On the influence of climate and land use change on monthly baseflow across the U.S. Midwest

This study evaluates the role played by changes in the climate system and land use in the observed monthly baseflow records from 1966-2015 for 458 U.S. Geological Survey sites across the U.S. Midwest. We developed parsimonious statistical models in which monthly baseflow is related to any combination of four predictors (precipitation, temperature, antecedent wetness, and agricultural intensity). We found that precipitation and antecedent wetness were the strongest predictors for all months, pointing to the role of water availability and infiltration in driving baseflow. Temperature was also an important predictor in the winter and spring where snow-melt processes are the most relevant. Agricultural intensity was selected in the Corn Belt region during the growing season (from April to August). Overall, the goodness-of-fit for our models and cross validation strongly support our modeling results for all months. Differences in model selection reported here can aid water managers in decision making for water availability, food security and economic growth.