Cloud top temperature information is important for many types of weather forecasts, including aircraft icing. Automated icing algorithms developed at NCAR use GOES infrared temperature data matched to model temperature and humidity output to determine the cloud top temperature and cloud top height. Sounding data can also be used to determine these parameters. A comparison of cloud top temperatures using GOES-EAST satellite and derived sounding cloud top temperatures was conducted to determine if a correlation existed between the two. Canadian and NWS sounding soundings taken at 1200 and 0000 UTC during the year 2001 were matched to GOES-EAST satellite data from 1115z and 2315z (approximate sounding launch times).

GOES-EAST cloud top temperatures were calculated for a 25-pixel box centered over each sounding site. Using various GOES-EAST fields, each pixel location was identified as being clear or cloudy. The infrared field was then used to determine the satellite cloud top temperature for each cloudy pixel and the mean value of all 25 pixels was used for the cloud top temperature.

Sounding-based cloud layers were calculated using relative humidity with respect to both water and ice. Instances where the relative humidity values with respect to water or ice were 87% and above were used to specify a cloud layer. Soundings where derived cloud top temperatures were below -40 Celsius were excluded from the dataset due to the unreliability of the humidity sensor below -40 Celsius. Results are examined using 10 degree Celsius temperature bins and sounding-derived cloud layer thicknesses from less than 500 meters to greater than 2000 meters.

Agreement between the derived sounding and satellite cloud top temperatures was best for relatively thick, warm clouds. When looking at seasonal variations, certain seasons yielded better agreements than other seasons.