NASA Technical Reports Server (NTRS) 20020021952: Measurements of Ice Particles in Tropical Cirrus Anvils: Importance in Radiation Balance

6 Pages·0.32 MB·English

by  NASA Technical Reports Server (NTRS)

Preprint Vol.
AMS Conference on Cloud Physics
January 15-20, Dallas, Texas
9B.6
MEASUREMENTS OF ICE PARTICLES IN TROPICAL CIRRUS ANVILS:
IMPORTANCE IN RADIATION BALANCE
*Theodore Foster, **William P. Arnott, **John Hallett,
and ***Rudi Pueschel
*California Polytechnic San Luis Obispo
San Luis Obispo, California
**Desert Research Institute
Atmospheric Sciences Center
Reno, Nevada
***NASA-Ames Research Center
Moffett Field, California
1. INTRODUCTION
that for strongly absorbing wavelenghts the average
projected area of the ice particles isthe important
Cirrus isimportant inthe radiation balance of parameter; in weakly absorbing regions it is the
the global atmosphere, both at solar and thermal IR volume (mass) of ice which is important. The
wavelength. In particular cirrus produced by deep shape of particles and also their internal structure
convection over the oceans in the tropics may be may also have significant effect on their radiative
critical in controlling processes whereby energy properties (Liou, 1986, Arnott et al. 1994ii).
from warm tropical oceans is injected to different
levels in the tropical atmosphere to subsequently In order to access the role of cirrus in the
influence not only tropical but mid latitude climate radiation budget it is necessary to measure the
(Ramanathan and Collins, 1991). Details of the distribution of ice particles sizes, shapes and
cloud composition may differentiate between a net concentrations in the regions of interest. A casual
cooling or warming at these levels (Lion, 1986). observation of any cirrus cloud shows that there is
The cloud composition may change depending on variability down to a scale of at least a few 10"m;
the input of nuclei from volcanic or other sources this isconfirmed by radar and lidar remote sensing.
(Sassen, 1992). Observations of cirrus during the Thus aircraft measurements designed to give insight
FIRE IIProject over Coffeyville, Kansas (Arnott, et into the spatial distribution of radiation properties
al, 1994i) and by satellite demonstrate that cirrus, of ice crystals must be capable of examination of
on occasion, iscomposed not only of larger particles concentration, size and shape over a distance ideally
with significant fall velocity (few hundred tzm, of 100 m or less and to detect particles down to a
'/2m st) but much more numerous small particles, size below which radiative effects are no longer
size 10- 20 _zm, with smaUfall velocity (cm s"), significant.
which may sometimes dominate the radiation field.
This is consistent with emissivity measurements 2. INSTRUMENTATION
(King et al, 1993). in the thermal IR, ice
absorption isstrong, so that ice particles only 10tzm Measurements were made during the
thick are opaque, at some wavelengths; on the other TOGA COARE Project, using the NASA DC-8
hand at other wavelengths and in the visible, ice is aircraft, in flights over the W. Pacific Jan./Feb.
only moderately to weakly absorbing. It follows 1993. Inview of the importance of tropical cirrus in
the radiation balance, these measurements were also possesses problems of interpretation, for such
specifically designed to cover the range of particles small particles. Particles greater than a few 100 _.m
important in the radiation balance with a sampling shatter on collection (which can be reconstructed);
rate of each instrument of some liters per second. more serious is that they may shatter on entry into
Larger particles (> 100/zm), were detected by the the slit itself. These fragmented particles can
2 DC PMS optical probe and smaller particles largely be excluded from analysis by a consideration
_>5_m by a formvar replicator which collected of the concoidal fracture consistent with high
particles through a 3 mm slit in a plastic solution to speed impact of a brittle solid - as occurs in
leave permanent casts. In parallel, measurements, shatter of a glass window. In parallel with the ice
were made of total nuclei (CN by a TSI 3010 measurements, data was obtained with a FSSP-
instrument), which detects particles down to 300 X with 15 channels centered between 0.4 and
< 0.01/zm; details of these measurements are being 20 #m. The data from this type of instrument in ice
published elsewhere (Hudson et al 1995). Data is is open to interpretation. Gardiner and Hallett
analyzed from a few seconds to 200 s of flight path (1985) showed that counting of supercooled cloud
representing V2km and 20 km path lengths. This droplets in the presence of some ice particles was
enables measurement of the lower concentration of enhanced: the shape and internal structure of small
larger particles to give more relevant statistical data. ice particles has potential for influencing counts.
Overall estimate isa potential for x 10enhancement
3. OBSERVATIONS from replica from the conditions for shatter, and a
comparable factor for FSSP total counts. A
Deep cumulonimbus convective clouds comparison of small particles by both techniques is
grow, subject to local dynamical forcing, over the made here, which give comparable results and some
west Pacific warm pool to altitudes > 50,000 ft and confidence in the measured concentrations.
top temperatures < -70°C. These clouds give ice
anvil as they extend through the .40_C level which
is spread away from the main vertical motion in 4. RESULTS
winds prevailing aloft. Instrumentation was carried
by the NASA-DC-8 which flew typically up to Measurements of replica, PMS - 2DC data,
42,000 ftwith local air temperatures down to -550C. and FSSP data (Flight 12 only) are shown in
Penetrations were made of anvils, not necessarily composite plots at specific times and temperature.
near the tops, which on occasion extended well Data represent integration over about V2minute,
above the highest flight level. Some penetrations representing about 5 km of flight path. Fig. I shows
were also made at lower levels. The data presented data from an anvil at a temperature of-55°C.
here were selected from flights where good quality Replica and 2 D-C are in sufficient agreement on
replica and PMS data were collected. Initial the region of overlap around 100/_m. It is noted
analysis consists of size, habit, and concentration that the concentration rises with decreasing size
from PM 2DC images and from replica. Caveats in to the effective limit of replica analyzed - in this
the data are that "null" images in the PMS (one case about 10/_m. (The smaller crystals were
pixei) need to be distinguished from noise, giving an examined using microscopy; the large ones by
effective lower limit for detection of two pixeis, and projection). The plot suggests a greater number
of habit estimation of some 10 plxels in diameter. than a straight line extrapolation log concentration
The replicator was mounted in a standard PMS pod plot of the larger size particles suggest. Composite
and transported film at 2 cm s"_, sampled through data including PMS counts are shown in Fig. 2,
a 3 mm slot. Thus each 16 nun fiJm frame collected which suggests even larger numbers at smaller size
crystals from a 200 m path. In principle, a "point" (nominally to 0.3/zm). The number concentration
analysis of 3 mm of film gives a minimum resolution near 10 /_m are consistent between FSSP and
of 50 m. Sampling of particles; below a few _m replica; there is also consistency in the overlap
size falls off because of reduced collection region between replica and 2D-C probe near
efficiency; particle "noise" in the solution and film 100 _m.
H IN Aavil outflow from Hurricane Olivet; 1711 52-
Figure la, b: Anvil from old convection; flight 15. 171234z. Hight 12,6 Feb.1993. Temperature -46*C. Replica -
Temperature -55"C. Tune: 22 34 59 -2235 15z. 17 Feb. 1993. 2D-C data overlap 75- 150 ttm; FSSP - replica overlap 10 -
LAT: 40" 20; LONG: 155"00 E. Replica and PMS -2D data,
201an. LAT: 11° 20 S. LONG: 155"15E. Squares are FSSP;
effective overlap "/5- 200_.rn. Replica and PMS data are others are replica.
conJilten| in overlap region. Square,s are PMS 2D-C, others
replicaforvarious analyses. 105
a)
a) 108 _
107 ,
104
N
o
106 mM
ea +t, a
_--'-- 103 * _ _.
105
{I
E I ."
104 Q ,
r_
Z 102 _,
Z o
"0
103 a
_ A
•a 4tA A
102 6
a
a
6
A 6
101 a 1011
100 0 I00 200 300
0 1O0 200 300
d (p.m)
d (IJm)
b) 10s
b)
10e
107 N
x _
* N
104
a
106 ao • N
o
a N
* =o
I0 s 103
'E • o_oeo 'E :,.;.
o ¢,'3
E
E 104
6e
"0 • 66
"0 z 102
z "U
"0 Cl
103
oa
102
•44at 10_
10_
100
10o I0t 102 103
10"I I00 10_ 102 103
d (p.m)
d (l.lm)
3
Figure 3 shows a composite data set of Figure 4a, b: shows a plot of area and
replica and 2D-C where consistency in counts occurs volume (equivalent circle diameter) related to sizes
near 100/.tin, but replica shows smaller numbers of suggesting a maximum contribution of particles
particles larger than this size by a factor of about 5. 15 - 20/_m diameter, with a replica fall-off for
Such discrepancy can be attributed to amore fragile large particles > I00 _m.
structure of the large crystals, which break prior to Neither PMS or replica showed any
replication observation of a large number of significant evidence for crystal aggregation at any
rejected shattered fragments are consistent with this level.
possibility.
104 .
2.107
oie
103 .
G,
---.
:L
A
E 102 o A • %
---.
1D
Z
"O A
E
1.107
a
101 -o
-o
m
100 .....
o
0 1O0 200 300 400 500 0
d(_m) o o o all Ko\
°el m _
o
10"1 100 101 102 103
Figure 3: Anvil from old convection; descending
d (gm)
cirrus plume. Hight 18; 24 Feb. 93. Temperature -
28°C, 22 37 21 - 22 37 49z. Approx: LAT 151
LONG: 2° S. Replica and PMS 2DC effective Figure 4a: Derived area of particles from in
overlap 75 - 200 _m. Replica gives fewer particles Fig. 2 from (fit 12) assuming an equivalent circle
than PMS for size > 150 _zm. Squares are PMS diameter. This data suggests a maximum area
2DC. contribution for particles near L5 - 20 jzm size.
3.108
120°C, strongly suggesting that their composition
was sulfuric acid. Such numbers are consistent with
observations of Clarke (1993). The ice particle
2 108 concentration at the smallest size examined is a few
o
cm 3, indicating that only a fraction of the aerosol is
'e
activated to become ice particles. This implies that
=. r_ o there could be present in the cirrus, in addition to
E g o the small ice particles, even larger number of
o
a a inactivated sulfuric acid haze particles. Should the
v:I. region of air containing such particles be
C3
subsequently lifted, further activation and freezing
>
13 1'108 could ensue as particles dilute and freeze
homogeneously. The origin of the initially large ice
g •
a concentration is of interest. Qualitatively, a
a
a significant parameter is the temperature at which
condensation begins on any aerosol present.
Cloud bases of deep convection are at temperatures
well above +20°C, so that the initial droplet
a spectrum isformed by the usual competitive process
at these levels, controlled by the surface CCN
10"_ 100 101 102 103 spectrum and cloud base updraft. Much happens
before these particles reach higher colder levels
d (pm)
where ice nucleation begins- warm rain removes
Figure 4b. Derived volume distribution showing a much water substance, nuclei are scavenged by
wide maximum contr_ution of particles 20 to various processes and incorporated by mixing from
100 _m. higher levels as the cloud builds upwards. The
ultimate ice concentration, as homogeneous
nucleation takes over after passing -40°C, can be
5. DISCUSSION envisioned to result from an identical competitive
process whereby nuclei are diluted and frozen
The most important result from these homogeneously, depending again on the tail of the
observations is the presence of large numbers of CCN spectrum and vertical velocity. A knowledge of
particles of size below 50 _.m which contribute the complete cloud properties (CN, CCN, vertical
significantly to the effective surface area and opacity velocity), it is vital in understanding the ice origin.
of the ice layer. It is important that the presence of The origin of the nuclei ALOFT can be speculated
the particles is shown in data obtained from two to be possibly from the surface - advected in strong
quite independent techniques- replicator and FSSP. convection, and possibly from moderate volcanic
It is to be noted that the numbers are well in activity which is always going on in the W. Pacific.
excess (x 100) of those found by the ER2 at higher The relative role of volcanic and ocean production
levels in the STEP Project (Knollenberg et ai 1993) is an area for potentially fruitful speculation.
using a different FSSP with a much lower response
to these sizes of particles. The presence of such Acknowledgment: The assistance of Duane Allen,
particles is apparent in all 3 cases chosen for Steve Howard, Guy Ferry in the instrument
analysis, at temperatures -28,-45 and -550C. In the installation and data acquisition was vital to the
former case, the penetration was through a project and is gratefully acknowledged. Patrick
descending region of ice from a higher layer aloft: Arnott and John Hallett were supported by NASA
this shows that such small particles are preserved in Grant No.NAG-I-1113. T. Foster was supported
such weak vertical motions. It is of interest that CN by a University of Stanford summer fellowship at
measurements made during these flights, showed, NASA Ames.
on occasion, counts as high as 5,000cm "j
(compared with "normal" background of a few
hundred cm_). Volatility analysis in a controlled
furnace showed that these particles evaporated near
a) b)
c) 0.)
Iq_ure 5. l(xamples ,3f lL,rmvar replica: a) partly
fragmented polycostal, shov, ing crystals s_3th hexagonal Gardiner. B.A. and J. llallett. 1985: De_adation of in-
p_.nphcry. _. b) hcxag, mal plate, w3th shatter fragments cloud forv, ard scattering spectrometer probe
from other c_'stals {these would not be counted as measurements in the presence of ice particles.
individual co'stals), c) columns, partly shattered on J. Oceanic and Atmos. Teeh., _, 171-1R_3.
collection (fit 12). d) cry'sial showing tngonal sTmmet W tludson. J.G., W.P. Arnott. J. tlallett and R. Pueschcll.
(flt. I"). 1994: In preparation.
King, M.D.. 19 93: Radiative Properties of Clouds.
,,. RI:_HiRt-N( '[.IS AerosoI-Cloud-Clim_te Interactions. editedP.V.
llobbs, Academic Press, inc. 123-149.
Arnott. _,'. I'., Y.Y. Dong and J. tlallett, 1994i: Knollenberg, R.G., K. Kelly. J.C. Wilson, 19')3:
l.lx'_ncmm F.fficiency in the [R (2 _um to 18_m) Measurements of high number densities of ice
,,f l..aborato_' Ice Clouds: Observations of c_'stals in the tops of tropical cumulonimbus. J.
Scattering Minima in the Christiansen Bands of Geophvs. Res., _. 8639-8664.
Ice. Accepted tier publication. J. of Applied l.iou, K.N., 1986: Influence of Cirrus Clouds on Weather
Optics, 3__33. and Climate Processes: A Global Perspective.
:\rnott. V,,. t'., Y.Y. l),,ng, J. llallctt and Michael R. Monthly Weather Review, 114, 110%1199.
Pt,clh,t. I','_411: ()bs_'_'atlons and importance of Ramanathan, V. and W. Collin._ 1991: Thermodynamic
small _cc erv',tals m ac_rrus cloud from FIRE II regulation ot ocean warming by cirrus clouds
data. 22 N,_s' 19')4. J. of Geophys. Res.. 9_..._). deduced from observations of the 1987 F.1Nifio.
13"7,1-1381. Nature. 35_.,127-32.
Clarke. A. I).. 19')3: Atmospheric Nuclei in the Pacific Sa.s..sen, K.. 1992: Evidence for Liquid-Phase Cirrus
Midtropsphere; l'hcir Nature. Concentration Cloud Formation from Volcanic Acro_qs:
and I:_volution. J. of (;e_phys. Re_, 9_1,2f).633- ('timatic Implications. Science, ZS7. 516-521.
. _.t>4.