Award Date

12-1-2016

Degree Type

Thesis

Degree Name

Master of Science in Engineering (MSE)

Department

Civil and Environmental Engineering and Construction

First Committee Member

Sajjad Ahmad

Second Committee Member

Sajjad Ahmad

Third Committee Member

David James

Fourth Committee Member

Haroon Stephen

Fifth Committee Member

Ajay Kalra

Sixth Committee Member

Ashok Singh

Number of Pages

108

Abstract

This thesis presents a comprehensive statistical analysis that determines the direction, rate, and interval of significant streamflow change patterns in the continental U.S. and correlates these changes with sea-surface temperature fluctuations. First, by using two non-parametric tests, namely, the Mann-Kendall trend test and the Pettitt’s test, the presence of long-term trends and abrupt shifts were determined at 10% significance level over continuously adjustable periods that stretched from 1903 to 2012. Modified versions of the tests were applied to account for the presence of persistence (autocorrelation) in data. Theil-Sen slope was determined to evaluate the rate of change across multiple temporal scales (water year and the four seasons). Intervals with a significant change in the historical data were identified utilizing the variable record length of data. Second, correlation and relative phase relationship between western U.S. streamflow and ENSO/PDO across multiple frequency (time-scale) bands were determined (from 1951 to 2010) using the concept of cross wavelet transformation (complex conjugation of independent continuous wavelet transformation) and wavelet coherency analysis. Significant regions of correlation in the time-frequency spectrum were determined with an error of 5% against the red noise (signal noise produced by Brownian motion).

The results revealed that the northeast and the upper-mid regions of the continental U.S. have experienced increasing trend over the study period, while the southwest and the northwest regions have undergone significant decrease. Shifts had similar spatial change patterns as trends with the higher number of stations with significance. Trends were observed across different intervals from 1910 to 2012, while shifts were prominent from 1961 to 2000. The highest variance in terms of rate of change was observed during spring, whereas summer showed the least variation. The cross wavelet transformation and wavelet coherency analysis between western U.S. streamflow and ENSO/PDO revealed that ENSO has a much higher association with the streamflow variance compared to PDO. Both ENSO and PDO showed higher correlation from 1980 to 2005, though their respective frequency bands of significance were different. ENSO had a higher correlation in the 10-12 years, while PDO showed a higher correlation in the 8-10 years band and bands beyond 16 years. The relative phase relationship revealed that both ENSO and PDO led streamflow with a certain lag.

The contributions of this study include a better understanding of the streamflow change patterns across different hydrologic regions of the continental U.S. and quantifying the effect of temporal scales and frequency parameters of change in association with sea-surface temperature fluctuations of the Pacific Ocean. The results may be helpful to regional water managers and the relationships determined can be useful predictors of climate forecasting models.

Keywords

Continuous Wavelet Transformation; Cross wavelet and wavelet coherency; ENSO; Mann-Kendall and Pettitt's Test; PDO; US streamflow

Disciplines

Civil Engineering | Climate | Hydrology

Language

English


Share

COinS