What We Can Learn from the Midlatitudes – Intercomparison ... talks/Barth.pdf · What We Can Learn from the Midlatitudes – Intercomparison Studies Mary C. Barth Mesoscale and

What We Can Learn from the Midlatitudes –Intercomparison Studies

Mary C. BarthMesoscale and Microscale Meteorology and

Atmospheric Chemistry DivisionsNational Center for Atmospheric Research

Si-Wan Kim, NCAR; now at NOAA/ESRL/CSDChien Wang, MITKen Pickering, NASA/GSFCLesley Ott, Univ. MarylandG. Stenchikov, Rutgers Univ.Maud Leriche, Sylvie Cautenet, CNRS/U. Blaise-PascalAnn Fridlind, Andy Ackerman, NASA/GISSJean-Pierre Pinty, Celine Mari, Lab. D’Aerologie, CNRSVlado Spiridonov, Hydrological Inst., MacedoniaJohn Helsdon, Richard Farley, SDSMT

Motivation

• Convective processing of chemical species isimportant to– Moving pollutants to upper troposphere

– Cleansing the atmosphere (rain out)

• Large-scale models produce inconsistentresults for convective transport of scalars

Results from the NCAR CCSM UsingDifferent Convection Parameterizations

Mixing ratio of surface tracer averaged over the TOGA-COARE region as afunction of day (December 18 – January 8) and pressure

From Phil Rasch, EGS talk, 2003

Motivation

• Convective processing of chemical species isimportant to– Moving pollutants to upper troposphere

– Cleansing the atmosphere (rain out)

• Large-scale models produce inconsistentresults for convective transport of scalars

• Convective-scale models produce reasonablyrepresent convective transport

Results From the COMMAS ConvectiveCloud Model Coupled With Chemistry

From Skamarock et al. (2000)

Motivation

• To improve sub-grid convective transport andwet deposition in large-scale models– multiple convective-scale models can be used to

obtain general characteristics of these processes.

• The Chemistry Transport in Deep ConvectionIntercomparison– means to calibrate a variety of convective-scale

models coupled with chemistry.

• Determine what the variability amongreputable cloud chemistry convective modelsis for a given storm.

Acknowledgment

The Chemistry Transport in Deep ConvectionIntercomparison

6th International Cloud Modeling Workshop

July 2004Hamburg, Germany

WMO

Simulate the 10 July 1996 STERAOstorm

Colorado

NebraskaWyoming

Chemistry Transport by Deep Convection

Primary Species:

Ozone (O3) – tracer

Carbon monoxide (CO) – tracer

Nitrogen oxides (NOx = NO + NO2) – enhanced by lightning

Soluble species:

Nitric acid (HNO3)

Hydrogen peroxide (H2O2)

Formaldehyde (CH2O)

affected by microphysics parameterization

Initialization

• Soundings for T, qv, u, and v

• Initiation via 3 warm bubbles

• Aircraft vertical profiles for chemical species

Initial profile

MOZART profile

Points from aircraftobservations

360x328x25 km

2x2 km horiz

0.500 km vert

3-30 s timestep

Offline chemistry transport

Gas chemistry on

Lightning-NOx param

Predict Mass only:

cw, rain, ice, snow, hail (Tao and Simpson, 1993)

No aerosols

3d GCE Model

(Tao and Simpson, 93)

U. Md.

Pickering, Ott

Stenchikov

120x120x20 km

1x1 km horiz

Aqueous chemistryPredict Mass only:

cw, rain, ice, snow, hail,

(Lin et al 83)

3d

No radiation

Spiridonov

120x120x20 km1x1 km horiz

51 vert levels

5 s timestep

Chemistry onLightning-NOx param

Predict Mass only,cw, rain, ice, snow, hail

No aerosols

3d, RAMSLeriche, Cautenet

120x120x20 km

1x1 km horiz

50 vert levels

Scavenge soluble species

Electrical scheme

Predict Mass only:6 hydrometeor classesNo aerosols

3d anelastic

MPDATA advec.

Meso-NH

Pinty, Mari

120x120x20 km

1x1 km horiz

0.250 km vert

No chemistryPredict Number, Mass & size

Prognostic CCN and IN

3d,

Interactive radiation

DHARMA

Fridlind, Ackerman

120x120x20 km

1x1 km horiz

0.4 km vert

Chemistry onPredict Number and Mass


3d pseudo-elastic


C. Wang

160x160x20 km

1x1 km horiz

50 vert levels

10 s timestep

Chemistry onPredict Mass only:

cw, rain, ice, snow, hail

(Lin et al, 83)

No aerosols

3d, flux form Runge-Kutta

No radiation

WRF-AqChem

Barth, Kim

ConfigurationChemistryCloud Microphysics and AerosolsDynamics, thermodynamics, radiationModel

360x328x25 km

2x2 km horiz

0.500 km vert

3-30 s timestep


Gas chemistry on

Lightning-NOx param

Predict Mass only:


No aerosols

3d GCE Model


U. Md.

Pickering, Ott

Stenchikov

120x120x20 km

1x1 km horiz



(Lin et al 83)

3d

No radiation

Spiridonov


51 vert levels

5 s timestep



No aerosols


120x120x20 km

1x1 km horiz

50 vert levels


Electrical scheme


3d anelastic

MPDATA advec.

Meso-NH

Pinty, Mari

120x120x20 km

1x1 km horiz

0.250 km vert



3d,


DHARMA

Fridlind, Ackerman

120x120x20 km

1x1 km horiz

0.4 km vert



3d pseudo-elastic


C. Wang

160x160x20 km

1x1 km horiz

50 vert levels

10 s timestep



(Lin et al, 83)

No aerosols


No radiation

WRF-AqChem

Barth, Kim


360x328x25 km

2x2 km horiz

0.500 km vert

3-30 s timestep


Gas chemistry on

Lightning-NOx param

Predict Mass only:


No aerosols

3d GCE Model


U. Md.

Pickering, Ott

Stenchikov

120x120x20 km

1x1 km horiz



(Lin et al 83)

3d

No radiation

Spiridonov


51 vert levels

5 s timestep



No aerosols


120x120x20 km

1x1 km horiz

50 vert levels


Electrical scheme


3d anelastic

MPDATA advec.

Meso-NH

Pinty, Mari

120x120x20 km

1x1 km horiz

0.250 km vert



3d,


DHARMA

Fridlind, Ackerman

120x120x20 km

1x1 km horiz

0.4 km vert



3d pseudo-elastic


C. Wang

160x160x20 km

1x1 km horiz

50 vert levels

10 s timestep



(Lin et al, 83)

No aerosols


No radiation

WRF-AqChem

Barth, Kim


360x328x25 km2x2 km horiz0.500 km vert3-30 s timestep

Offline chemistrytransportGas chemistry onLightning-NOx param

Predict Mass only:cw, rain, ice, snow, hail(Tao and Simpson, 1993)No aerosols

3d GCE Model(Tao and Simpson, 93)

U. Md.Pickering, OttStenchikov


Aqueous chemistryPredict Mass only:cw, rain, ice, snow, hail,(Lin et al 83)

3dNo radiation

Spiridonov

120x120x20 km1x1 km horiz51 vert levels5 s timestep


Predict Mass only,cw, rain, ice, snow, hailNo aerosols


120x120x20 km1x1 km horiz50 vert levels

Scavenge soluble speciesElectrical scheme


3d anelasticMPDATA advec.

Meso-NHPinty, Mari

120x120x20 km1x1 km horiz0.250 km vert

No chemistryPredict Number, Mass &sizePrognostic CCN and IN

3d,Interactive radiation

DHARMAFridlind, Ackerman


Chemistry onPredict Number and MassPrognostic CCN and IN

3d pseudo-elasticInteractive radiation

C. Wang


Chemistry onPredict Mass only:cw, rain, ice, snow, hail(Lin et al, 83)No aerosols

3d, flux form Runge-KuttaNo radiation

WRF-AqChemBarth, Kim

ConfigurationChemistryCloud Microphysics andAerosols

Dynamics,thermodynamics,radiation

Model








3dNo radiation

Spiridonov









Meso-NHPinty, Mari








C. Wang







Model








3dNo radiation

Spiridonov









Meso-NHPinty, Mari








C. Wang







Model








3dNo radiation

Spiridonov









Meso-NHPinty, Mari








C. Wang







Model








3dNo radiation

Spiridonov









Meso-NHPinty, Mari








C. Wang







Model








3dNo radiation

Spiridonov









Meso-NHPinty, Mari








C. Wang







Model








3dNo radiation

Spiridonov









Meso-NHPinty, Mari








C. Wang







Model








3dNo radiation

Spiridonov









Meso-NHPinty, Mari








C. Wang







Model








3dNo radiation

Spiridonov









Meso-NHPinty, Mari








C. Wang







Model

360x328x25 km

2x2 km horiz

0.500 km vert

3-30 s timestep


Gas chemistry on

Lightning-NOx param

Predict Mass only:


No aerosols

3d GCE Model


U. Md.

Pickering, Ott

Stenchikov

120x120x20 km

1x1 km horiz



(Lin et al 83)

3d

No radiation

Spiridonov


51 vert levels

5 s timestep



No aerosols


120x120x20 km

1x1 km horiz

50 vert levels


Electrical scheme


3d anelastic

MPDATA advec.

Meso-NH

Pinty, Mari

120x120x20 km

1x1 km horiz

0.250 km vert



3d,


DHARMA

Fridlind, Ackerman

120x120x20 km

1x1 km horiz

0.4 km vert



3d pseudo-elastic


C. Wang

160x160x20 km

1x1 km horiz

50 vert levels

10 s timestep



(Lin et al, 83)

No aerosols


No radiation

WRF-AqChem

Barth, Kim


Mary Barth and Si-Wan Kim (NCAR) – WRF–AqChem sensitivity sim.

Chien Wang (MIT) – C.Wang sensitivity sim.

Ken Pickering and Lesley Ott (U. Maryland), and Georgiy Stenchikov(Rutgers Univ.) – UMd/GCE

Ann Fridlind and Andy Ackerman (NASA/Ames) – DHARMA

Jean-Pierre Pinty and Celine Mari (CNRS--Toulouse) – Meso-NH

Maud Leriche and Sylvie Cautenet, (LaMP, U. B-P, Clermont-Ferrand)–Leriche or RAMS

Vlado Spiridonov, (Hydrometeor. Inst., Macedonia) – Spiridonov

John Helsdon, Richard Farley (South Dakota School M&T) – SDMST

Participants and their Models

Formulation of Models for the Convective-ScaleSimulations

• 3-d, fully compressible, non-hydrostatic– WRF-AqChem – WRF dynamics (flux form)– C.Wang – pseudo-elastic– UMd/GCE – GCE dynamics (anelastic)– DHARMA –Large Eddy Simulation– Meso-NH – anelastic, MPDATA advection– Leriche/RAMS – RAMS dynamics (anelastic)– Spiridonov – based on Klemp and Wilhelmson– SDMST – modified Clark-Hall model, MPDATA

Microphysics and Chemistry

vapor

ice

rain snow

cloudwater

Bulk water method:

qv, qaer, Naer, size of aerosols (16 bins)

qliq, Nliq, size of drops (16 bins)

qice, Nice, size of ice (16 bins)

Sectional method:

vapor

ice

rain snow

cloudwater

Two Moment method:

Nc

Nr Ns

Ni

NCCN, IN

WRF-aqchem, Meso-NH, RAMS,Spiridonov, UMd/GCE, SDSMT

C. Wang

DHARMA

Models Configured for an IdealizedConvective Case

• 160 km x 160 km x 20 km Domain: WRF-AqChem• 120 km x 120 km x 20 km Domain: C.Wang, DHARMA,

Meso-NH, RAMS/Leriche, Spiridonov, SDSMT• 360 km x 328 km x 25 km Domain: UMd/GCE

• Resolution: Δx = Δy = 1 km

stretched vertical grid (50 grid points): WRF-AqChem,Meso-NH, RAMS/Leriche

Δz = 500 m: Spiridonov, UMd/GCEΔz = 400 m: C. WangΔz = 250 m: DHARMA, SDSMT

Sample of the Results Analyzed for theIntercomparison

STERAO-1996From Dye et al. (2000)

Radar Reflectivityz = 10.5 km m.s.l.

t = 1 hr

observations

Model results

Radar Reflectivityalong-axis cross section

t = 1 hr

Maximum Updraft

Transectsacross Anvil

NOx production from Lightning

DeCaria et al. (2005)• Lightning interferometer data as input• Finds region of reflectivity > 20 dBZ• Distributes NO verticallyWRF-AqChem, UMd/GCE

Parameterized flash rate based on max updraftsRAMS/Leriche

Parameterized electric field based on microphysicsC. Wang

Predicts charge density in modelMeso-NH, SDSMT

Transectsacross Anvil

Left side has linearplots, Right side haslog plots

NO, NOxLinear scale

NO, NOxLinear scale

NO, NOxLog scale

NO, NOxLog scale

Cross sectionsCO

50 km from coret = 6000 s

Cross sectionsNOx


Cross sectionsO3


*Skamarock et al.(2003) JGR

Flux Through Anvil Average values through vertical cross-sections from t = 1 h to t = 2 h

6.58 +/- 3.492.24 +/- 0.536.86 +/- 1.30344.3 +/- 133.0avg +/- std dev

13.041.936.59196.9SDSMT(Helsdon et al.)

8.452.549.06274.0U. Md / GCE(Pickering et al.)

3.23.35.00444.0V. Spiridonov

5.302.297.68332.7RAMS(Leriche et al.)

2.841.595.41n/aMeso-NH(Pinty, Mari)

n/a2.397.69531.9DHARMA(Fridlind et al.)

5.971.946.72442.7C. Wang

7.231.946.75187.7WRF-AqChem(Barth, Kim)

5.81.905.9315Observations*

NOx Flux(10-8 mol m-2 s)

CO Flux(10-5 mol m-2 s-1)

Mass Flux(kg m-2 s-1)

Anvil Area(106 m2)

Model

Simulations of HNO3, H2O2, and CH2O

Initial profile

MOZART profile

Points from aircraftobservations

Chemistry and Aerosols

• CO, O3, NOx, H2O2, CH2O, HNO3

Chemistry simulated: WRF-AqChem, C. Wang,RAMS/Leriche, UMd/GCE

• Aerosols simulated:C. Wang

Microphysics and Chemistry

Species are transferredamong hydrometeorsaccording to the microphysics

Liquid to ice, snow, or hail: WRF-AqChem, C. Wang

Liquid to gas: UMd/GCE*, RAMS

vapor

ice

rain snow

cloudwater

Observations

WRF-AqChem(NCAR)

Chien Wang

U.Md/GCE

RAMS/Leriche

Transects acrossAnvil

Gas-phase mixing ratios

Observations

WRF-AqChem(NCAR)

Chien Wang

U.Md/GCE

RAMS/Leriche

No lightning – dotted lines

Liquid to ice, snow, or hail: WRF-AqChem, C. Wang

Liquid to gas: UMd/GCE*, RAMS

degas

adsorption

Transects acrossAnvil

Gas-phase mixing ratios

Observations

WRF-AqChem(NCAR)

Chien Wang

U.Md/GCE

RAMS/Leriche

Microphysical effects

Degassing during dropfreezing

No adsorption of gasesonto ice

Conclusions

• Tracer transport (CO and O3) are similaramong models and similar to observations.

• NOx is consistently underestimated when nolightning is included.

• Lightning-NOx parameterizations performreasonably well.

• Comparison of soluble species HNO3, H2O2,and CH2O shows we have much more toevaluate.

What’s next?

Observations Measurements of HOx precursors in both

the inflow air and the convective outflow arelacking.

Intercomparison of tropical convection andchemistry? Impact of aerosols?

Planning: Deep Convective Clouds andChemistry (DC3) Field Experiment

When: Summer 2009

Where: Central U.S.

Major Facilities:Aircraft (high altitude and low altitudeplanes),Radar (Doppler and polarimetric)Lightning Mapping Array

Initialization

Convection initiatedwith 3 warm bubbles

Sounding data camefrom Skamarock et al.(2000)

Initialization of Chemical Species

flux) mass air for 1 ( species of ratio mixing C

sectioncrossincellgridoflengthhorizontalwhere

==

!="

""

""

=#

# $

l

l

l

sanvil cell

cellsanvilflux

z

zCU%

Calculation is done on grid cells that contain cloud particles.The area of the anvil is the denominator.

!

z

x

y

Δx

l!

Δz

Flux Calculation

transport

washout and rainout

NO production from lightning

Processes in deep convection that affectchemical species

high photolysis rates

low photolysis rates

cloudchemistry

icechemistry

Phase of cloud particles¬cloud microphysics and

chemical species

Spatial Scales and Time Scales

1 s 100 s 1 hr 1 day 1 week 1 mon

Synoptic-scalestorms

1 yr

1 m

100 m

1 km

100 km

1,000 km

Short-lived speciesCH3OO

HO2

NO3OH

Moderately long-lived speciesCO

O3

SO2 H2O2

NOx

CH2O C3H6

C5H8

10 km

Fair weathercumulus

Cumulonimbus

Adapted from Brasseur et al. (1999)

Documents

What We Can Learn from the Midlatitudes – Intercomparison ... talks/Barth.pdf · What We Can Learn from the Midlatitudes – Intercomparison Studies Mary C. Barth Mesoscale and