George Kallos

With contribution from

S. Solomos, I. Kushta, M. Astitha, C. Spyrou, I. Pytharoulis

kallos@mg.uoa.gr

Key Processes in Regional Atmospheric Modeling in the Mediterranean Region

UNIVERSITY OF ATHENSSCHOOL OF PHYSICS, DIVISION OF ENVIRONMENT AND METEOROLOGY

ATMOSPHERIC MODELING AND WEATHER FORECASTING GROUP

http://forecast.uoa.gr

Physiographic characteristics are partially responsible for the formation of particular climatic conditions in the Mediterranean Region.

Regional climatic patterns are defined as a result of balancing between large scale flow and mesoscale circulations.

The resulted circulation has a general trend from North to South (pressure gradient and differential heating between land and water).

Air masses reaching the Mediterranean region are not clearly defined as pure maritime or continental because their characteristics are modified relatively fast.

The air masses in the area have a mixture of natural and anthropogenic origin aerosols with varying optical and hygroscopic properties.

Therefore, cloud formation is a complicated process.

MOTIVATION

Desert dust and sea salt are major sources of PM in the atmosphere.

Their impacts in the atmosphere are many and of course the feedbacks are considerable.

The impacts are ranging from modification of the radiative forcing to cloud formation and precipitation.

Therefore, perturbations in dust and/or sea salt particle production can have impacts on radiative properties, cloud formation and water budget.

These links are not one way but there are feedbacks that are critical for both meteorological and climatological – scale phenomena.

The links and feedbacks become more complicated because of the coexistence of anthropogenically-produced aerosols and chemical transformations.

MOTIVATION

Discuss the:

Long range transport of naturally (mainly Saharan dust and

sea-salt) and anthropogenically - produced aerosols.

Aerosol-cloud interaction processes

Other processes needed to be taken in the account

New modeling tools

OBJECTIVES

A

CB

M1

M2

M3

Saharan Dust Transport

I TCZ

Upper layer transport

D

A

CB

M1

M2

M3

Saharan Dust Transport

I TCZ

Upper layer transport

D

Kallos et al. 2007

JAMC, 46(8), pp. 1230–1251.

SUMMER SEASON

TRANSITION SEASONS

Transport ofSaharan Dust

Transport of Anthropogenic Pollutants

Transport of3rd generation species is a combination of the two transport paths

Transport ofSaharan Dust

Transport of Anthropogenic Pollutants

Transport of3rd generation species is a combination of the two transport paths

Paths and scales of transport

STRATOSPHERE

LOWER TROPOSPHERE

SO2 HCl

ash SO2 H2SO4

precipitation

warming

hv & OH

cooling of the surface

LW radiation

Cloud Modifications

NucleationDeposition

Dust NaCl

UPPER

TROPOSPHERE

Heterogeneous reactions

SO2 + O3 DSO4 (dust+sulfate)

NO2 DNO3 (dust+nitrate)

HNO3 DNO3 (dust+nitrate)

dust

dust

dust

PM AND GASEOUS POLLUTANTS IN THE ATMOSPHERE

PROCESSES AFFECTING AIR QUALITY AND CLIMATE

incoming-outgoing SW radiation

AEROSOLS AND CLOUDS

Aerosols in general act as CCN and IN in the cloud formation processes according to their physicochemical properties.

As it is known, high concentrations of aerosols result in haze and small-cloud droplet formation that do not necessarily lead in precipitation.

Although, this is not always the case.

Under certain conditions and mixture of naturally and anthropogenically-produced aerosols, gigantic CCNs and/or IN formation with enhanced hygroscopicity may lead in stormy weather conditions (Levin et al., 2006; Rosenfeld et al., 2008)

activation

drop growth (collision, coalescence)

Condensates (cloud,rain,ice,graupel,hail,

pristine ice,aggregates)

Natural particle emissions – transport

Aerosol

AEROSOLS AND CLOUDS

SEM analysis

Izaña Sta. Cruz de Tenerife

(Alastuey et al. 2005)

clay particles (around 1 µm)

clay particles(>10 µm)

general aspect of dust rounded quartz

fresh water diatom spore particles

natural marine aerosolsspongy carbonaceous anthropogenic particles

typical large and plate natural gypsum particles

potassium sulphate particles

covering clay aggregates

<5 µm Ca sulphates with crystalline habit

COEXISTENCE OF MINERAL DUST, SULFATES AND CLOUD DROPLETS

A photomicrograph of drops and aerosol particles collected inside clouds

D: drops,

P: dry interstitial particles

(Levin et al. 1996, 2006)

A photomicrograph of desert mineral dust with small sulfate particles on its surface.

COEXISTENCE OF MINERAL DUST, SULFATES AND CLOUD DROPLETS

NATURALEMISSIONS

AEROSOLS

DUST

SEA SALT (NaCl)

PARTICLES

COVERED

WITH GASES OR

AEROSOLS

ANTHROPOGENIC EMISSIONS

GASES

SO2, NO, NO2, VOC

CO, NH3, CO2

PARTICULATEMATTER

PM2.5PM10

SECONDARY POLLUTANTS

O3, SO4, NO3, NH4

2nd GENERATION POLLUTANTS

1st GENERATION POLLUTANTS

3rd GENERATION POLLUTANTS

SAHARAN DUST AND ANTHROPOGENIC POLLUTANTS

FORMATION OF VARIOUS GENERATIONS OF PM

Mechanical Processes

Molecular Processes

DUST

+

SEA SALT

+

SODIUM AEROSOL

+

CHLORIDE AEROSOL

sedimentation

rain

Chemical transformation of gases, coalescence,

condensation, homogeneous nucleation

SULFATES NITRATES

FINE PARTICLES COARSE PARTICLES

0.001 0.01 0.1 1 10 100

PARTICLE DIAMETER (μm)

SULFATES FORMED ON

DUST

NITRATES FORMED ON

DUST

Interaction

MASS-SIZE DISTRIBUTION OF PM

Aerosol-Cloud Interaction (Rosenfeld, 2008)

Aerosol effects on precipitation (Rosenfeld, 2008)

Aerosol concentration

Dry unstable situation

Mar

itime &

mod

erate

(wet)

conti

nenta

l clou

ds

Acc

umul

ated

rai

n

Precipitation

GAIN

(condensation+deposition)Precip = -

LOSS

(evaporation+sublimation)

Precipitation is often a small difference of large values

Aerosols affect both generation and loss of hydrometeor mass

Aerosol effects on precipitation

WHY WE NEED NEW MODELING TOOLS?

To study such complicated processes there is a need for advanced modeling tools that include physical and chemical properties directly coupled.

The Integrated Community Limited Area Modeling System – ICLAMS – has

been developed (still under development) at the framework of CIRCE

project (RL8)

It has been designed to be able to simulate links and feedbacks between

air quality and regional climate

ICLAMS development is on RAMS.6 modeling system

RAMS is a multi-scale modeling system and can be configured to run with

resolution from a few meters to tens of kilometers on a two-way interactive

nesting mode

It has detailed cloud microphysical scheme with 8 microphysical

categories, detailed surface parameterization and many other features that

make it the most advanced limited area meteorological model

The cloud microphysical scheme includes prognostic equations for mass

mixing ratios of the various forms of water species

ICLAMS

The new model development includes the following features:

Full dust cycle module following the formulation used in SKIRON/Dust with 8 dust

particle bins following lognormal distribution Kallos et al., (2005, 2007).

Sea salt production mechanism with 2 size bins following Gong et al., (1999).

Gas and aqueous phase chemistry (SAPRC mechanism as implemented in CAMx).

Gas to particle conversion and heterogeneous chemistry following the ISOROPIA

scheme and additional interactions with desert dust, sea salt and sulfates.

Impacts of aerosols and PMs on radiative transfer of the photochemically active

bands.

Visible and Infrared corrections due to aerosols and PMs.

Treatment of CCN and GCCN as predictive quantities (4-D).

All these new elements are directly coupled and executed together with the

meteorological modules.

ICLAMS

BASIC MODULES OF ICLAMS

The development is carried out on the RAMS ver. 6

It has two-way interactive nesting capabilities Explicit cloud microphysical scheme It can be nested inside of global systems It can run with configurations from a few meters horizontal resolution up to

hemispheric

It also includes:

Detailed soil surface and water interaction processes Detailed dust cycle description Detailed sea salt cycle Gas phase chemistry module with photochemical processes Aqueous phase chemistry Gas to particle conversion and heterogeneous chemical reactions Dry and wet deposition modules Aerosol-cloud-radiative transfer interaction module

It can be used mainly for case studies and scenario development related to aerosol cloud interaction

The modeling system handles dust, sea salt and three-generation particle formation on a direct way.

CCN, GCCN and IN are predictive quantities. Links and feedbacks between the direct and indirect aerosol effects are described.

Currently the following model components have been developed and tested :

Cloud effects: Prognostic CCN calculation based on aerosol number density and chemical composition. Three different approaches are used.

1. All fine aerosols and 33% coarse are efficient CCN (Levin et al., 2005)2. A parameterization based on Koehler theory (Nenes et al., 2007) is used to

calculate the number of natural particles that can be active CCNs and the number of cloud droplets nucleated based on the chemical composition of the aerosols, their size (lognormal distribution) and ambient conditions (temperature, updraft velocity).

3. Cloud droplets spectrum is obtained by previously (offline) generated lookup tables from detailed parcel-bin model as a function of CCN and GCCN number concentration, updraft velocity and temperature. (Cotton et al., 2007).

Prognostic cloud droplets spectrum is then used by the explicit microphysical scheme of RAMS (Meyers et al. 1997) to forecast the rest of the microphysical species (rain, pristine ice, snow, aggregates, graupel, hail) and precipitation.

Radiative transfer : Calculation of the heating rates based on aerosol load and type. Two different options are available.

1. Harrington radiation scheme 2. Rapid Radiative Transfer Model (RRTM)

ICLAMS Aerosol – Meteorology feedback

MEIDEX

experiment

MEIDEX TEST CASE

On 28 January 2003, a dust storm passed over the north east Mediterranean region.

On 29 JAN 2003 heavy rain and hail dispersed over the Middle East coastline and a few

km inland.

Flood events and agricultural disasters were reported.

Airborne measurements of this episode were obtained during the

Mediterranean Israeli Dust Experiment (MEIDEX)

Two cases have been considered:

1st case: Natural particles are treated as passive tracers. User specified constant CCN #.

2nd case: Dust plume and sea-salt sprayed interact with clouds. CCN and GCCN prognostic (from dust and sea salt concentrations and size distribution.

MODEL SETUP

RAMS6.0 with:

DUST MODULE (Zender at al., Marticorena and Bergameti)

8 Bin lognormal dust particles distribution - Dust cycle

SEASALT MODULE (Gong et al.)

2 Bin lognormal salt particles distribution - Seasalt cycle

DOMAIN SETUP

3 grids (36km-12km-4km) , 31 vertical levels, 120 hours run

Initial and boundary conditions – NCEP 1deg GFS analysis data

REFERENCE RUN (1st case)

ICLOUD=5 (Constant # of CCN)

Dust and Salt particles do not interact with the rest of the model

TEST RUN (2nd case)

ICLOUD=7 (3-D prognostic CCN and GCCN field)

Particles serve as efficient cloud condensation nuclei (CCN).

On 28 Jan 2003 a cold cyclone moved from Crete

through Cyprus accompanied by a cold front .

A second air mass transported dust particles from

NE Africa over the sea towards Israeli coastline.

Wind speed 28JAN2003 1200UTC

Dust load 28JAN2003 0900 UTC3h accumulated precipitation 28JAN2003 1200

UTC

LARGE SCALE FEAUTURES

DUST CONCENTRATION at 1500UTC (all bins)

27JAN2003 1500 UTC 28JAN2003 1500 UTC 29JAN2003 1500 UTC

1071 m6th model level

2087 m9th model level

3440 m12th model level

SEA SALT CONCENTRATIONS at 1500 UTC (both bins)

27JAN2003 28JAN2003 29JAN2003

97 m

2nd model level

530 m

4th model level

1071 m

6th model level

Case 1 - CONSTANT CCN (ICLOUD=5)

1hr accumulated precip. 29JAN2003 06 UTC 1hr accumulated precip. 29JAN2003 10 UTC 1hr accumulated precip. 29JAN2003 13 UTC 1hr accumulated precip. 29JAN2003 18 UTC

Case 2 - PROGNOSTIC CCN (ICLOUD=7)

1hr accumulated precip. 29JAN2003 06 UTC 1hr accumulated precip. 29JAN2003 10 UTC 1hr accumulated precip. 29JAN2003 13 UTC 1hr accumulated precip. 29JAN2003 18 UTC

PRECIPITATION PATTERN

3hr accumulated hail 29JAN2003 06 UTC 3hr accumulated hail 29JAN2003 21 UTC3hr accumulated hail 29JAN2003 15 UTC3hr accumulated hail 29JAN2003 09 UTC

Case 2 - PROGNOSTIC CCN (ICLOUD=7)

Case 1 - CONSTANT CCN (ICLOUD=5)

3hr accumulated hail 29JAN2003 06 UTC 3hr accumulated hail 29JAN2003 09 UTC 3hr accumulated hail 29JAN2003 15 UTC 3hr accumulated hail 29JAN2003 21 UTC

HAIL PATTERN

11:27 11:35 11:39 11:54 12:02 12:13 12:23 12:30 12:37 12:47 12:55 13:19 13:30 13:36 13:40

0

250

500

750

1000

1250

1500

1750

2000

Total Particles concentration (#/cm-3)

11:27 11:35 11:39 11:54 12:02 12:13 12:23 12:30 12:37 12:47 12:55 13:19 13:30 13:36 13:40

0250500750

1000125015001750200022502500275030003250

Time in (LT) Time in(UTC) height(m) Mean(Latitude) Mean(Longitude) Total concentration (cm-3)

9:27 11:27 3048 31.98 34.06 758.54

9:35 11:35 2590.8 32.02 33.67 640.54

9:39 11:39 1981.2 32.09 33.47 601.53

9:54 11:54 762 31.89 33.92 646.18

10:02 12:02 457.2 31.7 34.28 760.48

10:13 12:13 1798.32 31.81 34.14 639.53

10:23 12:23 1371.6 31.95 33.78 784.65

10:30 12:30 762 32.07 33.52 886.1

10:37 12:37 762 32.2 33.25 1414.86

10:47 12:47 457.2 32.27 33.18 1907.84

10:55 12:55 152.4 32.12 33.53 967.59

11:19 13:19 457.2 31.94 33.98 680.34

11:30 13:30 914 32.35 34 922.83 below cloud

11:36 13:36 1371.6 32.39 34.03 1082.56 in cloud

11:40 13:40 2286 32.32 34.09 323.15

Total Particles have been

measured at 15 locations along

the aircraft flight (11:27–13:40

UTC 28JAN2003)

aircraft altitude

particle concentration

MEIDEX OBSERVATIONS (Levin et al.)

For each one of the 15 measuring locations of MEIDEX experiment a time /

height plot of modelled total particles (dust & salt) concentration is created.

Maximum particles concentration during the flight period (10:00 – 13:00 UTC)

are ranging from 1000 #/cm3 near the ground to below 50#/cm3 above 3 Km.

These numbers compare well with the aircraft observations.

MEIDEX

experimental flight

ICLAMS PARTICLE CONCENTRATIONS

During the dust storm simulation the concentration profile

for coarse particles (d>1 μm) ranged from 40 (#/cm3)

near the ground to less than 5(#/cm3) above 2km

The simulated concentration for fine particles (d<1

μm) ranged from 1000 (#/cm3) near the ground to less

than 100 above 3km.

Fine particles – ICLAMS - 28JAN2003

Coarse particles – ICLAMS - 28JAN2003

COARSE & FINE PARTICLES

2nd CASE: Aerosols serve as efficient CCN

For simplicity we assume that all particles produced (dust and sea salt) are

efficient CCN.

However the CCN field will not retain the lognormal characteristics of

aerosol size distribution.

Dust and sea salt - born CCN will be added on the background value

(400 #/cm3) in order to enhance the effect of the dust storm in cloud

processes.

The rest of the RAMS microphysics scheme remains unchanged.

Constant CCN field

CCN concentration (#/cm3) CCN concentration 28JAN2003 1700 UTC CCN concentration 29JAN2003 0500UTC

CCN concentration 29JAN2003 1600UTC CCN concentration 30JAN2003 0000UTC

Prognostic 3D CCN field

ICLAMS CCN VERTICAL DISTRIBUTION

Constant CCN field

prognostic CCN field

Cloud concen. (#/cm3) 28JAN2003 16 UTC Cloud concen. (#/cm3) 29JAN2003 11 UTC

Cloud concen. (#/cm3) 28JAN2003 16 UTC Cloud concen. (#/cm3) 29JAN2003 11 UTC

CLOUD CONCENTRATION VERTICAL CROSSECTION

Prognostic CCN field

constant CCN field Total condensates (#/cm3) 29JAN2003 12

UTC

Total condensates (#/cm3) 29JAN2003 15

UTC

Total condensates (#/cm3) 29JAN2003 12

UTC

Total condensates (#/cm3) 29JAN2003 15

UTC

TOTAL CONDENSATES VERTICAL CROSSECTION

Prognostic CCN field

Constant CCN field

Ice mix ratio (gr/kgr) 28JAN2003 18 UTC Ice mix ratio (gr/kgr) 28JAN2003 21 UTC

Ice mix ratio (gr/kgr) 28JAN2003 18

UTC

Ice mix ratio (gr/kgr) 28JAN2003 21 UTC

ICE MIXING RATIO VERTICAL CROSSECTION

Maxim um updraft ve locity

0.00

1.00

2.00

3.00

4.00

5.00

6.00

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

28JAN2003 06 UTC - 28JAN2003 07 UTC

w (

m/s

ec)

ICLOUD7 w(m/s)

ICLOUD5 w (m/s)

MAXIMUM UPDRAFT VELOCITY

Significant delay on the updraft velocity maximum during cloud development and

also increase in maximum value in 2nd mode run (yellow line )

CCN Condensation Latent heat Updrafts

maximum updraft speed (m/sec)

0

2

4

6

8

10

12

1 2 3 4 5 6 7 8 9 10 11 12 13 14

cloud evolution time

w (

m/s

ec)

Constant CCN field

all particles are effective CCN

all f ine particles and 33%coarse particles are effective CCN

MAXIMUM UPDRAFT VELOCITY

Feature Constant CCN field All particles are

effective CCN

All fine particles and

33% from the coarse

are effective CCN

Max updraft 9.27 10.26 10.63

Time and height 10:00 UTC at 4km 13:00 UTC at 4km 14:00 UTC at 4km

OTHER PROCESSES THAT NEED SPECIAL ATTENTION

Surface properties and energy partitioning:

• SST

• Soil thermophysical and hydraulic properties

How useful is the high resolution SST on regional scale modeling in the Mediterranean Region?

Energy partitioning on the sea surface

What is the role on soil texture on the accuracy of the models?

SEP 2004

OCT 2004

NOV 2004

HiRes SSTs (1/16 deg) Coarse SSTs (1/2 deg

T+24Front over the Ionian Sea

RED: 1/16x 1/16 SSTs BLUE: 0.5x0.5 SSTs

2-m Temperature and 6-hr accumulated precipitation

Sea-surface energy partitioning

The SKIRON/Eta model calculates the surface parameters using a viscous sublayer scheme (Janjic 1994, MWR)

• The viscous sublayer over the ocean is assumed to operate in 3 regimes:

(i) smooth and transitional, (ii) rough, (iii) rough with spray,

depending on the Reynolds number which is a function of u*

Mean Differences in Heat Fluxes in Jan 2003

(VISCOUSyes-VISCOUSno)

Stronger latent and sensible heat fluxes over the water without the use of the viscous sublayer.

Higher precipitation amounts and stronger winds over the water without the use of the viscous sublayer.

This is in agreement with theory

Soil Characteristics - Thermophysical and Hydraulic Properties

Effects of Soil Thermophysical Properties on Saharan ABL

RMSE Rocky soil Sand

Α 2.07 4.02Β 4.07 5.31C 2.57 3.75D 2.82 4.07

AVG 2.98 4.34

Dust, sea salt and anthropogenically-produced aerosols are key players in radiation, cloud and precipitation formation.

The links and feedbacks are important on defining forcing and therefore studying regional climate.

In the new system - ICLAMS - such capabilities have been implemented:

Cloud formation through the treatment of CCN, GCCN and IN as predictive quantities.

Aerosol-cloud-radiation interactions. Treatment of naturally and anthropogenically-produced particles with

different properties. Simulation of cloud-scale phenomena.

The sensitivity tests carried out clearly showed that the CCN originated mainly from desert dust.

The storm generation and evolution is highly related to spatiotemporal distribution of CCNs.

In general, high concentrations of natural particles tend to delay rainfall but can have higher amounts of precipitation in certain locations.

More development (production of CCG, GCCN and IN from anthropogenic sources) and testing efforts are under way at the framework of CIRCE project.

Some Concluding Remarks

This work is funded by European Union FP6 project CIRCE

ACKNOWLEDGEMENTS