Periodic Time Series Models for United States Extreme Temperature Trends

Jaechoul Lee

Boise State University
 

Abstract

This paper studies trends in monthly maximum and minimum
temperatures recorded over the past 150 years in the
contiguous United States. Regression methods based on
periodic time series settings for observed extreme
temperatures are used to quantify linear trends in these
temperatures from 528 stations in the United States.
A brief overview of periodic time series will be given at
the beginning of the talk with a reason for needs of
periodic time series methods to temperature series.
Issues of temporal autocorrelations, periodicities, station
relocations (change points), and spatial smoothing algorithms
arise. The seasonal aspect of the analysis allows us to
investigate the issue of uniformity of temperature change
over varying season. Spatial contour maps of the estimated
trends will be presented for each of the four seasons and the
entire year.

Joint work with Robert Lund -- The University of Georgia