Award Date

5-1-2019

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Civil and Environmental Engineering and Construction

First Committee Member

Sajjad Ahmad

Second Committee Member

David James

Third Committee Member

Daniel Gerrity

Fourth Committee Member

Haroon Stephen

Fifth Committee Member

Ajay Kalra

Sixth Committee Member

Ashok Singh

Number of Pages

221

Abstract

The primary objective of the work presented in this dissertation was to evaluate the change patterns, i.e., a gradual change known as the trend, and an abrupt change known as the shift, of multiple hydro-climatological variables, namely, streamflow, snow water equivalent (SWE), temperature, precipitation, and potential evapotranspiration (PET), in association with the large-scale oceanic-atmospheric climate signals. Moreover, both observed datasets and modeled simulations were used to evaluate such change patterns to assess the efficacy of the modeled datasets in emulating the observed trends and shifts under the influence of uncertainties and inconsistencies. A secondary objective of this study was to utilize the detected change patterns in designing data-driven prediction models, e.g., artificial neural networks (ANNs), support vector machines (SVMs), and Gaussian process regression (GPR) models, coupled with data pre-processing techniques, e.g., principal component analysis (PCA) and wavelet transforms (WTs). The study was not solely limited to the hydrologic regions of the conterminous United States (U.S.); rather it was extended to include an analysis of northern India to appraise the differences in the spatiotemporal variation on a broader scale.

A task was designed to investigate the significant spatiotemporal variations in continental US streamflow patterns as a response to large-scale climate signals across multiple spectral bands (SBs). Using non-parametric (long-term) trend and (abrupt) shift detection tests, coupled with discrete wavelet transform, 237 unimpaired streamflow stations were analyzed over a study period of 62 years (1951 to 2012), looking at the water year and seasonal data, along with three discrete SBs of two, four, and eight years. Wavelet coherence analysis, derived from continuous wavelet transform, determined the association between the regional streamflow patterns and three large-scale climate signals, i.e., El Niño Southern Oscillation (ENSO), Pacific Decadal Oscillation (PDO), and Atlantic Multi-decadal Oscillation (AMO), across continuous SBs ranging from two to 16 years. The results indicated significant positive (negative) trends and shifts in the northeastern and north-central (northwestern) regions with an increase in the number of stations as the SB durations increased. The spatiotemporal association between regional streamflow and climate signals varied significantly (from no correlation, Rn2 ~ 0, to perfect correlation, Rn2 ~ 1.0) even amongst adjacent regions. Among the climate signals, ENSO showed the highest association (Rn2 ~ 1.0), having a consistent phase relationship with regional streamflow patterns, especially in the higher SBs. PDO (with the least influence among the three signals) and AMO showed stronger associations, in the lower SBs. These results may help explain the teleconnections between the climate signals and the US streamflow variations across multiple SBs, which may lead to improved regional flow regulations. The comparison among several data-driven models, e.g., ANN, SVM, and GPR models, preceded by PCA and WT, produced comparable results with significant accuracy (with R2 above 0.90) in short-term prediction of streamflow.

Later, the correlations between the western U.S. snow water equivalent (SWE) and the two major oceanic-atmospheric indices originating from the Pacific Ocean, namely, ENSO and PDO, were evaluated using continuous wavelet transform and its derivatives. Snow Telemetry (SNOTEL) data for 1 April SWE from 323 sites (out of which 258 are in six hydrologic regions) were obtained for a study period of 56 years (1961–2016). The results showed that ENSO had a much higher influence than PDO throughout the western U.S. SWE across the study period. Both ENSO and PDO showed a higher correlation with SWE at multiple timescale bands across different time intervals, although significant intervals in the higher timescales were of longer duration. ENSO showed a higher correlation in the 10-to-16-year band across the entire study period as well as in the lower timescales. PDO showed a higher correlation below the 4-year band. The relative phase relationship suggested that ENSO led SWE, with certain lags, while both were moving in the same direction in many instances. The lag-response behavior of SWE and PDO was not found to be uniform. Regional analyses, based on the western U.S. hydrologic regions, suggested significant variation across adjacent regions in terms of their correlation with ENSO/PDO. Association with ENSO was also observed to be higher compared to PDO among the regions. Regions close to the ocean and at lower elevation showed higher correlation compared to the inland regions with higher elevation.

The influence of ENSO on the north Indian temperature, precipitation, and PET change patterns was evaluated during the monsoon season across the last century. Trends and shifts in 146 districts were assessed using non-parametric statistical tests. To quantify their temporal variation, the concept of apportionment entropy was applied to both the annual and seasonal scales. Results suggest that the El Niño years played a greater role in causing hydro-climatological changes compared to the La Niña or neutral years. El Niño was more influential in causing shifts compared to trends. For certain districts, a phase change in ENSO reversed the trend/shift direction. The all-year (century-wide) analysis suggested that the vast majority of the districts experienced significant decreasing trends/shifts in temperature and PET. However, precipitation experienced both increasing and decreasing trends/shifts based on the location of the districts. Entropy results suggested a lower apportionment of precipitation compared to the other variables, indicating an intermittent deviation of precipitation pattern from the generic trend. The findings may help understand the effects of ENSO on hydro-climatological variables during the monsoon season. Practitioners may find the results useful, as monsoon, among the Indian seasons, experience the largest climate extremes.

A final task was designed that evaluated Coupled Model Intercomparison Project 5 (CMIP5) simulation models’ ability to capture the observed trends under the influence of shifts and persistence in their data distributions. A total of 41 temperature and 25 precipitation CMIP5 simulation models across 22 grid cells (2.5° x 2.5° squares) within the Colorado River Basin were analyzed and compared against the Climate Research Unit Time Series (CRU-TS) observed datasets over a study period of 104 years (from 1901 to 2004). Both the model simulations and observations were tested for shifts, and the time series before and after the shifts were analyzed separately for trend detection and quantification. Effects of several types of persistence were accounted for prior to both the trend and shift detection tests. The mean significant shift points (SPs) of the CMIP5 temperature models across the grid cells were found to be within a narrower range (between 1960 and 1970) compared to the CRU-TS observed SPs (between 1930 and 1980). Precipitation time series, especially the CRU-TS dataset, had a lack of significant SPs, which led to an inconsistency between the models and observations since the numbers of grid cells with a significant SP were not comparable. The modeled CMIP5 temperature trends, under the influence of shifts and persistence, were able to match the observed trends quite satisfactorily (within the same order and consistent direction).

Unlike the temperature models, the CMIP5 precipitation models detected the SPs earlier than the observed SPs found in the CRU-TS data. The direction (as well as the magnitude) of trends, before and after significant shifts, were found to be inconsistent between the modeled simulations and observed precipitation data. Shifts, based on their direction, were found to either strengthen or neutralize pre-existing trends both in the model simulations as well as in the observations.

The results also suggest that the temperature and precipitation data distributions were sensitive to different types of persistence. Such sensitivity was found to be consistent between the modeled and observed datasets. The study detected certain biases in the CMIP5 models in detecting the SPs (a tendency of detecting shifts earlier or later than the observed shifts) and also in quantifying the trends (overestimating the trend slopes). Such insights may be helpful in evaluating the efficacy of the simulation models in capturing observed trends under uncertainties and natural variabilities.

Keywords

Artificial Neural Network (ANN); Support Vector Machine (SVM); Gaussian Process Regression (GPR); Wavelet-Artificial Neural Network (WANN); Discrete wavelet transform; Continuous wavelet transform; Wavelet coherency; Cross wavelet transform; ENSO; PDO; AMO; Multi-resolution spatio-temporal change analyses; Trend; Shift; Entropy; Persistence; Autocorrelation; US streamflow; Snow water equivalent; Temperature; Precipitation; Potential evapotranspiration; Indian monsoon; Colorado River Basin; CMIP5

Disciplines

Artificial Intelligence and Robotics | Civil Engineering | Computer Engineering | Hydrology

Language

English


Share

COinS