Authors: Jennifer Boehnert*, NCAR, Mari Tye, NCAR, Olga Wilhelmi, NCAR, Kevin Butler, ESRI, Eric Krause, ESRI
Topics: Geographic Information Science and Systems
Keywords: GIS, climate, downscaling
Session Type: Virtual Paper
Start / End Time: 4:40 PM / 5:55 PM
Room: Virtual 28
Presentation File: No File Uploaded
Global Climate Models (GCMs) are mathematical equations that compute future projected climate at a global coarse resolution. GCM simulations are appropriate for global scale analysis however, are problematic when performing local scale studies which require much higher spatial resolution. Many diverse disciplines who already use GIS for analysis, need climate projections for local-scale studies. This talk will focus on a collaborative project between NCAR and ESRI, to review, test, and evaluate the use of tools in ESRI ArcGIS Pro’s called ‘Empirical Bayesian Kriging EBK Regression Prediction’ to downscale GCM output to a larger scale resolution for local applications and for use in GIS. We will discuss the steps taken to test and evaluate the results when downscaling the variable Mean Temperature from the Community Climate System Model (CCSM-4), which is natively a 1-degree spatial resolution, to larger resolutions in five distinct domains across the continental US. We will also discuss our evaluation of the impact that different output resolution have on temperature estimates and the tipping point when downscaling becomes re-sampling instead of producing more information.