Using Environmental Variables to Distribute Human Populations Spatially With an Evaluation in California

Climate change research, which focuses on anthropogenic factors, may be hampered by the lack of population data available at high spatial resolutions, particularly for developing countries or historical time periods. A refined model that distributes human populations spatially is developed and tested using California as the test case. The model is based on the concept of "habitability". Habitability is the propensity for a place to sustain a fixed human population. Habitability is forced by environmental factors, specifically elevation, topographic accessibility, river proximity, and annual water surplus. Model-estimated population distributions are compared to 1990 California census data at the block-group level to evaluate the efficacy of the habitability model. Results show that the model can realistically distribute California's total population based on the spatial distributions of a few environmental variables. Adding annual water surplus actually degraded the model's performance, which implies that it may not require a hydroclimatic variable. The model does not resolve well the very-high urban densities. Although further refinement is necessary to adequately represent the urban populations, the habitability model performance suggests that it is a promising approach to population mapping, as well as to quantifying population-environment variables.