Retrieval of Cirrus Radiative and Spatial Properties Using Coincident AVHRR and HIRS Data

Among the factors determining the efect of clouds on climate are cloud fraction, altitude, and emissivity. To properly assess the impact of clouds on climate, global measurements of these properties are needed. To this end, the International Satellite Cloud Climatology Project (ISCCP) was undertaken to measure global cloud properties suitable for assimilation in climate models. As part of the First ISCCP Regional Experiment Cirrus Intensive Field Observations (FIRE Cirrus IFO), coordinated measurements were made of cirrus clouds from satellite, aircraft, and ground-based platforms. We compare the results of cirrus property retrievals from two different algorithms: the High-Resolution Infrared Sounder (HIRS) CO2 slicing technique, and the Advanced Very High Resolution Radiometer (AVHRR) Infrared Cirrus (AIRC) model.

HIRS measurements, with a spatial resolution of 25 km, can be processed to retrieve the effective cloud emissivity Ne, which is the product of the cloud fraction N and the cloud emissivity e. The CO2 slicing method applied to HIRS data also yields estimates of cloud height. The AIRC model yields direct estimates of cirrus emissivity and temperature, which can be converted to height by comparison with coincident profile measurements. Additionally AVHRR data can be used to identify cloud fraction N at a relatively high spatial resolution (1 km) using an infrared thresholding technique.

The average radiative and spatial attributes obtained by the CO2 slicing and AIRC models compare well for the cirrus clouds observed during the NOAA overpass on 28 October 1986. The overall mean cirrus emissivities and altitudes observed by the CO2 slicing and AIRC models agree well with each other at 0.39 / 0.34 and 6.9 / 6.4 km respectively. However, the effective emissivity identified by HIRS can not always be used in conjunction with AVHRR-derived cloud fraction in order to retrieve emissivity values from HIRS observations. The reason is that the separability of effective emissivity Ne into separate cloud fration and emissivity components assumes that cirrus properties are homogeneous within a HIRS Field of View (FOV), which is not always the case. But when the high-resolution data indicate that cirrus properties are roughly uniform within a HIRS FOV, AVHRR-derived cloud fractions help to estimate cloud emissivity directly from HIRS CO2 slicing results. Thus the two models can be used in conjunction with one another to increase the accuracy of retrieved cirrus properties.