Online Program Home
My Program

Abstract Details

Activity Number: 649 - Recent Advances in Spatial and Spatial-Temporal Methods
Type: Topic Contributed
Date/Time: Thursday, August 1, 2019 : 10:30 AM to 12:20 PM
Sponsor: Section on Bayesian Statistical Science
Abstract #304573
Title: Spatio-Temporal Model to Predict Extreme Heat Events at Unobserved Locations
Author(s): Erin Schliep* and Alan E Gelfand
Companies: University of Missouri and Duke University

We develop a spatially-dependent multi-state autoregressive model to investigate the incidence, duration, and exceedance behavior of extreme heat events over a fixed spatial region. The data consist of maximum daily temperatures spanning multiple years and spatial locations. An extreme heat event is classified as one or more days with a maximum temperature above a pre-specified threshold that may vary spatially. We propose a two state model, where one state denotes the below threshold state and the second denotes the above threshold state; the latter defines the extreme heat events. The transition probabilities between states depend on both the previous state and previous observed temperature and are also spatially dependent. With data spanning more than 50 years at 18 locations, we investigate the changes in incidence and duration of extreme heat events over space and time. In addition, our model is able to predict, with uncertainty, the incidence and duration of extreme heat events at unobserved locations.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program