Predicted changes in climatic variability pose the most immediate threat to poorer nations and agrarian households that rely on rainfed agriculture. Many of the people who are impacted the most by climate change are also food insecure, particularly in drier areas of Africa. We use a combination of social and physical data to understand human behavior around climate change in a water scarce area of Zambia. Household surveys were conducted with approximately 1200 farmers across Zambia at the end of the 2015-6 growing season. We examine farmer’s perceptions of how to define the onset of the rainy season, historic changes in the onset of the rainy season, and heuristics that farmers use when determining when to plant their staple crop maize. This allows us to explore the relationship between heuristics and actual farmer behavior, particularly planting dates. The social data is then contrasted with historic hydroclimate data to examine whether farmer’s perceptions of the rainy season changes are consistent with the 5km-daily resolution satellite-derived precipitation from Climate Hazards Group InfraRed Precipitation with Station (CHIRPS) dataset (Funk et al., 2015). There is variation in the heuristics farmers use to determine rainy season onset and this influences the perceptions farmers have of when they believe the rains arrive and when to plant. We find that farmers perceive the onset of the rains to be getting later in most provinces, which impacts the dates they decide to plant their staple crops. Farmer perceptions are relatively consistent with CHIRPS data in drier provinces of Zambia-- demonstrating delayed onset of rainy season, slight decline in total annual rainfall, and an increase in rainfall variability over time. Investing in collection of precipitation data and developing a more robust information exchange system with farmers can improve agricultural decision-making among smallholder farmers.