Recent observational studies on ice nuclei and ice formation in clouds

1
Recent observational studies on ice nuclei and
ice formation in clouds

• Paul J. DeMott
• Colorado State University

2
Overview

• Focus primarily on some examples of studies ice
  and mixed-phase clouds in last 10 years in which
  IN and ice concentrations were measured.
• Some inferences about our understanding of upper
  tropospheric ice formation from studies of
  natural IN and clouds at low temperatures.
• A general realization/validation mineral dust is
  an important source of atmospheric IN.
• A few thoughts on future needs.

3
Wave cloud studies (1990-present)

• Heymsfield and Milosevich 1993-1995 papers
  suggest that there are few heterogeneous IN in
  upper troposphere and that ice formation by
  homogeneous freezing dominates in cold wave
  clouds
• Some during WISP studies (1993-1994) IN
  collected from around clouds and processed in
  CFDC and controlled expansion cloud chamber. IN
  concentrations reasonably consistent with ice in
  clouds, but method not sufficient to explain
  variability in time and space. Tests for
  evaporation IN in controlled expansion cloud
  chamber find no more than 2-3 enhancement.
• Wave clouds below -40C during SUCCESS (1996) show
  total ice concentrations consistent with
  homogeneous freezing, but also evidence
  consistent with presence of IN up to 100 per
  liter at low temperatures. Such high IN not
  always there and sometimes few deposition nuclei
  present.
• U.K. studies (Field, Cotton, et al.) using SID
  show evidence of strong ice formation mechanism
  in evaporation portion of modestly supercooled
  wave clouds. Does not always occur.
• WAVEICE studies (2000) Little apparent evidence
  for enhanced ice formation in downstream portion
  of modestly supercooled wave clouds. IN upstream
  of wave cloud are reasonably consistent with ice
  formed in cloud. Some evidence for springtime
  dust impacts on cloud ice formation.

4
Some early optimism that IN measurements are
meaningful Winter Icing in Storms Project

• WISP 1994
• NCAR Electra, Wyoming KingAir
• Six wave clouds
• Upwind-downwind penetrations
• Ice concentration from PMS 2DC and 1DC
• Upwind aerosol bag samples, analyzed at CSU lab
  with CFD Dynamic Cloud Chamber

5
WAVEICE 2000 March 17, 2000
6
WAVEICE 2000 Ice Concentrations in 35 Cloud
Passes - March 17
200X probe
7
Inference that, in absence of secondary
processes, and at T gt -38C IN Ice
Parcel model uses IN and CCN measurements
Data from Wyo. KA March 17, 2000
8
A number of observations have been made of
enhanced ice formation in evaporation region of
waves
Cooper (1995, AMS Cloud Physics Conf.). See also
Cotton and Field (2002, QJRMS)
9
Preliminary Inferences from AIRS-2 Studies on the
Role of IN in the Evolution of Mixed Phase Clouds

• Extremely inhomogeneous spatial distributions of
  IN may exist in the atmosphere prior to winter
  storms.
• This heterogeneity is reflected by the IN
  detected from cloud particle residuals (sampled
  by CVI) but the interpretation of these data may
  not be straightforward.

10
November 14 Lower clouds with and without ice,
deeper clouds and cirrus in some areas. Were in
process of transitioning CFDC conditions to equal
those in lower clouds at this time
11
November 14 Distinct layers with high IN aloft,
sometimes reaching down to lower clouds. Some
clouds had ice and some not. Likewise, some had
IN, some not.
12
IN in deep precipitating cloud system on
November 19, 2003 and relation to cloud residual
aerosol
CFDC processing T -12.5C RHw 102
13
Other Mixed Phase Cloud Studies Including IN
Measurements

• LAKE-ICE IN relate to cloud ice in lake-effect
  systems
• FIRE-ACE/SHEBA Generally lower IN in Arctic,
  possible sources from open ocean leads,
  silicate/sulfur chemistry of IN (Rogers et al.
  2001).
• North Dakota Tracer Experiment Bag samples from
  cumulus cloud base levels and surface sampling
  suggests agreement between IN and young updraft
  ice (Stith et al. 1994 DeMott et al. 1995)

14
Free tropospheric sampling of concentration and
composition of nuclei for cirrus formation
Storm Peak Laboratory (3220 m MSL Steamboat
Springs, CO, USA)
Cziczo et al. 2003, AST
INSPECT Nov. 2001
Aerosol processing methodology (Spring 2004)
15
Cirrus ice formation conditions and ice
concentrations nucleated on ambient tropospheric
aerosol particles
Homogeneous freezing
Heterogeneous ice nucleation
DeMott et al. 2003, December, PNAS
16
Inferences based on atmospheric observations of
RH-T conditions required for cirrus formation
NASA-SUCCESS RHi inside/outside cirrus, wlt1m/s
(Jensen et al., JGR, 2001)
Homogeneous freezing of pure sulfates from Chen
et al. (2000) or Koop et al. (2000)
water saturation
Ice saturation
17
Heterogeneous nucleation at low temperatures on
ambient tropospheric aerosol particles suggest
the range of cirrus types impacted
DeMott et al. 2003, PNAS
Smaller scale wave forcing and anvil cirrus
w
Synoptic lifting and Subvisual cirrus
Gierens (2003) critical concentration of
heterogeneous IN triggering a switch of
predominant mechanism from homogeneous freezing
to heterogeneous nucleation, as a function of T
and updraft speed
18
IN formulations for numerical modeling need for
better parameterizations and more fundamental
understanding
Meyers et al.
INSPECT (lt-38C)
INSPECT (gt-35C)
19
Do IN relate to aerosol particle concentrations
in a certain size range?
T -42 to -46C RHw 90-92
20
What is the composition of heterogeneous ice
nuclei active under cirrus conditions?
Statistics of PALMS cluster analyses of particle
types
Note Untold story about organic aerosol
components and ice nucleation
80 (1/4 with any detectable S)
20
21
Many recent results concerning cirrus from INCA

• Atmos. Chem. Phys., 3, 1791-1806, 2003 Freezing
  thresholds and cirrus cloud formation mechanisms
  inferred from in situ measurements of relative
  humidity
• W. Haag, B. Kärcher, J. Ström, A. Minikin, U.
  Lohmann, J. Ovarlez, and A. Stohl
• The analysis of field data taken at northern and
  southern midlatitudes in fall 2000 reveals
  distinct differences in cirrus cloud freezing
  thresholds. Homogeneous freezing is found to be
  the most likely mechanism by which cirrus form at
  southern hemisphere midlatitudes. The results
  provide evidence for the existence of
  heterogeneous freezing in cirrus in parts of the
  polluted northern hemisphere, but do not suggest
  that cirrus clouds in this region form
  exclusively on heterogeneous ice nuclei.
• Atmos. Chem. Phys., 3, 1807-1816, 2003 Cirrus
  cloud occurrence as function of ambient relative
  humidity a comparison of observations obtained
  during the INCA experiment
• J. Ström, M. Seifert, B. Kärcher, J. Ovarlez, A.
  Minikin, J.-F. Gayet, R. Krejci, A. Petzold, F.
  Auriol, W. Haag, R. Busen, U. Schumann, and H. C.
  Hansson
• Discusses the cloud presence fraction (CPF)
  defined as the ratio between the number of data
  points determined to represent cloud at a given
  ambient relative humidity over ice (RHI) divided
  by the total number of data points at that value
  of RHI. The CPFs taken at Southern Hemisphere
  (SH) and Northern Hemisphere (NH) midlatitudes
  differ from each other. Above ice saturation,
  clouds occurred more frequently during the NH
  campaign. Clouds during the SH campaign formed
  preferentially at RHIs between 140 and 155,
  whereas clouds in the NH campaign formed at RHIs
  somewhat below 130. Observed distributions of
  cloud water content differ only slightly between
  the NH and SH campaigns and seem to be only
  weakly, if at all, affected by the freezing
  aerosols.
• Atmos. Chem. Phys., 3, 1037-1049, 2003 In-situ
  observations of aerosol particles remaining from
  evaporated cirrus crystals Comparing clean and
  polluted air masses
• M. Seifert, J. Ström, R. Krejci, A. Minikin, A.
  Petzold, J.-F. Gayet, U. Schumann, and J. Ovarlez
• In-situ observations of aerosol particles
  contained in cirrus crystals are presented and
  compared to interstitial aerosol size
  distributions (non-activated particles in between
  the cirrus crystals). Size distribution
  measurements of crystal residuals show that small
  aerosol particles (Dplt 0.1 um) dominate the
  number density of residuals. On average the
  residual size distributions were shifted towards
  larger sizes and the calculated particle volume
  was three times larger in the Southern
  Hemisphere. The form of the residual size
  distribution did not depend on temperature as one
  might have expected considering different modes
  of nucleation. The observations of ambient
  aerosol particles were consistent with the
  expected higher pollution level in the Northern
  Hemisphere. The fraction of residual particles
  only contributes to approximately a percent or
  less of the total number of particles.

22
Dust and IN transports affect different parts of
the world at different times (E.g., Asian dust in
N. America)
Fine (PM2.5) soil concentration at the Mt. Zirkel
IMPROVE site (1993-2002).
VanCuren and Cahill JGR, 2002. Continental
transect of inferred fine Asian dust frequency
(top) and concentrations (bottom) in ng m-3.
23
Cirrus forming within Asian Dust layer Sassen
(2002)
Polarization lidar data in Salt Lake City, UT on
April 29, 2001 Considerable warmer and lower
than climatological means for cirrus
24
Possible dust impacts (Sassen 2002 GRL
introduced PDL evidence) Also evidence in
WAVEICE (2000)
Early morning 3/25 ruby lidar relative
backscattered power and linear depolarization
ratio at Salt Lake City (FARS-Ken Sassen)
March 25 Hazy day aloft
2D-c 16 to 206 l-1 T -14 to -37C
Mt Zirkel IMPROVE network sampler indicates dust
intrusion
25
WAVEICE 2000 Ice Concentrations in 40 Cloud
Passes - March 25
26
Recent Lab Studies Corroborate Ice Formation by
Dust Particles (resuspended Asian dust Cassie
Archuleta thesis)
Ca, Si, S, Mg
Homogeneous freezing points of sulfuric acid
aerosols
200 nm
Si, Al, Fe
Heterogeneous nucleation by dust
200 nm
27
Saharan dust aerosol sampled in-situ (DeMott et
al., GRL, 2003 and Sassen et al., GRL, 2003)
MODIS aerosol optical depth, July 20-27 July 29,
2003 back trajectory
Processing at T -37C, RHw 86, RHice 123
assured heterogeneous ice nucleation only
S. Florida PDL lidar data on 7/29
28
July 29 CFDC operating mostly in expected
homogeneous freezing regime at low temperature
during anvil ascent profile.

• High IWC contents
• No Citation FSSP data during period
• CFDC IN correspond with 2D within factor 2
• CPI concentrations also correspond well with IN
• IN up to 600/liter

CPI data C. Scmitt, A. Bansemer, A. Heymsfield
29
TEM analyses of IN from July 29, 2002
30
PALMS analysis of particles from high tropical
cirrus during CRYSTAL-FACE (Source D.J. Cziczo,
NOAA)

• Histogram of the area of the sodium peak in each
  positive polarity mass spectrum.
• This area can be used as a rough indicator of
  particle type.
• Most ice residue, particles outside cloud, and
  interstitial aerosols are sulfate / organics
  this is consistent with our understanding of
  homogeneous freezing.

31
PALMS analyses of anvil cirrus particles (July
29, 2002)

• Ice residue from July 28 -29 2002 have a much
  higher sodium signal than out of cloud particles
  or interstitial aerosol.
• 20 are consistent with frozen sea salt. lt10
  Sulfates and organics.
• Most of the remainder (70) are consistent with
  mineral dust or fly ash - heterogeneous
  freezing

32
Thoughts on future studies

• Still believe that wave clouds have much to offer
  in understanding ice formation mechanisms.
• Do not yet have IN measurements at appropriate
  conditions in and around cirrus (Tropical cirrus
  missions and high altitude capabilities coming).
• Need IN versus ice concentration in convective
  clouds. Likely missing an important ice formation
  mechanism.
• Validate the impacts of mineral dusts on clouds
  in programs such as AMMA?
• Continued need for laboratory studies of aerosol
  effects on ice nucleation and more fundamental
  work.
• Definition of ice versus water is still a
  critical issue for studies relating IN to ice
  formation.
• Need to take real-time combined IN and residual
  composition measurements aloft.

33
Dust versus no dust adiabatic simulation of
cumuli parcels (12 m s-1 updraft 15C cloud
base, maritime CCN)