Climate Impacts of Aerosols with Emphasis on Cloud Modification

1
Climate Impacts of Aerosols with Emphasis on
Cloud Modification
Evidence of Mineral Dust Altering Cloud
Microphysics and Precipitation

• Vernon Morris1, Qilong Min2, Rui Li2, Bing Lin3,
  Angelina Amadou1,
• Everette Joseph1, Yong Hu3, Shuyu Wang2
• 1NOAA Center for Atmospheric Sciences (NCAS)
  Howard University
• 2Atmospheric Science Research Center, State
  University of New York3Science Directories, NASA
  Langley Research Center

2
Outline of Presentation

• Justification of Study/Overview of Issues
• AEROSE Cruises
• Laboratory Simulations of Cloud Microphysics
• Use of Multi-satellite/Multi-sensor Observations
  for Understanding Cloud Microphysics

3
Indirect aerosol effect (I)
Few aerosols Low droplet concentration Less
reflective cloud
Numerous aerosols High droplet concentration More
reflective cloud (Cooler climate)
4
Indirect aerosol effect (II)
Smaller droplets
Lower Precipitation rate
Clouds are longer lived and retain higher liquid
water content
5
The semi-direct aerosol effect
6
Implications to Precipitation

• Inhibited precipitation or enhanced
  precipitation?
• It depends on the cloud temperature, the
  chemical nature of the advected material, the
  pressure-level of the interaction, cloud liquid
  water content, cloud dynamics, and likely several
  other factors

7
Overview of Problem

• Clouds are the largest modulators of the solar
  radiative flux reaching the Earth's surface
• Aerosols represent the source of greatest
  uncertainty in climate forcing and atmospheric
  chemistry
• One of the greatest current challenges in our
  understanding of atmospheric physics of aerosols
  is quantitatively relating radiative forcing,
  cloud modification, and chemical properties.
• Indirect effect of aerosols are understood
  theoretically but there are not many practical
  cases or experimental evidences to date
• Organic content of aerosols is significant but
  not well depicted in most atmospheric models
  incorporating ambient aerosol

8
Scientific Drivers

• Hypotheses
• Nucleation properties change as a function of the
  microphysics but not necessarily in linear
  fashion
• Microphysics varies as function of chemical
  composition
• Precipitation and rainfall structure will be
  impacted as function of aerosol chemical
  composition
• Goal and approach
• Systematic laboratory characterization of
  aerosols
• Comparisons of experimental and theoretical
  nucleation potentials
• Evaluation of trends in behavior as a function of
  composition

9
Summary/Science Traceability
10
Climate Change Impacts
How Do Number Distributions Change?
How Do Nucleation Properties Change?
How Do Size Distributions Change?
Microphysics
Composition
Microphysics
How Do Optical Properties Change?
How does the chemical composition of CN affect
cloud properties?
Optical Properties
11
Specific Tasks of Cloud-Aerosol Studies

• To develop a reliable laboratory technique to
  generate carbonaceous, aromatic, and mineral dust
  aerosols.
• To systematically investigate the basic
  nucleation properties of elemental carbon,
  organic carbon, and mineral dusts
• To systematically investigate the nucleation
  properties of aqueous/organic solutions which are
  capable of becoming cloud condensation nuclei.
• To clarify any structure-nucleation relationship
  identified through the study

12
Experimental Approach

• Investigate the influence of aerosol composition
  on cloud condensation nuclei based on changes in
• Size distributions
• Electrical mobility
• Number distributions of aerosols and CN
• Optical extinction

• Investigate perturbations by chemical class
• Aromatics
• Acids
• Ketones and aldehydes
• Esters
• Examine Homogeneous vs heterogeneous inclusions
• Graphite 100 ng/m3
• STM soot
• Mineral dust

13

• Baseline Data

14
Aerosol Nucleation Potential

• For a fixed R.H
• Mass of solute (dry diameter)
• Activation diameter (wet diameter)
• Need to estimate the effect on the Raoult term
  and Vant Hoff factor

Haze
Haze
Activated nucleus
Stable
unstable
15
Köhler curves for relative humidity of
100.55
chlorobenzene
hexane
benzene

• The surface tension is lowered
• hexane 70
• benzene 48
• chlorobenzene 51
• nitrobenzene 60

16
Nucleation properties of benzene

• Given
• Decrease of Surface tension of 48 from 0.073
  J/m2 to 0.035 J/m2
• Köhler curve gives
• Dry diameter is
• 0.029 µm
• Wet diameter is 0.123 µm
• The DMA size distribution
• ¼ nucleation particles
• All accumulation particles

17
Nucleation properties for nitrobenzene

• Given
• Relative Humidity of 100.55
• Decrease of Surface tension of 60 from 0.073
  J/m2 to 0.044 J/m2
• Köhler curve gives
• Dry diameter is
• 0.022 µm
• Wet diameter is 0.123 µm
• The DMA size distribution
• 1/5 nucleation particles
• All accumulation particles

18
Summary Data
19
What Are the impacts of Sahara Dust on Atlantic
Ocean Rainfall Structure?- as derived from
multi-satellite/multi-sensor observations

• Direct evidence for desert dusts effects on
  rainfall structure
• over the ocean has not been reported.
• Model simulations and observations of surface
  rain rate are inconsistent. Some observations
  show that dust suppress clouds and precipitation
  Rosenfeld 2000, Rosenfeld et al. 2001
• Observations of proposed AIE (aerosols indirect
  effects) from satellite platform are not
  consistent (Shao and Liu 2005).
• Giant CCN may enhance the collision and
  coalescence of droplets and therefore increase
  warm precipitation formation and decrease the
  clouds albedo Yin et al, 2000, van den Heever
  et al, 2005.
• How does one control for thermodynamic
  variation?

20
AEROSE is a series of field campaigns designed to

• Provide a set of critical measurements to
  characterize the microphysical and chemical
  evolution of Saharan dust aerosol during
  trans-Atlantic transport
• Obtain in-situ characterization of the impact of
  aerosols of African origin on energy balance and
  tropospheric chemistry in the tropical Atlantic
  Ocean,

• Obtain bio-optics and oceanographic observations
  in order to study the effect of the dust on the
  marine boundary layer, characterize water masses
  throughout the transects, as well as to
  investigate upwelling conditions off the
  Northwest coast of Africa.
• Provide additional complementary visible and near
  IR measurements that can support the validation
  and improvement of dust aerosol AVHRR SST
  corrections, the validation of MODIS data and
  products, upwelling activity, and associated
  biological signatures.

21
(No Transcript)
22
March 13
March 11
Evolution of Aerosol Surface Elemental
Composition
23
Observations from METEOSAT Visible imagery (Mar
310, 2004 - 1 frame / day)
24
Brief overview of the dust event

• On 3 March 2004, the massive storm formed a huge
  arc of thick dust that swept over the Canary
  Islands where it dropped a significant amount of
  dust. This event was captured by various
  satellites, including Meteosat-8 and NASA's Terra
  and Aqua. On 5 March 2004, the dust, still thick
  and well visible in the satellite images, reached
  the Cape Verde Islands and the shores of Western
  Europe. In the following days, the dust crossed
  the Atlantic Ocean and reached South America and
  the Caribbean Sea.  During this process, the dust
  got thinner and thinner (smaller dust particles
  and smaller aerosol optical thickness) making it
  less visible in the satellite images. However, on
  10 March 2004 large amounts of fine dust were
  still well visible in the area of the Gulf of
  Guinea.

Mar 03 UT1200
Mar 05 UT1200
Mar 03 Onset of event Mar 05 Reaches Cape
Verde Mar 07 Diffusing into Atlantic Mar 10
Still well visible throughout region
Clean
Clean
Mar 07 UT1200
Mar 10 UT1200
Dusty
Dusty
From http//oiswww.eumetsat.org/WEBOPS/iotm/iotm
/20040306_dust/20040306_dust.htmlpics
25
Principal Aims

• Use rainfall profiles and microphysicd derived
  from satellite-based passive microwave sensors
  measurements to examine the dusts impacts on
  rainfall structure.
• Determine a method to exclude confounding factors
  introduced by thermodynamic conditions when
  studying AIE using satellite measurements.

26
Data

• Geostationary satellite (Meteosat-8) images with
  high temporal resolution (4 frames/hr ) served as
  background to judge dust storms distribution.
• AERONET station in-situ observations serve as
  additional evidence to judge the location of dust
  storm.
• Low-orbit satellite (TRMM) observations/retrievals
  of rain profiles from multi-channel passive
  microwave sensor (TMI) and active microwave
  sensor (Precipitation Radar) provide us unique
  information of the inner structure of rain (TRMM
  standard products - 2A12 for TMI and 2A25 for PR)
  .
• TRMM Microwave Imager have 5 frequencies 10.7,
  19.4, 21.3, 37.0 and 85.5 GHz. All of them are
  dual polarized except for 21.3 channel, which are
  only vertical polarized.
• The TRMM standard product, 2A12, output vertical
  profiles of four kinds of hydrometeors at 17
  level. They are precipitation water, cloud water,
  recipitation ice and cloud ice.
• Polar-orbit satellite (AQUA) observations/retrieva
  ls of rain profiles by multi-channel passive
  microwave sensor (AMSR-E) add more useful samples
  into this study (non-official product, especially
  provided by Dr. Kummerows Group using the same
  algorithm of TMI, i.e. GPROF algorithm ).

27
Method - case study

• Several studies have reported that AIEs can be
  confounded by many parameters
• Cloud thermodynamic conditions and
  inhomogeneities
• Actual aerosol physical and chemical properties
• Biases in algorithms used to derive
  cloud/rainfall and aerosol characteristics.
• The Sahara dust distribution generally has a
  North-South gradients. If we can identify a
  rainfall system that is partially immersed in the
  dust - then we can get two dustier rainy areas
  and two clearer rainy areas by splitting the
  rain area into four quadrants (NE, NW, SW and
  SE).
• If we assume similar thermodynamic conditions and
  aerosol properties in each quadrant then the
  difference in rainfall structures among these
  quadrants should mainly be due to the dust.
• Combining the observations from Meteosat-8, TRMM
  and AQUA, we have identified two such cases in
  this study.

NE
NW
SW
SE
Clean Sky Rainy Dusty
28
Method statistical study

• Studies have shown that tropical rainfall
  vertical structures have large regional
  differences among those occurring along
  coastlines, inland and open oceans.
• To substantiate our assumption of controlled
  thermodynamics we compared the rainfall
  structure in given area during two different time
  period. One is dust-free period and the other
  is dusty period.

Mar 1 2004
Mar 6 2004
Mar 10 2004
Mar 5 2004
Dust-free period
Dusty period
29
Case study using TMI profile
(a)
(b)
Satellite observations of the dust and
cloud/rainfall system over Atlantic ocean. (a)
RGB compositing image using 0.6, 0.8 and 3.2-um
channel taken by Metsat-8 on 8 Mar, 2004 at UT
0912. The dust inflow from Sahara desert are
represented as white Smog which is clearly
distinct against other objects. The two major
cloud system represented as blue colors can be
seen in the image. Both of them are partially
invaded by the dust. One of them (the left and
larger one) are almost simultaneously detected by
TRMM. (b) The rainfall system detected by TRMM
TMI on 8 Mar, 2004 at UT 0911. (c) The four
quadrants we divided for the rainfall case
detected by TMI. Red and blue pixels indicate
convective and stratiform rain pixel follow our
definitions. Number of samples in each quadrant
is shown in bracket . (a) and (b) are made by
software UMARF native format reader and TRMM
orbit-viewer, respectively.
(c)
30
Observations from TRMM TMICase II. Clean
Rainfall System
TRMM TMI Surface Rain (mm/h) Orbit number
35881 Date Mar 02, 2004 UTC 0132 Condition
Clean (no dust)
About 3 days before the dust storms coming into
this area.
31
Observations from TRMM TMICase III. Dusty
(mostly) Rainfall System
METEOSAT IMAGE Date Mar 08, 2004 UTC 1200
TRMM TMI Surface Rain (mm/h) Date Mar 08,
2004 UTC 0911 Condition Dusty (partially)
32
Case study using TMI
Upper two plots Profiles of PW (Precipitable
Water), NPW (Normalized PW) and the
precipitation efficiency index - PEI (ratio of
PW to all hydrometeors in the air) in the four
quadrants. Lower four plots Additionally,
we use a mean clean case (March 2,
2004,UT132) and a mostly dusty case (March 7,
2004,UT 2139) as references.
33
Case study using PR
TRMM PR is the only satellite-based active
microwave instrument to detect rainfall. Profiles
derived from PR attenuation-corrected
reflectivity are regarded as more direct
measurements of the rainfall inner structure.
They have horizontal and vertical resolution of
4.3 km and 0.25 km, respectively. PRs swath
(about 220 km ) is less than one-third of TMI
swath (about 760km) so we only divide the rain
system into 2 sectors (using the same borderline
as described above). The north sector is regarded
as dusty and the south sector is regarded as
clean. PR cannot measure the rainfall
structure near the Earth surface because its
returns are contaminated. We only report the
profiles above 1.5 km. PR cannot distinguish
hydrometeors into ice or liquid water, rain or
cloud droplet. So we can not give the
precipitation efficiency index. Results derived
from TRMM PR are consistent with those of TMI.
The rain intensity is weaker in dust area at each
altitude.
34
Classification of rainy pixel and rain type

• TMI rainy pixel surface rain gt 0
• TMI convective rain pixel (convective rain /
  surface rain ) gt70
• TMI stratiform rain pixel (convective rain /
  surface rain ) lt70

35
Case I. Partially Dusty Rainfall System
Convective rain pixel Stratiform rain pixel
36
Case II. Clean Rainfall System
Convective rain pixel Stratiform rain pixel
37
Case III. Dusty (mostly) Rainfall System
Convective rain pixel Stratiform rain pixel
38
Stratiform Rain Profiles
Partially Dusty
Clean
Mostly Dusty
The mean intensity of precipitation water is
weakest in mostly dusty case (except for the SE
region). Large regional variations can be found
in partially dusty case. Mean intensity in NE
and NW region are significant weaker than those
in SW and SE region, and are close to those in
mostly dusty case. But there is no significant
difference between Clean case and Partially Dusty
case.
39
Convective Rain Profiles
Partially Dusty
Clean
Mostly Dusty
Mean intensity of precipitation water is weakest
in mostly dusty case (except for the SE region)
and strongest in Clean case.
40
Precipitation size growth PR Reflectivity
More precipitation-size ice here!
Weaker near surface radar reflectivity
41
Dusty Area
Convective Rain Stratiform Rain
advection
Saharan dusts act as ice forming nuclei to
produce more, small size cloud ice particles, but
unable to grow up to PR detectable ice particles
due to insufficient water vapor supply and short
life time.
More small ice particles continue to grow up
slowly, producing more PR detectable ice
particles as compared with its counterpart in
dust-free area.
Sahara Dust Layer Suppress the water vapor
supply And increase ice forming nuclei.
42
Summary

• Dusts, transported up by the strong convective
  updrafts acted as additional ice nuclei. Some of
  ice particles grow and contribute to convective
  precipitation, and others were advected into the
  neighboring stratiform region and enhance
  nucleation leading to precipitation in the
  stratiform region. Thus, dusts enhance stratiform
  precipitation.
• The microphysical effects of dusts in the
  convective regions were shifting the
  precipitation size spectrum from heavy to light
  and suppressing precipitation. Dusts also
  enhanced evaporation processes, which further
  reduced the precipitation reaching surfaces
• The cloud system adjusted itself to these changes
  and resulted in a weak but long lasting cloud
  system with increasing convective precipitation
  fraction and decreasing stratiform precipitation
  fraction.

43
Acknowledgments

• Dr. C. Kummerow
• Mr. Betty-Ann Garriques
• Mr. Temesgen Sahle
• Mr. Paul Nkansah
• Mr. Evans Dure

44
Extra (Supporting) Slides
45
Comparison of retrievals from TMI and AMSR-E
0.1 1.0 10.0
100.0 TMI Surface Rain rate (mm/h) Mar
2,2004 UT 0132
AMSR Surface Rain rate (mm/h) Mar 2,
2004 UT 0241

• TMI captured only one case which partially
  impacted by the dust. Fortunately, the AMSR-E
  aboard NASAs AQUA satellite capture another
  rainfall system which is also partially immersed
  in the dust storm.
• Because of requirement of extra CPU time and huge
  storage space ( Dr. Kummerow, private
  communication), so far, no vertical information
  are released by AMSR-E group.
• But the AMSR-E code is fundamentally the same as
  the TMI GPROF, and the observations of AMSR-E (12
  channels, 6 frequencies 6.925, 10.65, 18.7,
  23.8, 36.5, and 89.0 GHz. All are dual polarized)
  are enough to retrieve rainfall profiles.
• Before we use these profiles, we performed an
  examination of the consistency between TMI and
  AMSR-E. The case we select for this exercise is a
  rainfall system detected by AMSR-E on Mar 2 2004
  at UT 0241 and by TMI at Mar 2 2004 at UT 0132.
  Comparison are done in all four quadrants of this
  system.

46
Comparison of retrievals from TMI and AMSR-E
We observe poor agreement between the mean
profiles of precipitable water located in the
southeast quadrant is between the TMI rand
AMSR-Es retrievals. The other AMSR-Es
profiles, including PW, NPW and PEI, are very
close to those of TMI. Given the lag on
observation times (1 hour and 9 minutes) between
TMI and AMSR-E, we are fairly confident in
concluding that there is no systemic bias between
this two retrievals.
47
Problems

• Interpretation of the mechanism is based on
  inference, rather than observation.
• Some one point out GCCN will enhance
  precipitation while CCN will suppress
  precipitation. In this study, we can not
  distinguish CCN from GCCN because of lack of
  dusts size information.

48
A partially dusty Case UT 911, March 8, 2004

49
T and RH profiles derived from AIRS
In the two northern (dusty) quadrants, the air
temperature from 950 hPa to 700 hPa is a little
bit higher (Max 1.5 degree) than those in the
south two quadrants(dust-free). Additionally,
the difference of water vapor mass mixing ratio
and the saturate value is more negative from
surface to 700 hPa in the two northern (dusty)
quadrants. These plots indicated that there
is indeed a warm and dry layer in the dusty rain
area. This is an expected result of the dust
intrusion. Thus, we can infer that the
evaporation process is more intense and
suppresses convection in this sector.
Dust
50
Stratiform Rain Efficiency Index (REI)
Partially Dusty
Clean
Mostly Dusty
Clean
?
?
Dusty
Precipitation Water
REI
(precipitation water precipitation ice cloud
water cloud ice)
Generally, the REI in dusty cloud is significant
smaller than those in clean cloud. Clean REI
increases with decreasing altitude. Dusty REI
decreases with decreasing altitude at the layer
from 2.75 to 3.75km.
51
Dusty
Clean
When dust inject into cloud, raindrops evaporate
and result in a low increase speed toward earth
surface. In layer from 2.75 to 3.75km, the
relative Increase speed is even lower than that
of cloud water drop, thus product a positive
slope of .
From case 35979 (partially dusty)
52
Convective Rain Efficiency Index (REI)
Partially Dusty
Clean
Mostly Dusty
REIs in dusty clouds are significantly smaller
than those observed in clean clouds.
53
Statistical study (EQ-4N)
Samples Dust-free Convective rain pixel 3866
Stratiform rain pixel 8420 Dusty
Convective rain pixel 1580
Stratiform rain pixel 2549
54
Statistical Study of the Ice-forming process
The intensity of precipitation Ice is weaker for
dusty conditions (6 Mar to 10 Mar) than in
dust-free conditions (1 Mar to 5 Mar). This
indicates that the dust may inhibit the rainfall
not only in warm rain processes, but also in the
ice forming process. The implication of
suppressed convection is less water vapor being
transported to the upper layer. On the other
hand, in the upper layer, normalized
precipitation Ice for stratiform rain under dusty
conditions is larger than under dust-free
conditions. This means the growth speed is
quicker when dust exists for a given ice
intensity at 5.5 km altitude. This can be
explained if dust particles in the air increase
the concentration of ice nuclei, so that the
probability of ice crystal colliding with
super-cooled droplets or other ice particles
increases, thus, enhancing growth speed is This
phenomenon can not be seen in convective rain
mainly because ice is more common in stratiform
rain than in convective rain for a given surface
rain rate. The indirect effects of dust on ice
forming process is less important than its
suppressing on the convection intensity.
55
Case study of the Ice-forming process