modeling biomass burnings by coupling a sub-grid scale plume model with adaptive grid cmaq

Modeling biomass burnings by coupling a sub-grid scale plume

model with Adaptive Grid CMAQ Aika Yano

Fernando Garcia-Menendez Yongtao Hu

M. Talat OdmanGary Achtemeier (USDA Forest Service)

D. Scott McRae (NC State)

9th Annual CMAS Conference11 October, 2010

sub-grid scale plume model: Daysmoke

Adaptive Grid - CMAQ

AG-CMAQ vs Daysmoke

+ Eularian + Unit in

concentration- Uniform conc in

grid cell+ Constant data pts to calculate

+ Lagrangian+ Unit in mass+ No emission

dilution- Varying data pts to be calculated

per time stepA

GD

-CM

AQ

Objective: combine the two models to obtain the most detailed conc. in a grid model

Adaptor

CMAQ

Daysmoke

MM5 (WRF)

ChemicalCoupling Ambient concentrationsParcel/Puff/Plume concentrations

Grid

Grid

Grid

Particles

GridSupport?

Distributions

Parcels mixed in grid cells

Meteorology

MeteorologyParticles/CCN

Handover

d = downwind distance

Emissions

ConcentrationsMeteorology

How to combine Daysmoke in AG-CMAQ?

New CMAQ setup

Inside driver Inside sciproc VDIFF

CMAQ DAYSMOKEXADVRest of

CMAQ codes

CMAQ_DAYSMOKE: Between VDIFF and COUPLE

• Calculate locations of 1kg PM2.5 parcels per Daysmoke time step

DAYSMOKE

• Convert Daysmoke emissions to concentrations

• Determine when and where Daysmoke results can be carried over to AG-CMAQ

HANDOVER • Convert parcels beyond the boundary into concentrations in Adaptive grid (for each species)

• update grid conc.

OutConcPlus

Handover

0

100000

200000

300000

400000

02000400060008000100001200014000160001800020000

con

c differen

ce (m

icrog

/m3)

downwind distance (meters)

sum of conc error vs. downwind distance

Handover

Remain in DaysmokeEmission added to grid concentration

Atlanta Smoke Case Feb. 28 2007

Oconee NF (1542 acres) 5 yr old

10:00 am - 3:00 pm total 5hr

Piedmont NWR (1460 acres) 1 yr old

11:00 am - 2:00 pm total 3hr

Atlanta Smoke Case Feb. 28 2007

Measurement Sites

CMAQ vs AGD-CMAQ

Model Comparison

Mean Normalized Error

0%

20%

40%

60%

80%

100%

120%

CMAQ

AG-CMAQ

AGD-CMAQ

Summary• Inputting PB smoke emissions into grid

concentrations can be varied in time and place using Handover method in AG-CMAQ coupled with Daysmoke (AGD-CMAQ) .

• On average, AGD-CMAQ has less error from the measured concentration than CMAQ and AG-CMAQ

• AGD-CMAQ seems to pick up the measured concentration peaks better

Adaptor

CMAQ

Daysmoke

MM5 (WRF)

ChemicalCoupling Ambient concentrationsParcel/Puff/Plume concentrations

Grid

Grid

Grid

Parcels

GridSupport?

Distributions

Parcels mixed in grid cells

Meteorology

MeteorologyParticles/CCN

Handover

d = downwind distance

Emissions

ConcentrationsMeteorology

Future work: changes in grid adaptation

Using Fourier Transform for grid adaptation

Thank youQuestions/Comments

Mean Normalized Error

0.00%

20.00%

40.00%

60.00%

80.00%

100.00%

120.00%

1. McDonough

2. South Dekalb

3. Confederate Ave.

4. Ft.McPherson

5. Jefferson St

6. Fire Station 8

CMAQ

AGD-CMAQ

AG-CMAQ

Fort McPherson

0

20

40

60

80

100

120

140

160

9:00 AM 12:00 PM 3:00 PM 6:00 PM 9:00 PM

Feb 28 2007 EST

PM

2.5

(m

g/m

3)

CMAQAGD-CMAQObservationsAG-CMAQ