Regional Model Projections of Heavy Precipitation for Use in Stormwater Planning

Project Overview

Dynamically downscaled projections – in which Regional Climate Models (RCMs) are used to simulate local scale changes in climate – are still relatively rare, especially at the scales needed to assess changes in stormwater risk. Yet recent work has shown that RCMs are needed to accurately capture changes in heavy rainfall events in the Pacific Northwest (e.g., Salathé et al. 2014). The purpose of this work was to produce, obtain, and distill new RCM simulations for use in stormwater and CSO planning. This page combines results from a series of projects, in which tailored projections were produced for each participating jurisdiction.


RCM simulations were performed using the Weather Research and Forecasting (WRF, Skamarock et al. 2005) community mesoscale model. Building on extensive UW experience with this model, WRF was configured using the same approach used in previous work (Salathé et al. 2010, 2014). WRF simulations extended from 1970-2099. In order to facilitate stormwater analyses, results were saved at an hourly time step.

Global models are needed to drive the RCM simulations. All global climate model scenarios were obtained from the Climate Model Intercomparison Project, phase 5 (CMIP5; Taylor et al., 2012). All but one projection are based on a high-end greenhouse gas scenario (Representative Concentration Pathway (RCP) 8.5; van Vuuren et al., 2011).

WRF projections were bias-corrected to match rain gauge observations for a wide range of federally and locally-operated weather stations. Results were used to evaluate the change in the frequency, intensity, and duration of precipitation events. Changes were summarized as a function of duration (1 to 360 hours) and exceedance probability, for three future time periods: the 2030s, 2050s, and 2080s, relative to the recent past (e.g.: 1980-2009). Results are made available for download, are summarized in an interactive web tool and described in the reports linked below.


Results are designed to support both municipal and rural stormwater planning. The online tool allows users to evaluate projections as a function of precipitation intensity, duration, and frequency. The online repository includes summary statistics as well as time-series data for use in continuous simulation modeling.

 Project Outputs

  • Online repository of results for each rain gauge.
  • Google map for locating rain gauges and accessing online data.
  • Interactive web tool
  • Project Reports:
    • City of Everett 
      • Errata: An error was found in the GFDL CM3, RCP 8.5 projection. The error has been corrected in this link version, but the the previous results should be disgarded and replaced. Table 3 has been updated to reflect the corrected GFDL CM3 projections. Updated 8/20/2019.
    • King County
      • Phase 1 Report
        Errata: The Phase 1 King County report (Mauger et al., 2018) described results from two WRF projections: (1) ACCESS 1.0, RCP 4.5, and (2) GFDL-CM3, RCP 8.5. In creating the new larger ensemble of 12 RCP 8.5 projections (see Phase 2 King County report), an error was found in the WRF boundary conditions used for the GFDL-CM3 simulation. The error has been corrected in the new ensemble, and all of the reports have been updated to reflect this change. Previous results, based on the older GFDL CM3 RCP 8.5 WRF projection, should be discarded. Updated 8/20/2019)
      • Phase 2 Technical Memo
    • Port Gamble S’Klallam Tribe
      • Errata: 1) Correction was made to text on page 19. Previous text made incorrect references to tables and results (see report page ii for specifics). Updated October 30th, 2018. 2) Correction to results from high-end WRF projection: An error was found in the GFDL CM3 (RCP 8.5) WRF simulation. This has been corrected, and Table 5 has been updated accordingly. The data have also been replaced in the online repository. Updated 8/20/2019
    • Thurston County
  • Peer-reviewed publications: Lorente-Plazas et al. (2018)

Key Personnel

  • Guillaume Mauger, (Principal Investigator), University of Washington*
  • Jason Won, University of Washington*


The primary source of funding for this work was the King County Department of Natural Resources and Parks. Additional funding was provided by the Washington State Department of Ecology, City of Everett, Thurston County, and the Port Gamble S’Klallam tribe.

Back to Top