George Kallos With contribution from S. Solomos, I. Kushta, M. Astitha, C. Spyrou, I. Pytharoulis...
Transcript of George Kallos With contribution from S. Solomos, I. Kushta, M. Astitha, C. Spyrou, I. Pytharoulis...
George Kallos
With contribution from
S. Solomos, I. Kushta, M. Astitha, C. Spyrou, I. Pytharoulis
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