Date of Award

Fall 2019

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Civil/Environmental Engineering

Program/Concentration

Civil Engineering

Committee Director

Xixi Wang

Committee Member

Isao Ishibashi

Committee Member

Mujde Erten-Unal

Abstract

This thesis developed a statistical downscaling approach, which consists of a series of linear regression equations, to spatiotemporally downscale the rainfall predictions from North American Regional Climate Change Assessment Program (NARCCAP) in accordance with the 15-min observed rainfall data at the rain gauges across the state of Virginia. NARCCAP has generated twelve different region-global climate models (RCM-GCMs) with a temporal resolution of 3 hr and a spatial resolution of 50 km over the entire America. Although it has been downscaled already, such resolutions are still too coarse to represent the rain gauges. This means that the RCM-GCMs’ predictions need to be further downscaled for watershed planning and management as well as hydrologic engineering design. In this regard, this thesis developed the statistical downscaling approach using the predictions by one of the RCM-GCMs and then validated the applicability of the approach using the predictions by the other RCM-GCMs. The development and validation were implemented by comparing the RCM-GCMs’ predictions with the observed data. Future studies should better utilize the predictions of all twelve RCM-GCMs and try some nonlinear algorithms to minimize either underestimating or over predicting some extreme rainfalls for a duration of longer than 3 hr.

DOI

10.25777/87kt-yq93

Share

COinS