Numerical modeling and real time forecast of air pollution related to biomass burning on South America.

Saulo R. Freitas, Karla M. LongoSaulo R. Freitas, Karla M. Longosaulo.freitas@cptec.inpe.brsaulo.freitas@cptec.inpe.br -- http://meioambiente.cptec.inpe.brhttp://meioambiente.cptec.inpe.br

outline

– Biomass burning on South America– CCATT-BRAMS: mesoscale atmosferic-

chemistry-aerosol model– Near real time biomass burning emissions

estimate– Plumerise model for biomass burning smoke– Real time forecast

Local smoke plumeLocal smoke plume(deforestation fires)(deforestation fires)(picture from A. Andreae)(picture from A. Andreae)

smokesmoke

smokesmoke

smosmokeke

Regional smoke plumeRegional smoke plume~5~5 millions kmmillions km22

(Prins et al. 1998)(Prins et al. 1998)

GOESGOES--8 WF_ABBA 8 WF_ABBA (> 5000 fires)(> 5000 fires)

Biomass Burning and Smoke on South America

INPE developments on the atmospheric chemistry modeling

Ozone and PM2.5 biomass burning aerosol

Ozone at 1000 m ASL PM2.5 (bio. burn.) column

• SPACK: Pre-processor of chemical mechanism

• Pre-processor of emissions (anthropogenic, biogenic, biomass burning).

• Pre-processor of IC and BC for meteo-chemistry fields.

• 4DDA for meteo-chem fields.• Grid and sub-grid scale transport fully

coupled.• Plume rise model for fires and volcanoes

emissions.• Rad. CARMA and FAST-TUV (on-line

photolysis calculation).• Chemistry (RACM, CB07, RADM, etc).• Emission and deposition (dry and wet).• On-line with BRAMS regional model.• Being implemented in the CPTEC-GCM.

Coupled ChemistryCoupled Chemistry--AerosolAerosol--Tracer Transport Tracer Transport model to the Brazilian developments on the RAMSmodel to the Brazilian developments on the RAMS

CCATTCCATT--BRAMSBRAMS

Biomass burning emissions inventoryBiomass burning emissions inventoryRegional scale Regional scale –– daily basisdaily basis

[ ]

[ ] . . . ,veg firefveg vegM a

density of carbon data

land use data

near real time fire product

emission & combustion factors

mass estimation

CO source emission (kg m-2day-1)

Aboveground Biomass Density

Olson's carbon in live vegetation

Provides an estimate of the carbon content for each Olson’s vegetation class.

Saatchi et al., Global Change Biology (2007)

Amazon basin1 km resolution

Global data with 0.5 degree resolution

Fires position, timing and size using remote sensing products

Fires from AVHRR-MODIS-GOES: INPE (A. Setzer)

Fires WF_ABBA (GOES)CIMSS (E. Prins)

provides the diurnal cycle of the burning, each 1/2 hour.provides an estimate of

the instantaneous fire size.

However, the burned area and the AGB are the main source of uncertainties for biomass burning emissions estimates

Prep-Chem-Sources pre-processorBiomass burning sources

–Brazilian Biomass Burning Emission Model (Freitas et al., 2005; Longo et al., 2007): plume rise mechanism, daily and model resolution.–GFEDv2 (van der Werf et al., 2006): 8days/monthly – 1x1 degree.–Emission Factors from Andreae and Merlet (2001), Ward et al 1992, Yokelson et al (200X)

CO2 CO CH4 NHMC C2H2 C2H4 C2H6 C3H4 C3H6 C3H8 1_butene i-butene tr_2_butene cis_2_butene butadiene

n_butane i-butane 1_pentene 2_pentene n_pentane 2_Me_Butene 2_Me_butane pentadienes Isoprene cyclopentene cyclopentadiene 4_me_1_pentene 2_me_1_pentene 1_hexene hexadienes

n_hexane isohexanes heptane octenes terpenes benzene toluene xylenes ethylbenzene styrene PAH Methanol Ethanol 1_Propanol 2_propanol

Butanols cyclopentanolphenol Formaldehyde Acetald Hydroxyacetaldehyde Acrolein Propanal Butanals Hexanals Heptanals Acetone 2_Butanone2_3_Butanedione Pentanones Hexanones

110 speciesHeptanones Octanones Benzaldehyde Furan 2_Me_Furan 3_Me_Furan 2_ethylfuran 2_4_dime_furan 2_5_Dime_furan Tetrahydrofuran 2_3_dihydrofuran benzofuran Furfural Me_format Me_Acetate Acetonitrile AcrylonitrilePropionitrilepyrrole trimethylpyrazolemethylamine dimethylamine

ethylamine trimethylaminen_pentylamine 2_me_1_butylamineHFo HAc Propanoic H2 NOx NOy EF_N2OEF_NH3EF_HCNcyanogenSO2 DMS COS CH3ClCH3BrCH3I Hg PM25 TPM TC ,OC ,BC

Biomes: TropFor, ExtratropF, Savanna, Pasture, charcoal, waste, lab

inter-comparison of bioburn inventories

Longo et al., 2009 –under review (EGU-ACP)

Including emission in the model

Biomass burningBiomass burningand wildfiresand wildfires

Smoldering : mostly surface emission.Flaming: mostly direct injection in the PBL,

free troposphere or stratosphere.

1( )

air

air

E

E

F

Fz

Fz

first phys.model layer

injectionlayer

total emission flux: being the smoldering fraction

smoldering term :

flaming term :

Plume rise model

h

Injection Injection layerlayer

flamingflamingemissionemission

smolderingemission

Example in the model:

diurnal cycle of the burning:

( ) ( )E Et r t

0 4 8 12 16 20 24Hora Local (Fuso de Brasília)

0

250

500

750

1000

1250

1500

1750

2000

2250

Núm

ero

méd

iode

foco

sde

fogo

( )r t

t im e

15 LT

22

2

2

2

2, ,

( )

( )

e microp physics

v v vv ve

microphysics

c c cc

microphysics

ice rain ice rain

w ww gB wt z R

T T g Tw w w T Tt z c R t

r r rw w r rt z R t

r r rw w rt z R t

r rw w r

t z R

2

,, sedim

( , , , , ), sedim

ice rainice rain

microphysics

v c rain icemicrophysics

bulk microphysics:

Kessler,

entr

1969

rt

T r r r r t

only lateral (non-organized) entrainment

L wR

; Berry,1967

Ogura & Takahashi,1971

bulk bulk microphysicsmicrophysics

thermothermo--dynamicsdynamics

dynamics for dynamics for WW

Latham, 1994; Freitas et al., 2006, 2007

The 1D cloud model: governing equations (original formulation from the PLUMP model)

Including plume rise mechanism troughIncluding plume rise mechanism trough””supersuper--parameterizationparameterization”” conceptconcept

280tot

parcel env

al

Th kWm

rwT

230 total

parcel env h kWmT T

rw

2

23080

w h kWmw h kWm

Injection Injection layerlayer

plume toplower bound

plume topupper bound

1D plume1D plume--rise model for vegetation fires rise model for vegetation fires Biome: Biome: ForestForestTime duration: 50 mnFire size: 20 haHeat flux: 80 kWm-2 / 30 kWm-2

Example of CO source emission field with the plumeExample of CO source emission field with the plume--rise for rise for vegetation fires at the CATTvegetation fires at the CATT--BRAMS host modelBRAMS host model

South AmericaSouth Americamostly forest firesmostly forest fires

AfricaAfricamostly savanna firesmostly savanna fires

flamingflamingemissionemission

smolderingemission

CO with plume-riseCO without plume-rise

horizontal wind CO source emission field

Example of CO Example of CO withwith and and withoutwithout plumeplume--rise at level 5.8 kmrise at level 5.8 km::

22 SEP 200222 SEP 2002

CATTCATT--BRAMS comparison with AIRS 500 hPa COBRAMS comparison with AIRS 500 hPa CO

CO (ppb) from CO (ppb) from

AIRSAIRS at 500 hPaat 500 hPaMcMillan et al., GRL 2005.

1. Atmospheric InfraRed Sounder (AIRS) onboard NASA's Aqua satellite.

2. CO abundances are retrieved from AIRS 4.55 m spectral region.

Model CO (ppb) at ~5.8 kmModel CO (ppb) at ~5.8 kmwithout plume rise with plume without plume rise with plume riserise

Smoke plume rise under calm environment

Pictures taken by M.O. Andreae and M. Welling

Smoke plume rise under windy environment

Pictures taken by M. Welling.

The dynamic entrainment rate formulation

.

2 2

( )

( ) ( )( )

Consider a cylindrical volume of radius and depth , the in-cloud horizontal mass flux is:

The mass gained by cloud in is:

The definition of the m

h env e

h env e

f u u

m f

Rz

R z t u u z tt

R

2

21

2

1

(

( )( ) .

)

ass entrainment rate is

Assuming that cloud env

env

en

eentr

clou

tr e

d

u u R z tmm t tR

u uR

z

List of symbols: enviroment air, cloud mass densities

enviroment air and cloud horizontal wind velocities cloud radius at height z

, :, ::

env cloud

eu uR

22

2

2

2

2, ,

( )

( )

e microp physics

v v vv ve

microphysics

c c cc

microphysics

ice rain ice rain

w ww gB wt z R

T T g Tw w w T Tt z c R t

r r rw w r rt z R t

r r rw w rt z R t

r rw w r

t z R

2

,, sedim

( , , , , ), sedim

ice rainice rain

microphysics

v c rain icemicrophysics

bulk microphysics:

Kessler,

entr

1969

rt

T r r r r t

only lateral (non-organized) entrainment

L wR

; Berry,1967

Ogura & Takahashi,1971

bulk bulk microphysicsmicrophysics

thermothermo--dynamicsdynamics

dynamics for dynamics for WW

Latham, 1994; Freitas et al., 2006, 2007

The 1D cloud model: governing equations (original formulation from the PLUMP model)

The 1D cloud model: including the environmental wind effect on cloud scale dilution-

governing equations

2

2

2

2

2

( )

( ) ( )

(

(( )

)

)

emicrop physics

v v vv ve

microphy

entr

e entr e

entr e

entr

sic

v e

s

c

v

w ww gB w t z R

T T g Tw w w T T t

w

u uw w u u u ut z R

z cT T

r

R t

r r rw w r r t

rz R t

r wt

2

2

2

,,

, ,, sedim

c cc

microphysics

ice rain ice rain ice rainice rain

microp

entr e

entr c

entr ice r

hysic

ain

s

r rw rz R t

r r rw w r

dynamic entrainment

u uR

r

rt z t

R wt

R

6 15 2

( , , , , ), sedimv c rain icemicrophysics

entr

bulk microphysics:

Kessler, 1969; Berry,1967

Ogura & Takahashi,1971

R w R Rz R

T r r r r t

thermothermo--dynamicsdynamics

dynamics for dynamics for WW

dynamics for dynamics for UU

bulk bulk microphysicsmicrophysics

equation for equation for radius sizeradius size

See Freitas et al. (2009 ACPD) for 1d cloud model comparisons with fully 3D ATHAM simulations

Aerosol Optical Depth (550 nm) : MODIS x MODEL

Model evaluation with SMOCC/RaCCI 2002 using near surface measurements (CO and PM2.5)

Model evaluation with SMOCC/RaCCI 2002 with airborne measurements (CO)

Freitas et al., 2009

Effect of time resolution of the inventory

Longo et al., 2009(under review)

Air Quality forecast for South America: http://meioambiente.cptec.inpe.br

Surface level CO (ppb)12Z12SEP2007 500 hPa CO (ppb)

21Z14SEP2007

Biomass burning pollution

Mega Cities pollution new fresh plumeinjected by pyrocumulus

Forecast 21UTC14SEP2009

Field campaign to evaluate plume model: prescribed fire in Alta Floresta (aug – 2010)

The lower boundary condition

the closurethe closure

Morton, Taylor & Turner (1956): "Turbulent grav. convection from maintained and instantaneous sourc

buoyancy flux

plume

es

radiu

"

s65

(

p e

gF Ac p

R z

w z

boundary condition for

density correction

boundary condition for

where: entrainment coeffic

ient,

1/3

5/3

1/3

5 0.96

56 0.9

0.2

0.

) ( )

( )( )1

vv

v

e

e v

e

v

v

F

F

z

zFg

T z

w

z

T z T

virtual boundary height19 surfR

-17

-2convective energy from fire (Wm )

(McCarter & Broido, 1965)

1.5 to 2.1 10(heat flux) fuel load / combustion factor

plume area instantaneous fire size

joules kgflux

flux

A

E

0.4 0.8

h

t c

E

h c

flaming phase duration

(water flux

flux ) .5 ct

W 0

Example of the diurnal cycle of CO source emission fieldExample of the diurnal cycle of CO source emission field

CO withoutplume-rise

CO withplume-rise

Source

emission

00Z time 24Z 00Z time 24Z

Equiv. Pot.Temperature

Plume-rise of vegetation fires:typical energy fluxes (kWm-2)

Biome type Lower bound kWm-2

Upper bound kWmkWm--22

Flaming consumption

Tropical forest 30. 80. 45%

Woody savanna -cerrado

4.4 23. 75%

Pasture - grassland cropland

3.3 97%

Refs: Carvalho et al, 1995-2001-2005 (com. pessoal);

Riggan et al, 2004;

Ward et al, 2002;

Ferguson et al, 1998;

Cochrane et al; 200X-com. pessoal;

Miranda et al, 1993.

Directions to improve

1) Plume model needs the initial plume size (or fire size) and convective energy

Size of fire : 10 ha (dry/calm and wet/windy cases)

Main injection layer simulated by the ATHAM

model