Date of Award
Program or Major
Master of Science
According to the 2014 National Climate Assessment, annual rainfall has increased by approximately 10% from 1895 to 2011, and extreme precipitation events (the top 1% of all storms) have increased 70% since 1958 in the Northeast United States [Melillo et al., 2014]. These precipitation changes could have substantial impacts on the performance of bridges, stormwater drainage, pavement, water supply systems, and numerous other infrastructure projects. This study examines historical changes in Northeast rainfall and develops a computer model to simulate future precipitation. Historical precipitation trends were calculated for rainfall duration, depth, intensity and time between storms for 20 weather stations across the Northeast. In general, annual rainfall depth and intensity have increased over the past 50 years, while the annual time between storms has decreased over the same period. These changes differed among seasons. Winter storms have become shorter and more frequent. Summer storms have become longer and less frequent. Future precipitation at LaGuardia Airport was simulated using a three-step weather generation model. A generalized linear model was used to capture the seasonal pattern of precipitation occurrence. Events were identified as one of four storms types and simulated using a Markov Chain switching model. Precipitation intensity was simulated using a hybrid distribution, consisting of the Gamma and Generalized Pareto distributions, capable of capturing both mean and extreme rainfall events. The model projected annual rainfall increases, with especially large increases in winter precipitation. The frequency of extreme storms, exceeding the 97.5th percentile, was projected to increase substantially. Design storms, the two, 50 and 100 year storm depths, had limited change. This study shows that precipitation timing and intensity has changed in the past, and will likely continue to change in the future. Weather generators provide a means to understand the impacts from changing precipitation patterns beyond single extreme events. Their application can be used to determine the effectiveness and resiliency of existing infrastructure and to identify appropriate adaptation measures.
Goldman, Andrew Louis, "HISTORICAL PRECIPITATION TRENDS AND FUTURE PRECIPITATION PROJECTIONS USING A STOCHASTIC WEATHER GENERATOR IN THE NORTHEASTERN UNITED STATES" (2017). Master's Theses and Capstones. 1148.