Date of Completion

9-6-2018

Embargo Period

9-5-2023

Keywords

Weather forecasting, numerical weather prediction, storm wind-speed prediction, statistical post-processing technique, ensemble prediction

Major Advisor

Marina Astitha

Associate Advisor

Emmanouil N. Anagnostou

Associate Advisor

Luca Delle Monache

Associate Advisor

Brian M. Hartman

Associate Advisor

Guiling Wang

Field of Study

Environmental Engineering

Degree

Doctor of Philosophy

Open Access

Open Access

Abstract

Weather forecasting of extreme weather events has evolved with the use of various techniques in an attempt to improve the forecast skill (e.g. multi/single model ensemble, data assimilation and high-resolution techniques). Despite the noted improvements, inaccuracies caused by random and systematic errors are a continuous topic for research. Statistical post-processing techniques that are computationally inexpensive play a useful role to address this issue and contribute to the reduction of prediction errors. In this research, statistical post-processing methods using numerical weather prediction models and single-model ensemble techniques are implemented to improve the prediction of storms (e.g. tropical storm, thunderstorm, rain/wind events) that have impacted the northeast United States. The research is focused on the improvement of prediction accuracy of storm wind speed which is the main driving force behind tree failure and interruptions in the power grid in NE U.S. that can last from hours to days depending on the storm severity. The methods developed and described in this dissertation are applicable to any geographic location and not constrained for NE U.S., as long as there is a long record of historical storms in both observations and model simulations.

Available for download on Tuesday, September 05, 2023

COinS