Simulated Seasonal Temperature and Precipitation Patterns in WRF for the Western United States
Location
Eccles Conference Center
Event Website
http://water.usu.edu/
Start Date
3-30-2011 11:20 AM
End Date
3-30-2011 11:40 AM
Description
The Weather Research & Forecasting (WRF) model was initially developed as a mesoscale model intended to study features such as thunderstorm complexes and frontal boundaries; however, WRF has expanded into the realm of regional climate modeling. WRF requires lateral boundary forcing data; thus, necessitating the use of global climate model output to drive WRF for future climate simulations. This can be problematic as any biases in the global climate model can become greatly magnified in WRF. We set out to analyze these biases and determine a method to correct the forcing data from the global climate model prior to input into WRF. We performed sensitivity testing with WRF version 3.2 coupled with the Community Land Model (CLM) version 3.5 forced with National Center for Environmental Prediction (NCEP) reanalysis data to find the optimal combination of physics options. After we had achieved good results with the NCEP driven runs, we repeated the experiment except driven by output from the Community Climate System Model (CCSM) version 3.0 from the 20th Century experiment. We then quantified the biases in WRF caused solely by changing forcing data. A third simulation was then performed using CCSM data that had been biased corrected using a simple linear regression model (based on years 1949 - 2099 using the A2 emission scenario). All WRF simulations were performed for years 1989 through 1999 on a 110 X 110 grid with 32km resolution centered over northern Utah. Our results show that WRF produces good winter temperature and precipitation simulations when forced with NCEP data. However, using CCSM forcing data produces large wet biases, particularly over higher elevations. Bias correcting the data with a regression model helps reduce these biases, producing results that are much closer to those of the NCEP driven run. For summer simulations, all three runs tended to overestimate the North American monsoon. Here, we examine the causes of the summer biases in WRF and compare the results for all seasons with observational data.
Simulated Seasonal Temperature and Precipitation Patterns in WRF for the Western United States
Eccles Conference Center
The Weather Research & Forecasting (WRF) model was initially developed as a mesoscale model intended to study features such as thunderstorm complexes and frontal boundaries; however, WRF has expanded into the realm of regional climate modeling. WRF requires lateral boundary forcing data; thus, necessitating the use of global climate model output to drive WRF for future climate simulations. This can be problematic as any biases in the global climate model can become greatly magnified in WRF. We set out to analyze these biases and determine a method to correct the forcing data from the global climate model prior to input into WRF. We performed sensitivity testing with WRF version 3.2 coupled with the Community Land Model (CLM) version 3.5 forced with National Center for Environmental Prediction (NCEP) reanalysis data to find the optimal combination of physics options. After we had achieved good results with the NCEP driven runs, we repeated the experiment except driven by output from the Community Climate System Model (CCSM) version 3.0 from the 20th Century experiment. We then quantified the biases in WRF caused solely by changing forcing data. A third simulation was then performed using CCSM data that had been biased corrected using a simple linear regression model (based on years 1949 - 2099 using the A2 emission scenario). All WRF simulations were performed for years 1989 through 1999 on a 110 X 110 grid with 32km resolution centered over northern Utah. Our results show that WRF produces good winter temperature and precipitation simulations when forced with NCEP data. However, using CCSM forcing data produces large wet biases, particularly over higher elevations. Bias correcting the data with a regression model helps reduce these biases, producing results that are much closer to those of the NCEP driven run. For summer simulations, all three runs tended to overestimate the North American monsoon. Here, we examine the causes of the summer biases in WRF and compare the results for all seasons with observational data.
https://digitalcommons.usu.edu/runoff/2011/AllAbstracts/15