Evidence suggests that basin size and level of urbanization (LOU) are the watershed characteristics that correlate the most with peak discharges in urban watersheds. Consequently, in the U.S. virtually all regional models for peak flow prediction in developed basins consider information about these two variables. Determining watershed size is conceptually and technically trivial. In contrast, determining a basin’s urbanization level is not as easy as may seem at first glance. Of all lumped indices used to represent LOU, the most established (both in the U.S. and internationally) is the percentage of impervious area (PIA); it intuitively appears to be a reasonable measure of the degree of development in a watershed and has displayed decent predictive power in many case studies. However, it should be clear that PIA presents intrinsic conceptual limitations, as it cannot account for the spatial distribution of impervious surfaces in the basin. For example, even though two basins with similar PIA values but different spatial distributions of impervious areas relative to the stream network and the basin outlet would display different hydrologic responses and peak discharges, for the same precipitation event, they would still be considered as similar, if using PIA as a lumped descriptor. To overcome this limitation, we propose a new lumped urbanization index based on hydrologic connectivity, that enables to capture the relative contributions of impervious patches to the overall hydrologic response of the watershed, depending on their relative locations. The proposed index increases the explanatory power of regional regression models for peak flow estimation in ungauged basins.
.jpg "Francesco Dell'Aira, MSc photo")