Document
P6.10 COMPARISON OF SATELLITE AND AIRCRAFT MEASUREMENTS OF CLOUD MICROPHYSICAL PROPERTIES IN ICING CONDITIONS DURING ATREC/AIRS-II Louis Nguyen*, Patrick Minnis NASA Langley Research Center, Hampton, VA, USA Douglas A. Spangenberg, Michele L. Nordeen, Rabindra Palikonda, Mandana M. Khaiyer Analytical Services and Materials Inc., Hampton, VA, USA Ismail Gultepe Meteorological Service of Canada, Toronto, Ontario M3H 5T4 Andrew L. Reehorst Glenn Research Center, Cleveland, OH, USA 1. INTRODUCTION as described by Minnis et al. (2004a). The data of interest include cloud-top height, phase, base height, Satellites are ideal for continuous monitoring of temperature, effective droplet radius, liquid water path, aircraft icing conditions in many situations over and icing potential or risk derived every 15 or 30 extensive areas. The satellite imager data are used to minutes. These pixel-level data are averaged along the diagnose a number of cloud properties that can be used aircraft flight tracks and matched as closely as possible.
to develop icing intensity indices. Developing and Three aircraft, the NASA Glenn Twin Otter, the validating these indices requires comparison with Canadian National Research Council (NRC) Convair- objective “cloud truth” data in addition to conventional 580, and the University of North Dakota (UND) Citation- pilot reports (PIREPS) of icing conditions. Minnis et al.
II, carried a variety of in situ sensors and flew a number (2004a,b) examined the relationships between PIREPS of icing missions. The Citation-II sensor complement icing and satellite-derived cloud properties. The included a set of FSSP and 2-DC probes as Rosemount Atlantic-THORPEX Regional Campaign (ATReC) and and TAMDAR icing probes. The TAMDAR is a low cost the second Alliance Icing Research Study (AIRS-II) field sensor developed by AirDat for NASA and is designed programs were conducted over the northeastern USA to measure and report winds, temperature, humidity, and southeastern Canada during late 2003 and early turbulence and icing from regional commercial aircraft 2004. The aircraft and surface measurements are (Daniels et. al., 2004). The TAMDAR icing data consist concerned primarily with the icing characteristics of of three indices: no ice, heater warning, and ice. These clouds and, thus, are ideal for providing some validation indices are compared with the GOES icing products.
information for the satellite remote sensing product.
Figure 1 shows the Citation-II flight track over and into This paper starts the process of comparing cloud the clouds over Montreal, Quebec from Bangor, Maine.
properties and icing indices derived from the The flight track is overlaid on the GOES-12 infrared Geostationary Operational Environmental Satellite image, which shows little variation in brightness (GOES) with the aircraft in situ measurements of temperature over Montreal.
several cloud properties during campaigns and some of the The comparisons include cloud phase, particle size, icing intensity, base and top altitudes, temperatures, and liquid water path. The results of this study are crucial for developing a more reliable and objective icing product from satellite data. This icing product, currently being derived from GOES data over the USA, is an important complement to more conventional products based on forecasts, and PIREPS.
2. DATA The satellite data consist of 4-km GOES-12 pixels with associated spectral radiances and cloud properties * Corresponding author address: Louis Nguyen, NASA Fig. 1. UND Citation flight track on on GOES-12 infrared Langley Research Center, 21 Langley Blvd, MS 420, image, 1815 UTC, Nov. 30, 2003. Magenta squares indicate Hampton, VA 23681-2199. email: l.nguyen@nasa.gov. Mirabel and Bangor airport.
AMS 11th Conf Aviation, Range, and Aerospace. Hyannis, MA, October 4-8, 2004 Fig. 2. Effective droplet radius from GOES-12 with flight tracks of UND Citation, 30 November 2003. (a) 1745 UTC, (b) 1815 UTC, (c) 1845 UTC, (d) 1915 UTC.
The NASA Glenn Research Center Twin Otter and phase averaged over 4 pixels around the pixel Canadian NRC Convair-580 also carried a variety of corresponding to the aircraft location. The cloud phase probes including the King liquid water content and determined from the satellite (top of Fig. 3) indicate that Rosemount icing probes. the area is clear at the beginning of the flight while supercooled liquid clouds occur either at or below flight 3. RESULTS level for most of the flight. The icing risk (panel 2, Fig.
3) is also relatively high for much of the flight. The 3.1 Citation Flight, 30 November, 2003 passes over Mirabel are evident in the variation of re (third panel of Fig. 3) from 9 μm up to 19 μm or greater The 30 November 2003 icing mission presents a and back again several times between 1730 and 1920 unique opportunity to validate the NASA Langley’s Icing UTC. The effective cloud height appears to be relatively Products. Murray et al. (2004) provide a brief discussion constant around 3 km during most of the flight. This lack of the flight, which is examined in more detail here. The of variability is consistent with the nearly uniform Citation flight track is plotted over the most closely appearance of the clouds in Fig. 1.
matched images of the derived effective droplet radii re Figures 4 and 5 show the satellite derived icing given in μm. In Fig. 2a, the Citation spirals down into potential compared with the Citation Rosemount and the cloud deck passing through clouds with relatively TAMDAR icing indicators, respectively. The Rosemount large values of re (> 15 μm) then proceeds into an area data are flagged as icing if the change in the voltage is with smaller values as it begins to fly back and forth greater than 1.5 over a 2.6 min interval. The satellite over Mirabel Airport in Montreal. The clouds with large retrievals clearly show high icing probability was values of re move northeastward during the Citation detected during the first hour of flight where the Citation passes over Mirabel (Figs. 2b-d). The cycling of the climbed and maintained an altitude of over 9 km during aircraft through different cloud regimes is more readily the transit from Bangor to Mirabel. Satellite-derived apparent in Fig. 3, which shows the detected cloud cloud base and top were calculated at around 0.5 and 3 AMS 11th Conf Aviation, Range, and Aerospace. Hyannis, MA, October 4-8, 2004 Fig. 7. NASA Glenn Twin Otter flight on Nov 25, 2003 over Mirabel, Canada. Blue and red color denotes light and med- high icing probability from GOES respectively.
km, respectively. As the Citation made a spiral descent over the Mirabel runway, the Rosemount started detecting icing at 3 km while the TAMDAR picked up icing at ~2 km during the descent. Both sensors performed optimally and agree well with satellite measurement for most of the flight. However, at about 1818 UTC, the satellite shows a small gap where no icing was detected for a period of 15-20 min of flight.
Both the TAMDAR and Rosemount detected this gap Fig. 3. GOES-12 derived cloud products matched along UND but 10-15 minutes later. This lag needs further analysis.
Citation flight track, 30 November 2003.
3.2 Twin Otter Flight, 25 November 2003 The Twin Otter flew a coordinated icing mission with the UND Citation during 25 November 2003. The Twin Otter flew below 3 km in conditions with light-medium icing probability as determined from the satellite observations (Fig. 6). Figure 7 shows the GOES- derived icing probability compared with the Twin Otter Rosemount probe. For most of the flight where the Twin Otter flew below 2 km, the Rosemount detected icing whereas the satellite detected icing for the entire flight.
Satellite cloud base and top appear to be a little too high. Ceilometer measurements from weather stations Fig. 4. Comparison of GOES icing probability and Rosemount in the vicinity of the flight track revealed that the cloud icing probe, 30 November 30, 2003. Thick red line indicates bases were in the 1-2 km range (figure not shown), a icing from the Rosemount probe; two thin lines denote boundaries of cloud estimated for GOES.
Fig. 7. Same as Fig. 4, except for Twin Otter Rosemount Fig. 5. Same as Fig. 4, except for TAMDAR.
probe, 25 November 2003.
AMS 11th Conf Aviation, Range, and Aerospace. Hyannis, MA, October 4-8, 2004 Fig. 9. Same as Fig. 4, except for Canadian NRC Convair-580 Rosemount probe, 6 February 2004.
like to thank G. A. Isaac and research teams of the Cloud Physics and Severe Weather Meteorology Fig. 8. Comparison of GOES-12 and NRC Convair-580 FSSP Division of the Meteorological Service of Canada effective radius (in μm), 6 February 2004.
(MSC), the Canadian National Research Council (NRC) Convair-580, David Delene and Tony Grainger of finding consistent with the results between 1730 and University of North Dakota, and Tom Ratavsky (NASA 1850 UTC. The rise in the GOES cloud heights around Glenn Research Center) for proving in-situ data.
1900 UTC needs further investigation. Overall, the icing from the GOES is located in the correct location REFERENCES horizontally although there is some apparent vertical error.
Daniels, T. S., G. Tsoucalas, M. Anderson, D. Mulally, W. Moninger, R. Mamrosh, 2004: Tropospheric 3.3 Convair-580 Flight, 6 February 2004 Airborne Meteorological Data Reporting (TAMDAR) 3.4 Sensor Development. Proc. 11th Conf. on Aviation The Convair-580 FSSP-derived effective droplet Range, and Aerospace Meteorology, Hyannis, MA, radii are compared in Fig. 8 with the matched GOES Amer. Meteor. Soc., Hyannis, MA. October 4-8.
values during the Convair flight during 6 February 2004.
Minnis, P., et al., 2004a: Real-time cloud, radiation, and The GOES re values are slightly larger than their FSSP aircraft icing parameters from GOES over the USA.
counterparts near the beginning of the flight and th 13 AMS Conf. On Satellite Oceanography and Met., decreased throughout the flight. They appear to be in Norfolk ,VA, September 20-24.
good agreement with the FSSP data after 1800 UTC.
Minnis, P., W. L. Smith, Jr., L. Nguyen, D. A.
The differences early in the flight could be due to the Spangenberg, P. W. Heck, R. Palikonda, J. K. Ayers, presence of ice in the otherwise supercooled cloud. Ice C. Wolff, and J. J. Murray, 2004b: Near-real time particles cause an overestimate of re in the GOES cloud properties and aircraft icing indices from GEO retrieval. The icing from GOES (Fig. 9) is coincident and LEO satellites. Proc. 49th. SPIE Annual Meeting, with much of icing indicated by the Rosemount icing Earth Observing Systems IX Conference, Denver, detector except near the end of the flight. Again there CO, August 2-6 are some discrepancies in the cloud top heights.
Minnis, P., L. Nguyen, R. Palikonda, D. A. Spangen- berg, J. K. Ayers, Y. Yi, M. L. Nordeen, and C. Wolff, 4. CONCLUDING REMARKS 2004c: Toward a three-dimensional near-real time cloud product for aviationsafety and weather These preliminary results indicate that the GOES diagnoses. Proc. 11th Conf. on Aviation Range, and icing detection is generally consistent with the objective Aerospace Meteorology, Hyannis, MA, Amer. Meteor.
icing detection probes on three different aircraft. GOES Soc., Hyannis, MA. October 4-8.
only “sees” the tops of the clouds directly, so the Murray, J. J, P. R. Schaffner, P. Minnis, L. Nguyen, V.
occurrence of icing below cloud top must be inferred.
E. Delnore, and T. S. Daniels, C. A. Grainger, D.
Additional analyses using a three-dimensional estimate Delene, C. A. Wolff, 2004: Proc. 11th Conf. on of aircraft icing from the satellites (Minnis et al. 2004c) Aviation Range, and Aerospace Meteorology, should provide for a more direct comparison between Hyannis, MA, Amer. Meteor. Soc., Hyannis, MA.
the satellite and the aircraft, which samples at different October 4-8.
altitudes. The satellites are sometimes too high or low.
Gultepe, I., and G. A. Isaac: An analysis of cloud droplet number concentration (Nd) for climate Acknowledgments studies: Emphasis on constant Nd. Q. J. Royal Met.
The research was support by the NASA Advanced Soc., in print.
Satellite Aviation Products Program. The authors would AMS 11th Conf Aviation, Range, and Aerospace. Hyannis, MA, October 4-8, 2004