Date of Award


Document Type


Degree Name

Master of Science (MS)


Atmospheric Sciences

First Advisor

Xiquan Dong


Passive satellites often face difficulty when detecting clouds over snow and ice covered surfaces beneath them. The recent launches of active satellites, which directly measure cloud properties, have allowed scientists to gain a firsthand look at the complex cloud profiles across polar regions. To help quantify the differences between passive and active satellite retrievals, cloud properties derived for the Clouds and Earth's Radiant Energy System (CERES) project using MODerate Resolution Imaging Spectroradiometer (MODIS) data are compared with combined measurements from Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) and CloudSat (CC), and Department of Energy (DOE) Atmospheric Radiation Measurement (ARM) observations at the North Slope of Alaska site, from July 2006 to June 2010. The study was then extended to include the entire Arctic and Antarctic.

During the 4-year period, monthly mean cloud fractions (CFs) between ARM and CC differ by 5%. While CERES-MODIS CF retrievals agree well with ARM and CC during warm months (May-October), retrievals during the cold season (November-April) significantly underestimate CF. Annual mean cloud-base heights derived from ARM and CC agree within 200 m, while their cloud-top heights (Htop) differ by an average of 1.2 km, due largely to CC detecting more upper-level clouds during the warm months. Effective cloud heights from CERES-MODIS retrievals fall between CC and ARM cloud bases and tops, as expected.

Cloud fractions and heights across the span of the Arctic depict similar features as those shown at the ARM NSA site. During summer months, cloud fractions between CERES-MODIS and CC agree well, differing by no more than 10% across most regions of the Arctic. During this same season, however, cloud heights vary by as much as 5.2 km. This is largely due to multi-layer cloud systems, where CC measures the uppermost cloud layer, and CERES-MODIS detects lower cloud layers, resulting in lower CERES-MODIS cloud heights, but equal cloud fractions. Winter shows a contrast to the similarities in cloud fraction detected in summer, with CERES-MODIS underestimating CF by as much as 59%. Cloud heights between the two platforms, however, show better agreement during the cold months, when fewer high clouds occur.

The largest differences in CERES-MODIS and CC cloud fractions and heights occur during the cold season (JJA) in the Antarctic. During this time period, CC detects cloud fractions as much as 43% higher than CERES-MODIS, over regions coupled with cloud heights up to 12.3 km higher than CERES-MODIS. These extreme differences are caused by the presence of polar stratospheric clouds, which occur at altitudes between 15 and 25 km, and are nearly impossible for the CERES-MODIS sensor to detect.

Finally, single-layered low-level stratus cloud effective radius (re), liquid water path (LWP), and optical depth (τ) retrieved from CERES-MODIS and surface-based retrievals at the ARM NSA site were investigated. When surface snow and sea ice are not present, ARM and CERES-MODIS retrieved cloud droplet re, LWP, and τ agree well. However, when snow and sea ice are introduced, CERES-MODIS retrieved re values are higher than ARM results, while optical depths are lower. These differences suggest that CERES-MODIS cloud fraction retrieval algorithms during polar night and microphysical retrieval algorithms over snow and ice covered surfaces need future improvement.