Climate is getting hotter - both in fact and as a research topic. Yet there is little scientific consensus on the conflict potential carried by changes in climate and weather patterns. Previous studies investigating the links between drought and conflict have relied on precipitation-based measures of drought. However, in the same way that peace is not just the absence of war, drought is not just the absence of precipitation. What is needed is a greater focus on the relationship between the impacts of drought and conflict potential. In this thesis, I identify three contrasting drought indicators that are assumed to capture different theoretical concepts of drought: a precipitation-based measure (SPEI); a vegetation-based measure (NDVI); and a socio-economic measure (EM-DAT). These three measures are not only assumed to capture different theoretical concepts of drought, but also different stages in the drought cycle. Lack of precipitation may lead to less vegetation, loss of crops and deterioration of pasture, which in turn may spark a socio-economic disaster. By using these three measures, I answer the research question: Do different conceptualisations of drought affect the likelihood of communal conflict? Drawing on novel high-resolution data on communal conflict events and droughts in Sub-Saharan Africa from 1989 to 2014, this thesis evaluates the relationship between drought and communal conflict on the local level. Results from mixed-effects multilevel logistic regression show that all three drought measures are associated with a higher risk of communal conflict, but the effect differs across the various measures. The higher the measurable physical impact of drought, or the closer we get to measuring the socio-economic impacts of the drought, the higher the risk of experiencing communal conflict.