Determination of Aircraft Icing Threat from Satellite

1
Determination of Aircraft Icing Threat from
Satellite
William L. Smith, Jr.1, Patrick Minnis1,
Stephanie Houser2, Douglas A. Spangenberg2, J.
Kirk Ayers2, Michele Nordeen2 1Atmospheric
Sciences, NASA Langley Research Center, Hampton,
VA 23681 USA 2 SSAI, Inc., Hampton, VA 23666
USA
NASA Langley Research Center
P1.75
Cloud properties and icing severity index North
America, 1545 UTC, 17 April 2007
Introduction
AIRCRAFT ICING
Satellite imagery have been helpful in diagnosing
certain meteorological conditions both
subjectively and objectively in a nowcasting
mode. However, such diagnoses are somewhat
limited for automated usage because of a lack of
quantification of the physical properties,
especially those for clouds, observed in the
imagery. With the continuous availability of
well-calibrated research satellites and the
advent of more spectral channels and higher
resolution on geostationary satellite imagers
such the GOES-I (Geostationary Operational
Environmental Satellite) series, Meteosat Second
Generation SEVIRI (Spinning Enhanced Visible
InfraRed Imager), and soon the GOES-R ABI
(Advanced Baseline Imager) it has become possible
to provide better quantification of cloud
properties that are useful for diagnosing
conditions such as aircraft icing potential and
ceiling height or for providing other information
such as surface radiation, cloud water content,
or other parameters that would be useful for
energy, agriculture, or weather forecast
assimilation. For example, algorithms used to
analyze low-Earth orbit satellite data for Earth
radiation budget and cloud process studies have
been adapted to operate in near-real time on any
satellite having the requisite spectral channels
(Minnis et al. 2001). Smith et al. (2003)
demonstrated that several of the derived products
are useful for diagnosing aircraft icing
potential. This paper describes a suite of cloud
products that are now available within a few
minutes of satellite image times over much of
North America. Australia, and Europe. The
products are being used to develop an icing
threat algorithm for GOES-R . A prototype
algorithm and some early validation results are
presented
Field Program Validation Examples
Cloud icing parameters from GOES Twin Otter
along flight track, 1998
Fluctuations in Rosemont trace indicate icing
events
Icing risk from GOES
Trace of flight path over GOES cloud boundaries
Objective An Icing threat algorithm is being
developed for GOES-R. This poster describes a
prototype algorithm, early validation results,
applications and future work.
DRAFT
Data GOES-9/10/12 4-km imager data,
half-hourly Meteosat-8 SEVIRI 3-km imager
data, half-hourly Terra Aqua 1-km MODIS
data, 2/day Rapid Update Cycle (RUC) 20-km
forecast and analyses NASA Glenn Twin Otter
icing aircraft (1997 - 2004) - cloud
microphysical properties icing North Dakota
Citation (AIRS-II Atlantic THORpex flights,
2003) - cloud microphysical properties
icing Pilot Reports (PIREPS) hourly over USA,
icing intensity height
Above increased values of re and LWP from GOES
generally correspond to icing from the Rosemont
and TAMDAR probes and large LWC from King probe.
Results from many other flights are similar to
the 5 March case. NCAR scientists have found the
GOES derived satellite products a useful
nowcasting tool for directing icing research
aircraft and for icing certification Wolff et al.
(2005)

Moderate icing PIREPS confirm satellite-derived
icing threat
Cloud properties and icing severity index Europe,
1100 UTC, 21 March 2007
INTEGRATION OF SATELLITE PRODUCTS INTO CIP
Satellite Derived Parameters Daytime Methodology
VISST Visible Infrared Solar-infrared
Split-window technique 0.65, 3.9, 10.8, 12.0 µm
Cloud amount (each pixel cloud or clear)
Cloud phase, optical depth, effective
particle size (re), ice or liquid water path
Cloud effective temperature (Tc) height,
top/base height pressure Shortwave
albedo, longwave flux for clear and cloudy
pixels Surface skin temperature for
clear pixels Nighttime Methodology, SIST
Solar-infrared Infrared Split-window technique
3.9, 10.8, 12.0 µm Same cloud parameters
as daytime except optical depth (OD) limited to
thin clouds only, no albedos
RGB image, snow deep pink ice clouds
gray-white-magenta, low clouds peach-white GOES
Icing Severity boundaries similar to CIP, with
some missing (CIP low probability) extra (thin
cirrus over warm water cloud) areas. GOES Icing
Severity corresponds well to PIREPS intensities
Probability of icing from PIREPS over eastern USA
as function GOES VISST LWP at 1545 2045
UTC, 1-11 February 2004
The Current Icing Potential (CIP) product
(Politovich et al. 2004) currently uses PIREPS,
RUC output, a simple satellite retrieval, and
surface data for assessing icing potential and
severity in near-real time. To improve the
assessment of severity, the CIP is being reworked
to include the LWP from the above products as
another variable in the CIP decision process
(Haggerty et al. 2005). This is a significant
effort toward integrating satellite cloud data
into nowcasting products used by the US aviation
industry.
DATA PROCESSING
Cloud detection, base, and top heights gt Ceiling
determination
Comparison of cloud detection from GOES ASOS
ceilometer data 0715 UTC October 7, 2004 (night)
March 2004
Comparison of cloud base heights from GOES
retrievals ASOS ceilometer data 1900 UTC, 18
March 2004
Basis for Icing Detection Aircraft icing
conditions depend on supercooled liquid water
(SLW) in cloud liquid water content, LWC
presence of large droplets, SLD VISST detects
SLW Tc lt 272 K, phase water liquid water
path LWP f(LWC)  effective radius, re
f(SLD) Develop threshold criteria from PIREPS
in situ data (Minnis et al., 2004) Preliminary
Icing Classification Algorithm No icing Tc gt
272 K, clear, or ice cloud with OD gt
6  Indeterminate ice cloud, OD gt 6 Icing
probability (IP) For re 5 µm, IP 0.147
ln(LWP) 0.084 (1) For re 16 µm, IP
0.138 ln(LWP) 0.024 (2) For observed re
, IP(re) fIP(5), IP(16) low, IP lt 0.4
medium, 0.4 lt IP lt 0.7 high, IP gt 0.7 Icing
severity (IS) IS light, if LWP lt 432 gm-2 IS
moderate-heavy, if LWP gt 432 gm-2
Points to right of line ztop too low or cloud
too thick for zbase estimate Points to left ztop
too high or overlapped cloud layers
Cloud base height errors well distributed about
mean error of -0.35 km, scatterplot ot right
shows source of largest errors
Top GOES-12 cloud mask overlaid with
individual ASOS stations gray shaded by cloud
amount Lower left GOES-12 IR temperature Lower
Right GOES 3.9-11 µm temperature
difference Cloud mask picks up most clouds
including much of the valley fog in the mountains
at night that is not evident in IR image
Above plots show ratio of hourly positive
overcast results from both ASOS GOES to total
ASOS overcast for March 2004 Top Day Bottom
Night GOES near 100 over most areas during day,
80-90 at night with worst results over high
mountain west coastlines
Comparison of stratus cloud boundaries from GOES
ARM cloud radar, 1997-2003
Detection of low cloud boundaries is most
important for ceiling determination. Right Mean
results comparing boundaries of overcast clouds
below 3 km at the ARM Southern Great Plains site
derived from combined cloud radar ceilometer
data and from GOES-8/10 data using Langley
algorithm. Average top height from radar is 0.1
km higher than GOES. GOES provides a better
measure of cloud temperature because of inversion
problems in soundings.
Conclusions Future Work

• Prototype probability-based icing algorithm now
  operational in near-real time over CONUS being
  incorporated into CIP
• - some tuning required to minimize
  errors in PIREPS, thin cirrus cloud height
  errors
• - need further nighttime testing tuning to
  estimate severity at night
• Continue analysis comparisons with aircraft
  in situ, surface radar, PIREPS data
• Investigate use of multi-layered retrieval
  algorithms to minimize indeterminate cases
• - microwave over water, thin-over-thick
  multispectral over all surfaces
• Investigate use of multispectral effective
  radius retrievals
• Examine potential for assimilating cloud
  properties into RUC analyses

Corresponding author email address
william.L.smith_at_nasa.gov Access to
products in image or digital form http//www-ang
ler.larc.nasa.gov/satimage/products.html