CONTROL ID: 1489340
TITLE: Do high-resolution global models add value to climate model predictions? Results from a cluster analysis
AUTHORS (FIRST NAME, LAST NAME): Rui Mei1, Salil Mahajan1, Jitendra Kumar1, Forrest M Hoffman1, Moetasim Ashfaq1
INSTITUTIONS (ALL): 1. Oak Ridge National Lab, Oak Ridge, TN, United States.
ABSTRACT BODY: We apply the statistical technique of cluster analysis to the North American Regional Reanalysis (NARR) and model outputs at different resolutions from Community Atmospheric Model version 4 (CAM4) simulations to investigate the value added by higher-resolution in climate modeling. The CAM4 simulations were conducted at T341, T85 and T42 spectral truncations that roughly correspond to 0.35, 1.4 and 2.8 degree horizontal grid spacings. The atmospheric variables clustered include vertical profiles of horizontal winds, air temperature, specific humidity and geopotential height over the United States. Our preliminary analysis indicates that high-resolution model simulations perform better than coarse resolution simulations in capturing the frequency of occupation of atmospheric states derived from NARR over 1979–1983. We further test the hypothesis that higher-resolution climate models should outperform low resolution versions over the western United States during wintertime, where high resolution can potentially improve the simulation of cold season processes that are driven by the interactions between topography and large-scale circulation, and over the central and eastern United States during summertime, where high resolution can potentially improve the simulation of warm season processes that are governed by mesoscale convection systems. In addition, we will extend the cluster analysis to the Atmospheric Radiation Measurement (ARM) observations to investigate the uncertainties in the NARR data.
KEYWORDS:  INFORMATICS / Data mining,  GLOBAL CHANGE / Earth system modeling,  GEOGRAPHIC LOCATION / North America.
(No Image Selected)
(No Table Selected)
Previously Presented Material:
CONTACT (NAME ONLY): Rui Mei
CONTACT (E-MAIL ONLY): meir at ornl dot gov
TITLE OF TEAM: