CRGAQS: Meteorological Modeling

1
CRGAQS Meteorological Modeling

• prepared for
• Southwest Clean Air Agency

prepared by Alpine Geophysics, LLC ENVIRON
International Corp
19 June 2006
2
MM5 Simulations
Run6 similar to WRAP configuration
3
Analysis Regions
WG Western Gorge CG Central Gorge EG
Eastern Gorge PS Puget Sound NC North
Coast WV William EC East of Cascades
4
August 2004 4km Humidity Bias/Error
5
August 2004 4km Humidity Mean
6
August 2004 4km TemperatureBias/Error
7
Aug 2004 4km TemperatureMean
8
Aug 2004 4km WindRMSE/Error
9
Aug 2004 4km WindMean Speed
10
Aug 2004 12km HumidityBias/Error
11
Aug 2004 12km HumidityMean
12
Aug 2004 12km TemperatureBias/Error
13
Aug 2004 12km TemperatureMean
14
Aug 2004 12km WindRMSE/Error
15
Aug 2004 12km WindSpeed Mean
16
Nov 2004 4km HumidityBias/Error
17
Nov 2004 12km HumidityMean
18
Nov 2004 4km TemperatureBias/Error
19
Nov 2004 4km TemperatureMean
20
Nov 2004 4km WindRMSE/Error
21
Nov 2004 4km WindMean Speed
22
Nov 2004 12km HumidityBias/Error
23
Nov 2004 12km HumidityMean
24
Nov 2004 12km TemperatureBias/Error
25
Nov 2004 12km TemperatureMean
26
Nov 2004 12km WindRMSE/Error
27
Nov 2004 12km WindMean Speed
28
MM5 Configuration Selection
Based on Overall Synthesis of the Candidate
Simulations, Run 6 is chosen as best performing
29
Run 6 Performance Evaluation

• Qualitative Precipitation Analysis
• Comparison with 0.25o (27km) CPC Episode Total
• CPC analysis does not extend over water
• Gorge Mean Value Analysis
• Compare Time Series of Spatial Mean Model/Obs.
• Wind Vector Analysis

30
Qualitative Precipitation4km Aug 2004
31
Qualitative Precipitation12km Aug 2004
32
Qualitative Precipitation4km Nov 2004
33
Qualitative Precipitation12km Nov 2004
34
Qualitative PrecipitationSummary

• MM5 generally captures spatial extent and
  magnitude of precip.
• MM5 shows smaller scale structure that can not be
  verified with the coarse CPC analysis
• MM5 underestimates precip. in
• Southeastern Oregon in August
• Oregon coast and Portland in November

35
Aug. 2004 Mean Humidity4km Gorge
West
Central
East
36
Aug. 2004 Mean Temperature4km Gorge
West
Central
East
37
Aug. 2004 Mean Wind Speed4km Gorge
West
Central
East
38
Nov. 2004 Mean Humidity4km Gorge
West
Central
East
39
Nov. 2004 Mean Temperature4km Gorge
West
Central
East
40
Nov. 2004 Mean Wind Speed4km Gorge
West
East
41
Mean Value AnalysisSummary

• MM5 generally overestimates humidity in August
  and underestimates in November
• Significant humidity phase difference in eastern
  gorge in mid-August.
• MM5 tends to underestimate daytime and
  overestimate nighttime temperatures
• Typical pattern but larger than normal
• Model overestimating temperatures in central
  gorge early in Nov. episode.
• Wind speed trends generally captured

42
Wind Vector Analysis

• Hourly wind vectors have been prepared
• Brief subset for presentation
• Black vectors MM5 winds
• Red vectors Obs. winds

43
Aug 2004 4km Wind Vector
44
Aug 2004 4km Wind Vector
45
Nov 2004 4km Wind Vector
46
Nov 2004 4km Wind Vector
47
Nov 2004 4km Wind Vector
48
Nov 2004 4km Wind Vector
49
Nov 2004 4km Wind Vector
50
Nov 2004 4km Wind Vector
51
Nov 2004 4km Wind Vector
52
Wind Vector AnalysisSummary

• MM5 generally captures up-gorge flow in Aug. and
  down-gorge flow in Nov.
• Flows follow gorge quite closely
• Gorge monitors show more variation between nearby
  monitors than MM5 fields
• Indication that obs. influenced by small scale
  features

53
Summary

• Significant Effort was put into determining the
  best performing options in MM5
• MM5 is still performing less well than
  performance benchmarks based on historic MM5
  applications.
• Gorge more complex than other areas. Monitors
  more likely to be influenced by small scale flows
• Gorge analysis regions have fewer sites than were
  used for benchmarks
• Nov. case has weak synoptic forcing. MM5
  traditionally performs better under stronger
  forced conditions
• MM5 Captures August Up-gorge flows
• MM5 Captures November Down-gorge flows

54
Next Steps

• Prepare emissions using WRAP 2002 inventory
  projected to 2004 replaced by ODEQ/SWCAA data for
  selecteed WA and OR Counties
• 36 and 12 km WRAP 2002 projected to 2004
• 4 km add OR and WA inventories
• Some issues/inconsistencies with OR/WA data
• CMAQ and CAMx base case modeling and model
  performance evaluation
• Evaluate need/usefulness of 1.33 km MM5 modeling