Online Program Home
My Program

Abstract Details

Activity Number: 403 - SPAAC Poster Competition
Type: Topic Contributed
Date/Time: Tuesday, July 30, 2019 : 2:00 PM to 3:50 PM
Sponsor: Section on Statistics and the Environment
Abstract #304478
Title: Testing Exchangeability in Spatiotemporal Random Processes
Author(s): Trevor Harris* and Bo Li and Nathan Steiger and Jason Smerdon and Naveen Naidu Narisetty and Derek Tucker
Companies: University of Illinois Urbana Champaign and University of Illinois at Urbana-Champaign and Lamont-Doherty Earth Observatory and Lamont-Doherty Earth Observatory and University of Illinois at Urbana Champaign and Sandia National Laboratories
Keywords: paleoclimate; spatiotemporal; data depth; functional data; inference; central regions

Statistical inference on spatiotemporal processes is a fundamental problem in many fields including Ecology, Oceanography, and Climatology. Of particular interest to the paleoclimate community is the study of Climate Field Reconstructions (CFRs) with seasonal to annual resolution spanning the last several millennia. CFRs attempt to recover spatiotemporal fields of climate variables, using proxy records of past climate variability, and have emerged as important tools for studying the mechanisms of climate change. Motivated by assessing differences between CFRs, we propose a new method for evaluating the differences in the distributions of two spatiotemporal processes by using the notions of data depth and functional data. Our test is robust, computationally efficient, distribution free and and has a convenient asymptotic distribution. We apply our test to study global and regional proxy influence on a Data Assimilation based CFR by comparing its background and analysis states. We find that there is a steadily increasing divergence between the state’s distributions over time, indicating increasing proxy influence, and that proxy influence can extend far beyond collection sites.

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

Back to the full JSM 2019 program