12km MM5 Issues Mar8 9 2005

12 km MM5 Modeling Issues: 

12 km MM5 Modeling Issues Susan Kemball-Cook, Yiqin Jia, Chris Emery and Ralph Morris, ENVIRON Corporation Int., Novato, CA Zion Wang and Gail Tonnesen University of California, Riverside WRAP Modeling Forum Meeting, San Francisco, CA March 8-9, 2005

New WRAP 12 km Run Configuration: 

New WRAP 12 km Run Configuration 12 km configuration identical to 36 km grid, except for cumulus scheme where the Betts-Miller scheme was used with the 36 km configuration

Slide3: 

All 12 km subdomains within benchmark WRAP January Wind 12km vs. 36km

Slide4: 

All 12 km subdomains within benchmark WRAP March Wind 12km vs. 36km

Slide5: 

WRAP July Wind 12km vs. 36km

Slide6: 

WRAP October Wind 12km vs. 36km

Slide7: 

WRAP December Wind 12km vs. 36km

Slide8: 

WRAP January Temp 12km vs. 36km

Slide9: 

WRAP March Temp 12km vs. 36km Cold bias, slight improvement with 12km grid

Slide10: 

WRAP July Temp 12km vs. 36km

Slide11: 

WRAP October Temp 12km vs. 36km

Slide12: 

WRAP December Temp 12km vs. 36km

Slide13: 

WRAP Jan Humidity 12km vs. 36km

Slide14: 

WRAP March Humidity 12km vs. 36km

Slide15: 

WRAP July Humidity 12km vs. 36km

Slide16: 

WRAP Oct Humidity 12km vs. 36km

Slide17: 

WRAP Dec Humidity 12km vs. 36km

Conclusions: 12 km vs. 36 km MM5 Performance for Surface Met: 

Conclusions: 12 km vs. 36 km MM5 Performance for Surface Met Improvement in surface wind performance Improvement in temperature performance, reduces, but does not eliminate, winter cold bias Generally, slight improvement in humidity (reduces wet bias in SW)

Slide19: 

January Precipitation Precipitation predicted with good skill overall Overprediction in North subdomain Rainfall too intense in parts of pacNW subdomain Little difference between 12km an 36km 12km 36km

Slide20: 

March Precipitation Good agreement with observations for both MM5 12 km and 36 km simulations 12km 36km

Slide21: 

July Precipitation Widespread over prediction of rainfall—no cumulus parameterization in 12km MM5 precipitation field more granular than observations Maxima are too intense, although general pattern ok in southwestern US 12km better performance than 36km 12km 36km 12km 36km

Slide22: 

October Precipitation Good agreement for both MM5 simulations with observations 12km 36km

Slide23: 

December Precipitation Fairly good agreement with observations MM5 too intense in lower Sierra Nevada and Vancouver area 12km 36km

Summary of 36/12 km Results: 

Summary of 36/12 km Results The new 36 km and 12 km runs show improvement in the modeled precipitation fields, particularly in the summer in the southwest and in winter in the Pacific Northwest. Reduction in the overprediction of summer convective rainfall in both 36 and 12 km runs. Wet biases in the surface humidity and cold biases in the surface temperature reduced in the North and DesertSW subdomains in the summer months. New WRAP run was optimized for summer performance over the WRAP region, but winter performance in the west did not deteriorate significantly.

Summary of 36 /12 km Results: 

Summary of 36 /12 km Results Surface wind performance improved throughout the year due to observational nudging of surface winds.   Showed improved temperature performance in summer in the west, and slightly worse performance in winter in western U.S.   Small overall degradation in performance in the east. Some of this was the result of eliminating surface analysis nudging of temperature and moisture that was done in WRAP_0. Also due to change from Kain-Fritsch to Betts-Miller CUP.

Summary of 12 km Results: 

Summary of 12 km Results 12 km run is within or near performance benchmarks for surface wind and humidity over the annual cycle of 2002. Surface temperature generally falls outside the bias benchmark. It is possible that this is due to terrain resolution effects – improved temperature performance is seen going from 36 km to 12 km grid. The new 12 km WRAP run has been significantly improved in terms of its surface performance relative to the WRAP_0 12 km run. Improvement in summer rainfall and surface humidity performance in the WRAP region reduces biases of original 12 km run.   

Conclusions: 12km vs. 36km MM5: 

Conclusions: 12km vs. 36km MM5 Better surface meteorology and precipitation model performance in western US using a 12 km grid Whether this results in better air quality model performance remains to be seen 12 km MM5 runs are available, so decision on need for 12 km modeling should be based on emissions and air quality modeling issues