Research indicates a warming global climate leads to change in the spatial and temporal characteristics of precipitation. Although precipitation is inherently variable through time and space, for some water-sensitive stakeholders, the evenness with which precipitation is distributed through a time interval rivals the importance of total precipitation amount and frequency within that period. This study uses a relatively new approach of analyzing inequity in the temporal distribution of precipitation to examine the recent historical record of precipitation across the United States. The Gini coefficient (GC), which has been commonly used in the field of economics to measure wealth distribution, was used here to assess inequity in the temporal distribution of daily precipitation through seasonal and annual timeframes. Additionally, the Lorenz asymmetry coefficient (LAC) was used to assess the magnitude of daily precipitation events (light, heavy) associated with inequity in the temporal distribution of precipitation. The concept of using these two metrics together to quantify changes in the character with which precipitation occurs across a time interval has yet to be documented for areas within the United States. Therefore, this study expands upon previous research of long-term hydroclimatic change and variability by illustrating the combined ability of these two relatively under-utilized metrics, the GC and the LAC, to enhance quantification of recent change in the characteristics of the temporal distribution of daily precipitation across the United States.
The first element of the research presented here is demonstration of the utility of the GC and LAC metrics using data from the physically diverse mid-Atlantic sub-region of the United States. This research used station-level daily precipitation data to compute historic time series of intra-annual and intra-seasonal precipitation amount, precipitation frequency, GC, LAC, variance (V), and interquartile range (IQR). The results of this portion of the research show that when compared to other simpler measures of characterizing variability (i.e., V and IQR), the GC is relatively robust to both the number of days with precipitation and the total precipitation received in a temporal increment (i.e., season or year). The research expanded in scale to the continental United States, requiring data integration to a regional level to facilitate data analysis and physical understanding. The analysis used gridded seasonal means (1981 – 2010) of four precipitation characteristics: precipitation amount, precipitation frequency, GC, and LAC to delineate regions of homogenous precipitation characteristics. To accomplish this, a multi-step regionalization technique was employed. Specifically, the historic seasonal means were subjected to a Principal Components Analysis (PCA), and the resulting component scores were subjected to several cluster analysis techniques. The average linkage clustering technique produced the most logical clustering solution, indicating that 15 regions of homogenous precipitation exist within the contiguous United States. It is argued that the regions better serve hydroclimatic analyses than the nine climate regions designated by the United States National Centers for Environmental Information (NCEI).
The third element of the research integrates the first two research elements in study of recent United States hydroclimate variability and change. For the 15 United States hydroclimate regions, regionally averaged water year time series (1949 – 2018) of precipitation amount, precipitation frequency, GC, and LAC were computed using in-situ precipitation data gathered from the NCEI's Global Historical Climatology Network (GHCN)-Daily database. The time series of all precipitation characteristics for each region were then subjected to the nonparametric Mann-Kendall trend test to assess the significance of each trend, and the Sen's slope estimator was used to quantify the magnitude of the trend. Time series that characterize two key atmospheric characteristics, total column water vapor and static stability, were also computed for each region. For most of the 15 study regions, water year total precipitation and precipitation frequency increased through the latter half of the 20th century. The largest magnitude of change in water year total precipitation and precipitation frequency occurred in the time series of regions located within the eastern and northern portions of the contiguous United States. Results also show that inequity in the temporal distribution of water year precipitation increased through the 70-year study period for most of the 15 study regions. Combined, these results indicate that days with light and heavy precipitation are becoming more prevalent at the expense of days with moderate precipitation. Furthermore, variability in the time series of some precipitation characteristics for several regions coincide with variability in the atmospheric variables that characterize total column water vapor and static stability, however the dominant driver of hydroclimatic change across the contiguous United States remains elusive. / Doctor of Philosophy / Research indicates a warming global climate leads to change in the spatial and temporal characteristics of precipitation. These changes could adversely affect some water-sensitive stakeholders who are concerned not only with the amount of precipitation received over time, but also with the manner in which the precipitation is distributed through time – all at once, or spread evenly. The Gini coefficient (GC), which has been commonly used in the field of economics to measure wealth distribution, was used here to assess inequity in the temporal distribution of daily precipitation through seasonal and annual timeframes. Additionally, the Lorenz asymmetry coefficient (LAC) was used to assess the magnitude of daily precipitation events (light, heavy) that were primarily responsible for inequity in the distribution of daily precipitation amounts through each time interval. The research presented here used gridded seasonal means (1981 – 2010) of four precipitation characteristics: precipitation amount, precipitation frequency, GC, and LAC to delineate regions of homogenous precipitation characteristics. Through this process, 15 hydroclimatic regions were delineated within the contiguous United States. Regionally averaged annual time series (1949 – 2018) of precipitation amount, precipitation frequency, GC, and LAC were computed for each region using station-level precipitation. The time series of each precipitation characteristic, and for each region were then examined for statistical trends through the 70-year study period. Regional time series which characterize two key atmospheric characteristics, total column water vapor and static stability, were also computed for each region. For most of the 15 study regions, water year total precipitation and precipitation frequency increased through the latter half of the 20th century. The largest magnitude of change in water year total precipitation and precipitation frequency occurred in the time series of regions located within the eastern and northern portions of the contiguous United States. Results also show that precipitation became less evenly distributed across the water year through the 70-year study period for most of the 15 study regions. Combined, these results indicate that days with light and heavy precipitation are becoming more prevalent at the expense of days with moderate precipitation. Furthermore, variability in the time series of some precipitation characteristics for several regions coincide with variability in the atmospheric variables that characterize total column water vapor and static stability, however the dominant driver of hydroclimatic change across the contiguous United States remains elusive.
| Identifer | oai:union.ndltd.org:VTETD/oai:vtechworks.lib.vt.edu:10919/106952 |
| Date | 19 June 2020 |
| Creators | Marston, Michael Lee |
| Contributors | Geography, Ellis, Andrew, Shortridge, Julie Elizabeth, Robinson, David A., Shao, Yang |
| Publisher | Virginia Tech |
| Source Sets | Virginia Tech Theses and Dissertation |
| Detected Language | English |
| Type | Dissertation |
| Format | ETD, application/pdf |
| Coverage | United States |
| Rights | In Copyright, http://rightsstatements.org/vocab/InC/1.0/ |
Page generated in 0.0033 seconds