Skip to main content
U.S. flag

An official website of the United States government

Predicting the likelihood of altered streamflows at ungauged rivers across the conterminous United States

July 23, 2013

An approach is presented in this study to aid water-resource managers in characterizing streamflow alteration at ungauged rivers. Such approaches can be used to take advantage of the substantial amounts of biological data collected at ungauged rivers to evaluate the potential ecological consequences of altered streamflows. National-scale random forest statistical models are developed to predict the likelihood that ungauged rivers have altered streamflows (relative to expected natural condition) for five hydrologic metrics (HMs) representing different aspects of the streamflow regime. The models use human disturbance variables, such as number of dams and road density, to predict the likelihood of streamflow alteration. For each HM, separate models are derived to predict the likelihood that the observed metric is greater than (‘inflated’) or less than (‘diminished’) natural conditions. The utility of these models is demonstrated by applying them to all river segments in the South Platte River in Colorado, USA, and for all 10-digit hydrologic units in the conterminous United States. In general, the models successfully predicted the likelihood of alteration to the five HMs at the national scale as well as in the South Platte River basin. However, the models predicting the likelihood of diminished HMs consistently outperformed models predicting inflated HMs, possibly because of fewer sites across the conterminous United States where HMs are inflated. The results of these analyses suggest that the primary predictors of altered streamflow regimes across the Nation are (i) the residence time of annual runoff held in storage in reservoirs, (ii) the degree of urbanization measured by road density and (iii) the extent of agricultural land cover in the river basin.

Citation Information

Publication Year 2013
Title Predicting the likelihood of altered streamflows at ungauged rivers across the conterminous United States
DOI 10.1002/rra.2565
Authors Ken Eng, Daren Carlisle, David M. Wolock, James A. Falcone
Publication Type Article
Publication Subtype Journal Article
Series Title River Research and Applications
Series Number
Index ID 70046061
Record Source USGS Publications Warehouse
USGS Organization National Research Program - Eastern Branch