Date of Award

10-29-2009

Document Type

Restricted Access Thesis

Degree Name

Master of Science in Applied Meteorology

Department

Department of Atmospheric Sciences and Chemistry

Thesis Advisor

Sam Miller

Committee Member

Barry Keim

Committee Member

Dan St. Jean

Abstract

A classification of climate zones in New England was developed using multivariate analysis of meteorological data collected from surface weather stations. This was part of a larger project to define winter severity indexes for New England that will help determine the cost-benefit of new winter season technologies, crew training, and the efficiency of winter road maintenance practices. Data were gathered from ASOS and COOP stations throughout New England during the winter months from 2001-2006. The data consisted of precipitation type, temperature, and wind. The mean and maximum of each variable were computed over each month. The results were interpolated onto a 25km grid using Barnes Analysis. Principle component analysis (PCA) was performed on the gridded data to reduce the dimensionality of the dataset. Three components were retained, and the associated component scores were incorporated in cluster analysis. Nine distinct clusters were identified, and grid points were assigned a cluster membership. This resulted in nine climate zones as each cluster represents a zone. The exception was cluster 5 which was discontinuous. An additional climate zone was created as a result to keep continuity among the other zones. Stepwise discriminant analysis was performed to determine the successfulness of the classification of grid points to clusters. ArcGIS was used to graphically represent the climate zones and their respective boundaries. Finally, boundaries were delineated based on grid point cluster assignments.

Share

COinS