Patterns and Periodicities of the Continental U.S. Streamflow Change

Editors

C.S. Pathak, D. Reinhart (Eds.)

Document Type

Conference Proceeding

Publication Date

1-1-2016

Publication Title

World Environmental And Water Resources Congress 2016: Hydraulics and Waterways and Hydro-Climate/Climate Change - Papers from Sessions of the Proceedings of the 2016 World Environmental and Water Resources Congress

Publisher

American Society of Civil Engineers (ASCE)

First page number:

658

Last page number:

667

Abstract

This study evaluated the change patterns of the continental U.S. streamflow across 600 stations having a minimum of 30 years of continuous data. To detect the presence of long term trends, several non-parametric Mann-Kendall tests (with modifications accounting for persistence in data) were used. Non-parametric Pettitt's test was used to evaluate the abrupt shifts and change points across the study period. A subset of the data (237/600 stations) was further analyzed with discrete wavelet transformation (DWT) to evaluate the effect of periodicities on the change patterns using multiple dyadic scales (i.e., two years, four years, and eight years). Water year analyses (along with its dyadic scales) revealed that regions in the northeast and upper-central have experienced an increasing change while regions in the northwest have undergone decreasing change. The seasonal analyses (with dyadic scales) showed wide-ranging changes at different seasons and revealed the seasonal variations. DWT analyses showed that there is a presence of stronger change at a periodicity of 8 years or above at all temporal scales. Coupling effect of multiple climate signals were also observed with the change in streamflow pattern. The study provided comprehensive analyses on the change behavior of U.S. streamflow. Results of the current study can be used in building forecasting models and can also be used for taking water management decisions under changing climate conditions.

Keywords

Mann-Kendall; Pettitt; Step; Streamflow; Trend; Wavelet

Language

English

UNLV article access

Share

COinS