmer adaptation. Using the MOSAIKS machine learning process developed by Dr. Carleton and collaborators, we will use publicly available satellite imagery to produce data that can be statistically regressed against surveyed agricultural information. We will then utilize this regression to model and predict regional agricultural adaptation in response to various climate change scenarios.