Http:// SOLAS Dust workshop (Reading) Overview of dust modelling from Leeds global aerosol group Graham Mann, Ken Carslaw, Dominick Spracklen,

http://www.ncas.ac.uk

SOLAS Dust workshop (Reading)

Overview of dust modelling from Leeds global aerosol group

Graham Mann, Ken Carslaw, Dominick Spracklen, Hannele Korhonen, Joonas Merikanto, Maria FrontosoPaul Manktelow, Dave Ridley, Matt Woodhouse,

Zongbo Shi, Mike Krom, Lianne Benning

Institute for Climate & Atmospheric Science,School of Earth & Environment

University of Leeds.



Several community global aerosol models developed at Leeds (soon available)

TOMCAT global CTM

GLOMAP-binaerosol module

(sectional aerosol)Unified ModelHadGEM climate modelQUEST Earth System Model

Observations Field campaigns

APPRAISE Community Model

Climate Prediction

GLOMAP-bin (no. & mass conc. in size bins) The most complete aerosol scheme that can be run globally

GLOMAP-mode (no. & mass conc. in log-normal modes) A reduced GLOMAP, cheap enough to run in the UKCA composition-climate model & coupled to Earth System.

The Global Model of Aerosol Processes (GLOMAP) was developed in Leeds to model aerosol microphysics & chemistry with detailed size-resolved composition.

2 aerosol schemes

GLOMAP-modeaerosol module(modal aerosol)

or

GLOMAP-modeaerosol module(modal aerosol)

UKCA Community Model

Schemes in 3 model environments

1) chemical transport model (TOMCAT)

2) nested to particular region in CTM

3) coupled to climate in GCM (UM)



Number size distributions from GLOMAP-bin & GLOMAP-mode (annual mean)

GLOMAP-modeGLOMAP-bin

Arctic N. Atlantic Sub-tropics

TropicsSouthern Ocean Antarctic

Modal scheme (as implemented into UKCA) matchesthe number size distributions from the more detailed GLOMAP-bin scheme remarkably well.

Some discrepancies in coarse mode mode overestimates(sedimentation?).



GLOMAP-bin & -mode being tested against 20 years of published observations

Spracklen et al. (2005) “microphysical models…to understand the factors that control the number of CCN at a given supersaturation”



GLOMAP CCN against in-situ observations

GLOMAP-bin GLOMAP-mode

Both models compare fairly well against observations but both tend to underestimate at high CCN numbers new particle formation in BL not included.



Dust emissions scheme used in GLOMAP include in UKCA

Size distribution of dust emissions is a key uncertainty in dust modelling and is critical for the climate impact of dust.

Smallest dust particles (<1m) have long lifetimes and are the most important particles with respect to climate.

Have implemented scheme of Alfaro & Gomez (2001) which represents the physics of the sand-blasting process.

Experiments in a wind tunnel (Alfaro et al, 1998) showed that saltating aggregates impacting on surface dust release smaller sized particles only in stronger wind events

mode 1 (14.2 m mass median diameter) mode 2 (6.7 m mass median diameter) mode 3 (1.5 m mass median diameter)

Modes only released once binding energies exceededby energy of saltating aggregates.Small modes have larger binding energies (more energy required)



GLOMAP-bin (A&G) dust against Univ. Miami observations

mass (gm-3)Annual mean surface dust concentration

Global surface fields of soil moisture, leaf area index, snow/ice cover & preferential source areas [paleolake beds] from Tegen et al (2002) determine horizontal saltation flux

ISLSCP global soil texture dataset drives size distribution of saltating aggregates

Influence of particle size and soil moisture on threshold friction velocity following Woodward (2001)

From Manktelow PhD thesis.



GLOMAP-bin (A&G) mass conc. vs observations

Overestimates at some remote sites

Underestimates near source

Annual mean

Modelled spatial dust mass distribution in reasonable agreement with observations(log correlation coefficient of 0.8). Captures timing of Asian dust maximum (not shown).




GLOMAP 2-mode vs 7-bin representation of dust transport

GLOMAP-mode run using 2 size modes (4 tracers)GLOMAP-bin run using 7 dust bins (14 tracers)

Here both use prescribed daily dust emission fluxes for AEROCOM by Paul Ginoux (GFDL)

Use GLOMAP-mode and –bin driven by Alfaro & Gomez sandblasting scheme for predicting dust deposition to ocean over several years.Collaboration with UEA (Le Quere, Buitenhuis) to model ecosystem response

Shift in size distribution examined at 8 sites going out from Sahara over ocean.



Influence of dust on sulphate aerosol

Some studies (e.g. Bauer & Koch, 2005) have found large enhancements of sulphate due to heterogeneous oxidation of SO2 on dust.

Despite high concentrations of dust in ACE-Asia “Perfect Dust Storm”, GLOMAP-bin sulphate still not significantly enhanced due to quenchingof sites for oxidation once dust particles are aged (Manktelow, PhD thesis).

However, mixing of dust with pollution affects particle lifetime, hygroscopicity & solubility

SO4 (gm-3) % increase



Size-dependence of dust deposition processes also important

Need to test deposition of GLOMAP modal dust and coarse bin schemes to see how they perform against bin-resolved schemes with large number of bins.



GLOMAP dust deposition to ocean using A&G vs DIRTMAP

Generally encouraging spatial distribution of dust deposition vs observations

Sandblasting scheme seems to give too short a lifetime too low deposition.

Not enough smaller particles with longer lifetimes but this could be enhanced by sub-grid wind speed variability causing more smaller particles to be ejected

Marine sediment trap measurements of dust deposition as compiled by Tegen et al (2002)

Deposition to ocean provides nutrients to ocean phytoplankton

Dust deposition (gm-2yr-1)




SOLAS KT project on dust deposition & DMS emissions

• DMS production in “high-nutrient low-chlorophyl” areas of the ocean can be stimulated by dust-deposition as a result of Fe limitation.

• Consequent increase in CCN and uptake of CO2

Uptake of acids and cycling through clouds affect Fe solubility & bio-availibility

Matt Woodhouse, Graham Mann, Ken Carslaw (University of Leeds)



GLOMAP-mode 2xDMS patch experiments show CCN impact is not local.

CCN change <2% in perturbed patch

CCN changes seen far from patch



SOLAS project – factors controlling dust bioavailability

• When rocks containing iron minerals weather under desert conditions they tend to produce chemically refractory iron oxides (haemetite and goethite).

• When dust is taken up into the atmosphere it interacts with acid in clouds.

• The cloud processes may convert the refractory iron into dissolved Fe.

• When the dissolved Fe reprecipitates we suggest it does so as chemically reactive Fe nanoparticles

Mike Krom, Zongbo Shi, Lianne Benning, Ken Carslaw, Graham Mann, Alex Baker, Tim Jickells

(University of Leeds and University of East Anglia)



What we are going to do?

● Carry out controlled lab experiments using realistic simulated atmospheric conditions to determine the fraction of unreactive Fe converted to Fe nanoparticles.

● Examine dust samples collected in SOLAS cruises using modern microscopic and spectroscopic techniques with conventional methods and relate this to measures of Fe solubility.

● Use a global model of aerosol processes from UKCA which will be evaluated against the available and new/ongoing understandings of processes.

● Understand the contribution of physical size sorting and chemical processing of dust particles towards changing the solubility & bioavailability of iron



Summary

Range of environmental applications involving global modelling of dust aerosol to be carried out at Leeds with GLOMAP, UKCA and QUEST models.

Climate impacts of dust via direct radiative effect (UKCA)

Impact on aerosol & climate feedbacks from enhanced DMS by large-scale dust deposition (SOLAS KT PhD)

Impact of dust deposition to oceans on marine ecosystem (QUEST ESM & collaboration with UEA)

Chemical transformation of dust during transport and impact on bioavailibility to ocean ecosystem (SOLAS)

Impact of dust on sulphate aerosol properties and interaction with other aerosol components



The influence of DMS on CCN concentrations at Cape Grim

Around 50% of CCN at Cape Grim in SH winter caused by sulfate particles from enhanced new particle formation as a result of DMS from phytoplankton (Korhonen et al., in preparation)

ObservedModel

Model, no DMS



“Standard” UKCA Aerosol Scheme

Mode name Mean rad range nm

Composition Production Comments

nucleation soluble

< 5 SO4 nucleation Currently only binary H2SO4-H2O, later BLN

Aitken insoluble

5 – 50 BC, OC primary BC/OC emissions Separation to handle BC/OC ageing (necessary if timescale for ageing > lifetime in grid box)

Aitken soluble

5 – 50 SO4, BC, OC growth of nucl. soluble, condensation (sol,ins Ait), primary SO4 ems, coag.

Accum. insoluble

50 – 500 DU primary DU emissions Separation to handle dust ageing (as above

Primary accum. mode BC/OC ageing?

Accum. soluble

50 – 500 SO4, SS, DU, BC, OC

growth of Aitken soluble, condensation (sol,ins acc) primary SS ems, coag.

Coarse insoluble

>500 DU primary DU emissions Separation to handle dust ageing

Coarse soluble

>500 SO4, SS, DU, BC?, OC?

growth of accum soluble, primary SS emission

Equates to 23-25 advected aerosol tracers



Observed and modelled nucleation events

Observed(using 2 DMPS’s)

GLOMAP-bin

modelled(using BL nucleation param’n)

See Spracklen et al (2006)

GLOMAP simulates observed nucleation bursts at Hyytiala remarkably well given that it is from a global model run.



Impact of including BL nucleation on CCN concentrations.

Absolute change

Baseline CCN

We find that BL nucleation has a significant impact on regional CCN,

Spracklen et al.(2006) Spracklen et al (submitted 2007)



GLOMAP & UKCA simulate internal mixtures and ageing of aerosol

SO4

BC

OC

Sea saltCoagulation

CCN

Primary BC/OC distribution (assumed insoluble)

SOA & H2SO4

Condensation of

“Ageing” 1 monolayer

Two simulations with and without “condensation” ageing

Condensation and chemical reaction

Insoluble More solublePartially soluble



Impact of particle ageing on model CCN against observations

BC/OC condensation agedBC/OC not condensation aged

S Africa, SAFARI-2000

Probably not ‘aged’

Korea, Germany, Amazon

Aged particles can act as CCN

Particle chemical ageing is fundamentally important for climate change (it can change CCN by a factor 10)



Changes in global & regional sulfate aerosol since the 1980s.

Considering annual regional burdens, find:

1% reduction in European SO2 emissions. 0.65% decrease in regional SO4 burden

1% increase in E. Asian SO2 emissions 0.88% increase in regional SO4 burden

Stronger oxidant limitation in-cloud over Europe & N. America compared to S. Asia & E. Asia means 12% decrease in global SO2 emissions resulted in only a 3% decrease in global SO4 aerosol burden.

-4mgS m-2-10mgS m-2

+6mgS m-2-2mgS m-2

-4mgS m-2

+5mgS m-2

see Manktelow et al (2007)

SO2 emissions

SO4 production

dryoxdn

wetoxdn



Summary & conclusions

A number of global aerosol models based on GLOMAP have been developed at Leeds & will soon be available as community models in CTM (offline) and GCM (coupled) frameworks.

GLOMAP-bin is a detailed bin-resolved model for short timescales.

GLOMAP-mode is a reduced version for longer integrations and for GCM simulations coupled to atmospheric dynamics and clouds.

The UKCA sub-model now has GLOMAP-mode incorporated in the UM environment (although not yet coupled to radiation & clouds).

A nested version of GLOMAP will enable high-resolution simulations of regional aerosol including long-range transport and other important global impacts (e.g. FT nucleation).

The models are being used for a variety of applications including new particle formation, biomass burning, air quality, volcanic aerosol, dust-chemistry & DMS-CCN Earth System interactions.



CN & CCN from GLOMAP-bin & GLOMAP-mode (December 2000)

GLOMAP-bin surface CN (cm-3) GLOMAP-bin surface CCN (cm-3)

GLOMAP-mode surface CCN (cm-3) GLOMAP-mode surface CN (cm-3)



Number size distributions from GLOMAP-bin & GLOMAP-mode (December 2000)

GLOMAP-mode

GLOMAP-binModal scheme (as implemented into UKCA) reproduces the number size distributions from the more detailed GLOMAP-bin scheme quite well.

Some discrepancies in nucleation & Aitken modes.

Arctic N. Atlantic Sub-tropics

TropicsSouthern OceanAntarctic

Documents

Http:// SOLAS Dust workshop (Reading) Overview of dust modelling from Leeds global aerosol group Graham Mann, Ken Carslaw, Dominick Spracklen,