Regional Air Quality Modeling Progress at NOAA/NWS/NCEP

Jeff McQueen, Pius Lee, Jianping Huang, Ho-Chun Huang, Perry Shafran,

Eric Rogers, Geoff DiMego, Ivanka StajnerJune 10, 2015

CMAQ V4.6.4 CB05/AERO-4• V4.6.2 : April 28, 2014

• Inclusion of latest EPA Carbon Bond 5 (CB05) chemical mechanism.• Inclusion of AERO-4 aerosol chemistry.• Updated anthropogenic emissions with 2014 Dept. Energy projections.

• V4.6.3: June 13, 2014• Modulate fugitive dust emission: suppress over ice/snow.• Incorporate NESIDS HMS wild fire smoke. CONUS • Incorporate real-time surface dust emissions (wind dependent). CONUS • NTR, organic nitrate photolyzed and removed quicker.• Layer specific time step was added to speed up code.

• V4.6.3v2: June 27, 2014• Correction to overestimates of dust emissions

• repartition percent going to PM2.5, mix beyond first level

• V4.6.4: July 16, 2014 With NAM Parallel• Turned off gas emissions from fires.• Ozone predictions will not be impacted by inclusion of smoke emissions.

• OPERATIONAL: January 29, 20153

CMAQ V4.6.5 CB05/AERO-4

• V4.6.5 : May 1, 2015• All emissions based on EPA 2011 Inventories (NOAA/ARL)• Updated anthropogenic emissions with 2015 Dept. Energy projections• Increased vertical levels from 22 to 35 (NOAA/ARL)• Using operational NEMS Global Aerosol Capability (NGAC) for dust lateral boundary

conditions (J. Wang)• Testing PM2.5 ESRL bias correction (J. Huang)• Update BlueSky smoke emission system (H-C Huang)

• Evaluations• Ozone: case studies and FVS statistics

• July 7, 2014: Southern California• June 12, 2015: North East U.S.

• PM: • NGAC boundary impact cases

• May 10, 2015: Dust• June 10, 2015: Smoke

• Bias Correction: FVS statistics• Potential Impact of higher resolution meteorology 4

1 hour average Ozone PredictionsJuly 08 , 2014 Case

5

EXP

Operational

Experimental : V4.6.3 Ozone Daily max vs obs

Met BIAS NAM, NAM Nest, RAPSouth West Coast (July 2014)

6Warm, Dry bias with lower PBL than observedConus Nest improved DPT and Winds

2 m Temperature 2 m Dew point T

10 m windDaily PBL Height (WEST U.S.)

7

NAM Parent, Nest, DNG vs URMA Temperatures

July 8, 2014 21 UTC (33 H forecast)

Temperatures cooler than observed near San Bernadino Dewpoints drier than observed near coast

Dew Point Temperature2 m Temperature

URMA 2.5 km

NAM 12 km

8

NAM Parent, Nest vs URMA Wind Vectors July 8, 2014 21 UTC (33 h forecast)

- Complex flow fields delineated by URMA- Turning and slowing of winds around high terrain near San Bernadino not captured by NAM 12

URMA 2.5 km

NAM 12 km

Nest 4 km

Ozone AQI

10

NAM Parent, NestJune 8, 2015 LA Basin

Dew Point Temperature

10 m wind/topo

URMA 2.5 km

NAM 12 km

Connecticut Department of Energy and Environmental Protection

CMAQ continues its 10 ppb early-season ozone under-prediction

13

NAM Parent, Nest vs Obs TemperaturesJune 12, 2015 18 UTC (12 H forecast)

NAM-12: Temps ~ 2-4 ° cooler & 4-8 ° moister in Upstate NY/CT Can contribute to lack of ozone productionNest: Better inland predictions, Strong drying near the coast : downward mixing ?

Dew Point Temperature2 m Temperature

NAM Nest

NAM 12 km

Improvements with High Res Met ?May 15 –June 14, 2015 9h forecast : CONUS

14

Wind Speed 2m Temperature

Improvements with High Res Met ?May 15 –June 14, 2015: NE/SW Bias

15

Dew point T

2m Temperature BIAS

NE U.S. SW U.S.

16

Saharan Dust in GulfMay 10, 2015

17

Saharan Dust in GulfMay 10, 2015

18

Canadian Smoke over E. U.S.June 10, 2015

HYSPLIT BlueSky V3.5.1

HYSPLIT Oper BlueSky

Total Column smoke Concentration (ug/m3) CMAQ Oper Sfc PM2.5 Prediction

Connecticut Department of Energy and Environmental Protection

Aqua/Terra Images from June 9- June 11

20

Canadian Fire smoke over DMV June 10, 2015

1 hour Avg PM2.5 PerformanceSkill Score: Prod, para, bias corrected

21

East US

West U.S.

CONUS

Bias correction improvement but :Not capturing day to day variability & larger events

22

Summary• Smoke, dust, ozone, anthropogenic PM : unified system using CMAQ V4.6.4

• Ozone predictions improved with latest changes to V4.6.4– Although strong under-prediction in California still evident

• PM predictions:– Main positive impact from updated emissions and NGAC LBCs (V4.6.5)– Bias Correction: More evaluation needed (J. Huang)– Better emissions from wild fire smoke (H-C Huang)

• NWS Un-Restricted Mesoscale Analysis (2.5 km) proved useful for evaluating NAM 12 and 4 km predictions

• High Res Met useful for identifying processes not resolved with 12 km NAM - 4 km nest captured blocking flow in San Bernadino Valley, California

- 4 km nest performance of AQ sensitive weather fields best so far (what about fire wx nest ?)

• What is the impact of smoke/dust on Meteorology ?- possible impact for June 9-11

• Met performance should be carefully evaluated while proceeding to address other system errors

- Concern that we are making changes to chem/emissions that mask 1st order met errors

23

Future plans• Short term (1-2 years)

– Include NGAC real-time full aerosol boundary conditions– Improve smoke emissions

• Update USFS BlueSky emissions – (forest load, consumption, spread emissions)– Smart fire behavior using real-time met

• Evaluate plume rise (additional met constraints)

– Improve dust emissions• NAM gust vs speed• DNG winds• Soil moisture impact

– Include ESRL bias correction for O3/PM• At stations • spreading technique to grid

– Short-range High Resolution smoke/dust modeling– Partial inline NMMB-CMAQ V5.1

• Dust first• Impact on weather• Highly optimized

HYSPLIT smoke driven with NAM 4 kmJuly 19, 2014