NorWeST Stream Temperature Points


Description Spatial Attributes

Theme: stream, water, temperature, hydrography

These data represent modeled stream temperatures for a portion of a larger dataset known as the Great Northern Landscape Conservation Cooperative (GNLCC) ( This metadata record is a combined description for two spatial data feature types, vector lines and points, which cover the same geographic area. The line features are derived from NHDPlus ( (USEPA and USGS, 2010) stream lines and the point data represent 1 km intervals along the NHDPlus stream network. Both datasets contain identical modeled stream temperature attributes. These modeled stream temperatures were generated as part of the U.S. Forest Service NorWeST stream temperature project

These data reside in ESRI shapefile format, ArcGIS version 9.3. The line and point shapefile extents correspond to NorWeST production units, which generally relate to 6 digit (3rd code) hydrologic unit codes (HUCs). 

August mean stream temperature was the metric selected to be modeled in the NorWeST temperature model. Use of this metric allowed the largest proportion of data in the NorWeST observed temperature database to be used (~80%), which facilitated calibration of the model to thousands of unique stream sites across the region.

The vector stream line data were derived from the NHDPlus data through a process referred to as reconditioning. This reconditioned data set was modified from the original NHDPlus data to ensure stream connectivity, which was required to fit spatial statistical models to the stream network data. Braided channels, most canals, and disconnected streams were deleted from NHDPlus. Additionally, where three or more segments converged into a single downstream segment the stream layer was manually edited to offset two of the three segments. Because many stream segments were deleted, this dataset does not contain all of the line features of the original NHDPlus data. The FLoWS and STARS toolboxes ( were used to identify topological errors and generate the final spatial layer. The stream lines were further processed into 1 km segments to be used as input for the NorWeST stream temperature model.

The point shapefiles correspond to the mid-point location for each 1 km stream segment. Stream temperatures were modeled at each point location. Modeled temperature values were subsequently attributed back to the 1 km stream line dataset.

Stream temperatures were modeled from a set of covariate predictors using spatial statistical software called SSN and STARS (Ver Hoef et al. 2006). (

The spatial covariates used for modeling stream temperature were derived from various sources as described below:
1. Air temperature_August (?C). Mean August air temperature across the river basin derived from the dynamically downscaled NCEP RegCM3 reanalysis (Hostetler et al. 2011). Data were downloaded from the USGS Regional Climate Downscaling website ( 
2. Stream discharge_August (m3/s). Mean August stream discharge across the river based derived from USGS flow gages on streams with minimal water abstraction or storage reservoirs. Data were downloaded from the NWIS website (
3. Elevation (m). Elevation at stream temperature sites was used to represent the vertical trend towards cooler air temperatures. Data were obtained from the 30-m resolution digital elevation model associated with NHDPlus (USEPA and USGS, 2010). Data were downloaded from
4. Latitude (m). The y-coordinate at stream temperature sites from the Albers Equal Area projection was used to represent latitude and the poleward trend towards cooler air temperatures.
5. Canopy %. The percent canopy variable from the 2001 version of the National Land Cover Database (NLCD; Homer et al., 2007) was used to represent stream shade at each temperature site.  Canopy % values in areas with recent wildfires between 2001 and 2008 were reduced based on U.S. Forest Service burn severity data following procedures developed by Miller et al. (2009).  NLCD data were downloaded from
6. Cumulative drainage area (km sq.). The value of CUMDRAINAG in NHDPlus (USEPA and USGS, 2010) at each stream temperature site was used to represent stream size and amount of insolation. It was assumed that larger streams had been exposed to insolation over a greater length and were less shaded by adjacent riparian vegetation.  Data were downloaded from
7. Stream slope %. The stream slope value in NHDPlus (USEPA and USGS, 2010) at a stream temperature site. It was assumed that slope affects flow velocity and equilibration time to local heating conditions. Steeper slopes, therefore, should negatively affect stream temperatures because conditions further upstream at higher elevations have greater influence on local temperatures.  Data were downloaded from
8. Mean annual precipitation (mm). The value of AREAWTMAP in NHDPlus (USEPA and USGS, 2010) at each stream temperature site. Areas with high annual precipitation may have higher water yields that cool streams.   Data were downloaded from
9. Base flow index (BFI). The value of the base flow index (Wolock, 2003) at a stream temperature site. Streams with larger baseflows and groundwater contributions are thought to be colder than other streams and potentially less sensitive to climate warming. Data were downloaded from
10. Glacier %. The percentage of the catchment area classified as glacier at each temperature site. Summaries were computed using a standard flow accumulation routine. This covariate represents the local cooling effect that glaciers may have on downstream temperatures.  Data were downloaded from
11. Lake %. The value of NLCD11PC in NHDPlus (USEPA and USGS, 2010), which is the percentage of the catchment area classified as open water, at a temperature site. This covariate represents the warming effect that natural lakes and many reservoirs have on downstream temperatures. Data were downloaded from
12. Tailwater. Categorical predictor variable coded as 0/1 to indicate whether a stream temperature site is downstream from a reservoir that creates an anomalously cold tailwater.

Using the SSN and STARS tools along with the covariate predictors, various mean August stream temperature scenarios were modeled. The scenarios include the 19 year average from 1993-2011, the 10 year average from 2002-2001, and single year scenarios for the years 1993 through 2011.

Referenced Cited:

Homer, C., Dewitz, J., Fry, J., Coan, M., Hossain, N., Larson, C., Herold, N., McKerrow, A., VanDriel, J.N., and Wickham, J. 2007. Completion of the 2001 National Land Cover Database for the Conterminous United States. Photogrammetric Engineering and Remote Sensing, 73:337-341.

Hostetler, S.W., J.R. Alder, and A.M. Allan. 2011. Dynamically downscaled climate simulations over North America: Methods, evaluation and supporting documentation for users: U.S. Geological Survey Open-File Report 2011-1238, 64 p. website:

USEPA and USGS, 2010. NHDPlus Version 1 (NHDPlusV1) User Guide, available online at

Miller, J.D., E.E. Knapp, C.H. Key, C.N. Skinner, C.J. Isbell, R.M. Creasy, J.W. Sherlock, 2009. Calibration and validation of the relative differenced Normalized Burn Ratio (RdNBR) to three measures of fire severity in the Sierra Nevada and Klamath Mountains, California, USA. Remote Sensing of the Environment, 113:645-656.

Ver Hoef, J.M., E.E. Peterson, and D. Theobald. 2006. Spatial statistical models that use flow and stream distance. Environmental and Ecological Statistics 13:449-464.

Wolock, D.M. 2003. Base - Flow Index Grid for the Conterminous United States. U.S. Geological Survey open-file report 03-263, USGS, Lawrence, KS.

These data were originally intended to be used for managing biological resources and predicting species distributions that are affected by August mean stream temperature.

Supplementary Information


Status of the data

Time period for which the data is relevant

Publication Information

Data storage and access information

Details about this document