Precipitation and evapotranspiration controls on daily runoff variability in the contiguous United States and Puerto Rico

Daily runoff variability is an important driver of fluvial erosion but is difficult to incorporate into landscape evolution models due to limited observations and incomplete understanding of hydroclimatic controls on runoff distributions. Prior work in the contiguous U.S. showed how limitations can be overcome when mean runoff is correlated with the shape of the right tail of runoff distributions. However, which probability distribution functions best capture geomorphically important events and whether patterns in the contiguous U.S. transfer to other settings remain important open questions. Our analysis of large hydroclimatic data sets from the contiguous U.S. and Puerto Rico reveals that stretched exponential distributions provide a common probabilistic framework to evaluate daily rainfall and runoff variability. In both settings, daily runoff variability is correlated with the evapotranspiration ratio, aridity index, and the ratio of wet to dry days. Surprisingly, mean storm depth (estimated from average daily precipitation during wet days only) and storm depth variability are uncorrelated with daily runoff variability in either data set. These findings suggest that first-order controls on runoff variability are processes that reduce runoff during intermediate frequency flows rather than processes that enhance the magnitude of rare floods. However, by normalizing local runoff variability by storm depth variability, some correlations collapse onto a single trend for the contiguous U.S. and Puerto Rico, suggesting a secondary role for rainfall variability on runoff variability. Taken together, this analysis provides a rationale for how hydroclimatic controls on runoff variability can be better incorporated into landscape evolution models from readily available data.