New Projections of Changing Heavy Precipitation in King County


Mauger, G.S., J.S. Won, K. Hegewisch, C. Lynch, R. Lorente Plazas, E. P. Salathé Jr., 2018. New Projections of Changing Heavy Precipitation in King County. Report prepared for the King County Department of Natural Resources. Climate Impacts Group, University of Washington, Seattle.


Executive Summary

Recent research suggests that future heavy rain events will be more intense in the Pacific Northwest (e.g., Warner et al. 2015). Past studies have not accounted for this change, either because the methods cannot reliably capture changes in rainfall intensity (e.g., Abatzoglou and Brown 2012), or because previous studies did not evaluate changes in short-duration precipitation (e.g., hourly) that are of relevance to stormwater planning (e.g., Salathé et al. 2010).

The purpose of this project was to develop projections of 21st century changes in precipitation that can be used to inform stormwater and wastewater management in King County. The work was funded by the King County Department of Natural Resources and Parks and the Washington State Department of Ecology. Additional support came from the Critical Infrastructure Resilience Institute (CIRI), a U.S. Department of Homeland Security S&T Center of Excellence.

The specific objectives of the project were to:

1. Produce two new Regional Climate Model projections: Evaluate global models and develop projections representing a low- and a high-end scenario for 21st century change in precipitation. Archive hourly precipitation and other fields (e.g., evapotranspiration, wind speed and direction), for the entire model domain.

2. Synthesize projections to support wastewater conveyance and treatment system impacts assessment: Evaluate changes in the intensity, duration, and magnitude of heavy precipitation events in the region. Explore the possibility of developing a statistically-derived pseudo-ensemble of projected changes by relating large-scale global model projections to local-scale exceedance probabilities.

3. Synthesize projections to support countywide stormwater design: Develop hourly precipitation time series (1970-2099) for hydrologic model input points.

Using the Weather Research and Forecasting (WRF, Skamarock et al. 2005) regional climate model, we produced two new “dynamically downscaled” projections of future climate. A key feature of these projections is that they provide hourly estimates of future weather conditions (temperature, precipitation, humidity, wind, etc.). This is critical, given that many stormwater facilities are designed based on short-duration rainfall intensities. Results from each simulation were evaluated for 83 rain gauge sites that are currently operated by Seattle and King County for stormwater planning.

Our results show the potential for large increases in future rainfall intensity, for example showing a 7 to 54% increase in the 10-year hourly rainfall extreme at Sea-Tac, by the 2080s. However, results differ substantially among seasons and for the two climate projections considered. Although most projections indicate an increase in precipitation intensity, results for the low-end model project a decrease in precipitation intensity for some statistics and durations (e.g., the 25-, 50-, and 100-year extremes in hourly precipitation at Sea-Tac). Similarly, most of the projected changes for summer suggest a decrease in precipitation intensity, although this may be affected by the model’s limited ability to capture convective events such as thunderstorms. All data, documentation, and findings are made available online, including an interactive tool that allows users to easily navigate to the station and results of interest.

There are two ways in which future work could build on these results. First, by producing additional regional climate model projections. Additional simulations could be obtained via a collaboration with UW Professor Cliff Mass, who is currently producing several new regional model projections under a grant from the Amazon Catalyst program. These new projections would help elucidate our results in two ways: by both increasing the sample size, thereby improving the statistics; and by evaluating results for new global models, which may have different model representations for key processes. Second, we believe there are a number of methodological choices that could be investigated further, including (a) an improved treatment of the precipitation statistics and (b) further optimization of the regional climate model to improve the representation of precipitation extremes.