Risky business: Temporal and spatial variation in preindustrial dryland agriculture

Traditional dryland agriculture in the Pacific island was often labor-intensive and risky, yet settlement and farming in dry areas played an important role in the development of Polynesian societies. We investigate how temporal and spatial climatic fluctuations shape variation in agricultural production across dryland landscapes. We use a model that couples plant growth, climate, and soil organic matter dynamics, together with data from Kohala, Hawai'i, to understand how temperature, rainfall, nitrogen availability, and cropping activity interact to determine yield dynamics through time and space. Due to these interactions, the statistical characterization of rainfall alone is a poor characterization of agricultural yield. Using a simple linear model of human population dynamics, we show that the observed yield variation can affect long-term population growth substantially. Our approach to analyzing spatial and temporal fluctuations in food supply, and to interpreting the population consequences of these fluctuations, provides a quantitative evaluation of agricultural risk and human carrying capacity in dry regions.