# Limitations

The StreamStats Web application provides access to automated procedures and very large, complex data sets. These data sets are known to contain occasional errors. Users are advised to carefully check all results for accuracy and to exercise their own professional judgment in evaluating the appropriateness of the results for their application. Basin delineations, in particular, frequently have been found to be erroneous. StreamStats provides tools and base maps useful for verifying the accuracy of the basin delineations and for correcting them, if necessary.

Estimates of streamflow statistics for user-selected ungaged sites can be obtained by two methods in StreamStats. The *Estimate Flows Using Regression Equations* tool provides estimates by solving USGS-developed regression equations. The *Generate Flows Based on Similar Gages* tool estimates streamflow statistics by applying the flows per unit area for streamflow statistics at a nearby gaging station to the drainage area for the ungaged site. The theoretical basis for these methods and the outputs provided by StreamStats are explained in more detail on the Ungaged Sites page.

## Limitations for Estimates Obtained by Use of the *Estimate Flows Using Regression Equations* Tool

Estimates obtained by use of the *Generate Flow Statistics* tool assume natural flow conditions at the ungaged site. If human activities such as dam regulation and water withdrawals substantially affect the timing, magnitude, or duration of flows at a selected site, the estimates provided by StreamStats should be adjusted by the user to account for those activities.

The *Generate Flow Statistics* tool can be used to obtain regression equation-based estimates of streamflow statistics under natural conditions for USGS data-collection stations that are affected by human activities. Users should not assume, however, that the differences between the data-based estimates for the stations and the regression equation-based estimates are equivalent to the effects of the human activities on streamflow at the stations because there are errors associated with both sets of estimates.

Extrapolation occurs when one or more of the basin characteristics needed to solve the applicable regression equations for an ungaged site are outside the ranges of basin characteristics for the sites used to develop the regression equations. When extrapolation occurs, StreamStats provides a warning in the output to indicate that the basin characteristics are out of range. StreamStats will provide extrapolated estimates for ungaged sites, as those estimates still are often the best estimates that can be obtained for the site; however, the errors associated with extrapolated estimates are unknown. As a result, StreamStats does not provide indicators of the errors for the estimates.

**Example of StreamStats basin characteristics table with warning because the drainage area is out of the range of applicability for the low-flow equations**

Extrapolation also can occur when a basin characteristic that is not included as an explanatory variable in the regression equations is out of the range of the values of that basin characteristic for the stations used in the regression analysis. For example, an ungaged site may be located just downstream from a large lake, such that the percentage of the basin covered by open water is extraordinarily large compared to the percentages of open water for the stations used in the regression analysis. If the percentage of open water does not appear as an explanatory variable in the regression equations applicable to the location of the ungaged site, then StreamStats will not detect that extrapolation is occurring.

Users should carefully evaluate their sites of interest to determine if the available regression equations for that location are suitable for their intended purpose, and if extrapolation is occurring because of a basin characteristic that does not appear in the applicable regression equations for the location. In particular, numerous reports that contain regression equations provide limits to the applicability of the equations based on the percentage of the basin that is either urbanized or regulated. As a result, it is important that users should read the reports for the regression equations that apply to the area of interest, be aware of any such limitations placed on the regression equations, and carefully evaluate the drainage basins for their selected sites to determine if the limitations are violated.

## Limitations for Estimates Obtained by Use of the *Generate Flows Based on Similar Gages* Tool

The *Generate Flows Based on Similar Gages* tool allows users to decide whether or not to allow use of regulated streamgaging stations as a basis for estimating flow statistics for an ungaged site. This decision should be made carefully. If estimates of natural flows are desired, then regulated stations should not be used. If the user-selected site is downstream from a dam, a diversion, or some other activity that has a major effect on streamflow, then estimates based on regulated streamgaging stations may provided more realistic estimates of actual flow conditions at the user-selected site than natural flow estimates.

If regulated stations are used in the analysis, then users should also get a StreamStats report of the published streamflow statistics for the streamgaging station and review the publication from which the statistics were obtained to understand any special conditions used to compute the statistics for the station. In some cases, the computed statistics may have been determined from a period during which the streamflow was unregulated at the station. Directions for obtaining reports of streamflow statistics for USGS data-collection stations are provided below.

By default, StreamStats will provide estimates only if the drainage area for the user-selected site is within 0.5 to 1.5 times the drainage area for an upstream or downstream streamgaging station. These limits can be modified for an individual state application when requested by the local USGS state office.

The *Generate Flows Based on Similar Gages* process usually should be done after estimates based on regression equations are obtained because if regression estimates have been obtained first, then StreamStats will combine the regression estimates with the estimates based on the nearby streamgaging stations to provide improved estimates. Improved estimates are available only when the years of record are available for the streamgaging station in the StreamStats database of streamflow statistics and the equivalent years of record are also available for the regression estimates.

The U.S. Geological Survey expressly disclaims responsibility for damages or liability that may arise from use of these data. Feel free to contact us if you have any questions.