Modeled Streamflow Metrics on Small, Ungaged Stream Reaches in the Upper Colorado River Basin: Data Release

Product Type: 

Data Release

Year: 

2016

Author(s): 

Reynolds, L.V. and Shafroth, P.B.

Suggested Citation: 

Reynolds, L.V. and Shafroth, P.B., 2016, Modeled Streamflow Metrics on Small, Ungaged Stream Reaches in the Upper Colorado River Basin: Data: U.S. Geological Survey Data Release, http://dx.doi.org/10.5066/F7H9938M.

Modeling streamflow is an important approach for understanding landscape-scale drivers of flow and estimating flows where there are no streamgage records. In this study conducted by the U.S. Geological Survey in cooperation with Colorado State University, the objectives were to model streamflow metrics on small, ungaged streams in the Upper Colorado River Basin and identify streams that are potentially threatened with becoming intermittent under drier climate conditions. The Upper Colorado River Basin is a region that is critical for water resources and also projected to experience large future climate shifts toward a drying climate. A random forest modeling approach was used to model the relationship between streamflow metrics and environmental variables. Flow metrics were then projected to ungaged reaches in the Upper Colorado River Basin using environmental variables for each stream, represented as raster cells, in the basin. Last, the projected random forest models of minimum flow coefficient of variation and specific mean daily flow were used to highlight streams that had greater than 61.84 percent minimum flow coefficient of variation and less than 0.096 specific mean daily flow and suggested that these streams will be most threatened to shift to intermittent flow regimes under drier climate conditions. Map projection products can help scientists, land managers, and policymakers understand current hydrology in the Upper Colorado River Basin and make informed decisions regarding water resources. With knowledge of which streams are likely to undergo significant drying in the future, managers and scientists can plan for stream-dependent ecosystems and human water users.

The random forest models we used to project flow metrics to small, ungaged streams varied in model performance between 45.3- and 82.55-percent variance explained. We successfully projected low flow metrics to small, ungaged reaches across the Upper Colorado River Basin using random forest models developed in Reynolds and others (2015) and described in the “Methods” section of the accompanying Digital Data Series publication. We produced this collection of seven flow metric datasets, one for each modeled flow metric, for small stream reaches across the Upper Colorado River Basin. We produced an eighth dataset for the collection showing modeled intermittency status for each stream reach, including streams that are potentially threatened with intermittency because of drier conditions.'

For a complete description of this data collection see the USGS Digital Data Series 974 (Reynolds and Shafroth, 2016).

 

Related Projects

      

Lindsay Reynolds
Lindsay ReynoldsPat Shafroth