U.S. Environmental Protection Agency


Date of this Version



Ecological Indicators 40 (2014) 109–116



US government work


The urban stream syndrome may not be limited to streams in urbanized watersheds. We measured thespatial pattern of impervious cover in ∼82,800 small watersheds across the conterminous United Statesby comparing watershed-based and stream-based measures of imperviousness. The watershed-basedmeasure was the commonly used watershed percentage impervious cover. The stream-based measurewas the percentage of watershed stream length flowing through impervious cover. Spatial pattern ofimpervious cover was classified on a watershed basis as proximal to streams, distal to streams, and uni-form by comparing the two measures of impervious cover. We used a classification threshold of ±5%to assign watersheds to the three classes (i.e., stream-based minus watershed-based ≥5% = proximal;watershed-based minus stream-based ≥5% = distal; else = uniform). We then applied the classification totwo impervious cover thresholds, ≥5% and ≥15%. For ≥5% and ≥15% thresholds, impervious cover wasdistributed uniformly across ∼70% and ∼86% of the watersheds, respectively. For the remaining water-sheds, the proximal spatial pattern was ∼12x and ∼4x greater than the distal spatial pattern for the ≥5%and ≥15% impervious cover thresholds, respectively. The proximal spatial pattern of impervious coveroccurred predominantly in non-urbanized watersheds, resulting in a widespread occurrence of a rela-tively high percentage of streams flowing through relatively high impervious cover in watersheds wherethe total percentage impervious cover was relatively low. The spatial pattern of change in imperviouscover between ca. 2001 and ca. 2006 did not avoid streams. Impervious cover increased in the vicinitystreams in ∼55% of the watersheds with increases in impervious cover. During this period, the lengthof streams flowing through ≥5% and ≥15% impervious cover increased by ∼9800 km and ∼6900 km,respectively.