Food security and social stability in Africa: New estimation methods for data-driven climate projections in data-sparse regions

  • Climate change threatens food security and social stability in Africa, but data scarcity makes it difficult to quantify these threats.
  • Our approach uses outcomes data (on agricultural yields, conflict incidence, and crime rates) from around the world to flexibly estimate how sensitive a location is to temperature based on its adaptive capacities.
  • By pairing these results with high-resolution seasonal climate forecasts, we can predict future climate-induced crime and conflict risk and reduction in crop yields. 
  • Our findings suggest that 24 hours of exposure to a temperature of 40°C results in maize yield losses of up to 12% (depending on location), relative to the yield at the optimal maize growing temperature of 29°C.
  • A one standard deviation increase in temperature is expected to cause a 10.8% average increase in conflict incidence and 16.2% average increase in the violent crime rate. 
  • The International Rescue Committee is prototyping this system, which can guide crime and conflict prevention efforts immediately. 

A warming climate threatens food security and social stability in many parts of Africa. However, these locations often lack the data necessary for adequately assessing such climate risks. In this project, we develop methods to overcome this scarcity of data and obtain Africa-wide estimates of temperature sensitivities at high spatial resolution.

Our approach uses outcomes data (on agricultural yields, conflict incidence, and crime rates) from around the world to flexibly estimate how sensitive a location is to temperature based on its adaptive capacities, determined using measures of physical and socioeconomic characteristics such as long-run climate, income per capita, prevalence of irrigation, and how urban the location is. We generate high-resolution datasets of these measures and thereby extrapolate locally-relevant temperature sensitivities throughout Africa.

We demonstrate how these estimates can be used to develop climate early warning systems by pairing our results with high-resolution seasonal climate forecasts to generate predictions of future climate-induced crime and conflict risk. The International Rescue Committee (a global humanitarian aid, relief, and development nongovernmental organisation) is prototyping this system, which can guide crime and conflict prevention efforts immediately.

Results from a sister project focused on India are available here.

Outputs