Regional climate model evaluations of long-term changes in total precipitation and high precipitation events

atmospheric science;regional climate model;high precipitation events;total precipitation;trends in precipitation;long term variations

Jia, Wenjing

关键词: atmospheric science;    regional climate model;    high precipitation events;    total precipitation;    trends in precipitation;    long term variations;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/42479/Wenjing_Jia.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Extreme events, including precipitation extremes, can have severe impacts on human society and on ecosystems. In the late 20th century, heavy precipitation events tend to occur with increasing frequency or intensity proportion to total rainfall over most land areas in the world. Precipitation, however, is the single most difficult variable to simulate in numerical weather/climate models. There are very limited studies on modeling evaluation of long-term trends in extreme precipitation events over the contiguous U.S.This study examines the downscaling skills of a state-of-the-art regional climate model, the Climate extension of Weather Research and Forecasting Model (CWRF), in evaluating long- term changes (1982-2008) in total precipitation and high precipitation events (e.g., 75th, 85th, 95th percentiles) over the contiguous U.S. The CWRF, driven by the NCEP-DOE AMIP II Reanalysis (R-2), was simulated for the period 1982-2008 over the contiguous United States. All the 27-yr interannual and interseasonal (half years of warm season: March-August; cold season: September-February) total and high precipitation events were calculated from the observational data and model outputs. The observed and simulated trends for the total and high precipitation events were calculated by using Kendall’s tau based slope estimator (Theil-Sen regression), a powerful alternative to the simple least squares linear regression slope. The Mann-Kendall Rank- based nonparametric significance test was applied to evaluate the statistical significance of the trends in the total and high precipitation events. The CWRF not only captured the magnitudes of total and high precipitation evens overall in the contiguous U.S., but simulated their trends well relative to those derived from observations, even in higher-elevated subregions (e.g., the West Coast and some areas of Rocky Mountains), where it is difficult for a model to properly determine precipitation. In general the trends in total and high precipitation events simulated by he model were not as intense as the observed ones. Better agreements between model and observations in both total and high precipitation events and their trends were found in the warm half years (warm seasons: Mar.-Aug.) than in the cold half years (cold seasons: Sep.-Feb.) during 1982-2008. Thus, CWRF shows significant improvements in warm-season convective precipitation relative to many other numerical weather/climate models. It is also noticed that trends in the high precipitation events are highly sensitive to the spatial scale chosen (at least for a case study of the lower-elevated regions from the Midwest (MW) to Illinois (IL), and to central Illinois (CIL) in the contiguous U.S.) from both model and observations.This is the first study to evaluate a regional climate model’s capability in capturing long- term historical trends in precipitation extremes (or high precipitation events) over the contiguous United States. Previously only Kunkel et al. (2002) evaluated a 10-yr (1979-88) heavy precipitation events over the U.S. using the second-generation Regional Climate Model (RegCM2), but did not examine the trends in the heavy precipitation. The findings have important implications on the capabilities of projections of total and extreme precipitation events in the future climate. Our study also provides important information for the capabilities of using this model to assess long-term variations of climatic and hydrologic extremes and hazards (e.g. floods, droughts, etc.) and to support mitigation strategies.