Climate change, spatial spillovers, and sustainable productivity of Teff, Maize, and wheat across Ethiopian districts using integrated climate and agricultural data from 2000 to 2024

Abstract

Climate change poses growing threats to cereal-based food systems in Sub-Saharan Africa, where agricultural output relies heavily on weather patterns and on interconnected geographic areas. In Ethiopia, key crops such as teff, maize, and wheat are crucial to national food security, but there is limited empirical evidence on how climate variability and spatial spillovers jointly affect their productivity. This research tackles this issue by investigating the spatial and temporal effects of climate change on cereal yields at the district level throughout Ethiopia from 2000 to 2024. The study uses a dynamic Spatial Durbin Model with harmonized district-level yield data from the Central Statistical Agency (CSA), high-resolution rainfall data from CHIRPS, temperature and heat-stress metrics from ERA5, and vegetation dynamics from MODIS NDVI. The modeling framework clearly accounts for the persistence of yields over time, spatial dependence, and spillover effects across districts. Climate factors comprise seasonal precipitation, precipitation anomalies, average and peak temperature, consecutive dry days, and days of heat stress. The findings show robust spatial concentration of yields and significant statistical effects of climate on all crops. A 1 °C increase in maximum temperature reduces maize yields by approximately 6–8%, wheat yields by 4–6%, and teff yields by 3–5%. Seasonal precipitation boosts yield by 4–6% for every 100 mm, with the greatest impact observed for teff. Spatial spillovers are significant and mainly detrimental for temperature-related factors: heat stress in one area lowers adjacent maize yields by as much as 9%, whereas beneficial rainfall spillovers raise nearby teff yields by approximately 5–6%. Long-term effects surpass short-term impacts, indicating accumulated climate stress and spatial feedback processes. These results show that climate threats to cereal yields in Ethiopia are not limited to specific areas and that adaptation measures have effects that extend beyond district boundaries. Socioeconomic and soil conditions were identified as factors that could either lessen or intensify these effects. Disregarding spatial spillovers leads to an underappreciation of climate impacts and a misallocation of adaptation funds. The study provides robust, spatially explicit evidence to inform climate-smart agricultural planning, emphasizing the need for coordinated, landscape-level adaptation strategies to enhance resilience and food security under climate change, which are crucial for maintaining productivity.