Decadal changes in thin cirrus height measured by MISR

Abstract

Thin cirrus clouds are difficult to detect and the operational cloud-height product from the Multiangle Imaging Spectro Radiometer (MISR) may miss thin cirrus when it overlies thicker, lower-level cloud. To detect thin cirrus over clear skies with MISR, cloud-heights above 10 km were retrieved using a stereo matching technique on the oblique camera pairs and re-projected onto the nadir camera after correction for parallax and winds. A merged dataset, including measurements from the CERES broadband radiometer, was used to compare measured outgoing longwave radiation (OLR) to the OLR from a column model of longwave radiative transfer that uses MISR cloud-top heights and reanalysis data as input parameters. After accounting for uncertainties in modeled OLR of ± 4 W m−2 due to uncertainties in the input properties, especially for known high cloud, surface temperature and specific humidity, the average difference of 17 W m−2 cannot be directly explained, and is accounted for by the addition of thin cirrus with a coverage of 60%. The analysis of tropical MISR orbits with standard processing shows a decline in cloud-top height since 2000. After validation of oblique-detected thin cirrus with ground-based sensors, the additional cirrus detected over the tropics also showed a decline in cirrus-top height. The decadal changes and climatological comparisons in cloud-top height of thin cirrus in the tropics are also reported.