Coarse resolution (30-m) impervious cover (IC) tends to underpredict total IC compared to fine (1-m) resolution. Furthermore, using % IC as a metric to describe urban watersheds is problematic because parameters such as flashiness that correlate linearly over the full range of % IC (including non-urban and highly urban watersheds) do not correlate in the middle range between 10 to 40% IC. Including end members in the metrics applied when comparing watersheds in the middle range of % IC could create a bias based on non-urban and highly urban watersheds.
Read more in a paper published in Hydrological Processes with coauthors from the NSF-funded Urban Critical Zone project.