Paleoclimatic Reconstruction of Annual Precipitation in the East, North, and Central Division of Midwest Region: A Bayesian Hierarchical Model for Sparse Spatial Data

• Activity Number: 290
• Type: Topic Contributed
• Date/Time: Tuesday, August 3, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics and the Environment
• Author(s): Rajib Paul*+ and Kwok Wai Lau and William R. VanHove
• Companies: Western Michigan University and CSIRO and Western Michigan University
• Keywords: dendroclimatology ; dimension reduction ; nonseparable correlation

Abstract

Recent researches based on global climate models project an upward trend for precipitation in Midwest Region. We reconstruct the annual rainfall of this region for the past millennium using tree-rings and lake-levels proxy data. Due to proximity, Great Lakes influence the Midwest climate by acting as a big source of heat sink. Despite heavy rainfall, due to increased temperature and high evaporation rates, Great Lake levels are declining. The current and past characteristics of precipitation within the region are important for understanding the driving factors behind the decline of Great Lake levels. Our paleoclimate reconstruction is based on the hydrologic continuity equation which characterizes the lake effect on precipitation, temperature, and evaporation rates. Our inferences are based on a Markov Chain Monte Carlo algorithm.