|dc.description.abstract||Climate change impacts on precipitation characteristics will alter the hydrologic characteristics, such as peak flows, time to peak, and soil erosion potential. However, many of the currently available climate change datasets are provided at temporal and spatial resolutions that are inadequate to quantify projected changes in the hydrologic characteristics of a watershed. Therefore, it is critical to temporally disaggregate coarse-resolution precipitation data to finer resolutions for studies sensitive to precipitation characteristics. First, this study generated novel 15-minute precipitation datasets from hourly precipitation datasets obtained from five NA-CORDEX downscaled climate models under Representative Concentration Pathway (RCP)8.5 scenario for the historical (1970-1999) and projected (2030-2059) years over the Southeast United States using a modified version of a stochastic method. The results showed the conservation of mass of the precipitation inputs. Furthermore, the probability of zero precipitation, variance of precipitation, and maximum precipitation in the disaggregated data matched well with the observed precipitation characteristics. The generated 15-minute precipitation data can be used in all scientific studies that require precipitation data at that resolution.
Secondly, the datasets generated from objective one were further applied to estimate projected erosivity (R-factor), which determines the erosive power of rainfall. The magnitude and scope of these future changes in the erosive power of rainfall remain largely unknown, particularly at finer resolutions and local scales because previous studies have relied on aggregated (hourly, daily) rainfall data. The erosivity for the region was calculated using the RUSLE2 erosivity calculation method without data limitation. Ensemble results for projected values (as compared to historical values) showed a projected increase in annual average precipitation, erosivity, and erosivity density by 14%, 47%, and 29%, respectively, over the southeast region from 2030 to 2059. The future ensemble model showed an average annual R-factor of 11237±1299 MJ mm ha-1 h-1 yr-1. These findings suggest that changes in rainfall intensity, rather than only precipitation amount, may drive future changes in erosivity. However, the study is associated with limitations due to bias correction and downscaling of the precipitation dataset and obscuring the result of projected erosivity. In general, coastal and mountainous regions are expected to experience the greatest absolute increase in erosivity, while other inland areas are expected to experience the greatest relative change. This study offers a novel examination of projected future precipitation characteristics in terms of erosivity and potential future erosion.
Furthermore, future projected rainfall Intensity-Duration-Frequency (IDF) curves were developed for the Southeast United States using Generalized extreme value (GEV) distribution with disaggregated sub-hourly (15-, 30-, and 45-min) monthly maximum rainfall from 2030 to 2059 using the five climate models under the RCP8.5 scenario. A computationally efficient feed-forward back-propagation Artificial Neural Network (ANN)-based approach was found to be significantly superior for disaggregating rainfall compared to the stochastic model with an average Nash–Sutcliffe efficiency (NSE) ranging from 0.67 to 0.84. Kolmogorov-Smirnov (KS) test confirmed at a 5% significance level that the annual maximum rainfalls come from Gumbel extreme value distributions. The study found that there is an increasing intensity of future projected rainfall in the range of 7% to 36% in comparison to the historical period. The spatial variation in future projected extreme rainfall depths showed that the Gulf-Atlantic coast and the Appalachian Mountains are expected to receive more extreme rainfalls.
Finally, uncertainties associated with climate studies are of considerable interest, particularly extreme rainfall analysis for providing confidence to the relevant stakeholders for designing hydrological structures. Therefore, this study quantified the uncertainties associated with the rainfall IDF Curves for the largest cities of the Southeast United States using the ANN and bootstrapping resampling technique along with Gumbel distribution. Results showed no significant differences while disaggregating hourly to sub-hourly (15-, 30-, and 45-min) monthly maximum rainfall with or without hyperparameter tuning using random search. Additionally, bias correction significantly improved the rainfall IDF curves rejecting the null hypothesis of no difference using Welch two sample t-test. Overall, the minimum and maximum of annual maximum rainfall intensities were found in a range of 38 to 55 mm h-1 and 143 to 210 mm h-1, respectively. Further studies are needed to improve the uncertainty quantification of human error in climate modeling and the randomness of natural processes in datasets.||en_US