Comparing satellite and ground based observations of cloud over the Southern Ocean.

• Main Author: McErlich, Cameron
• Format: Other/Unknown Material
• Language: English
• Published: University of Canterbury 2020
• Subjects: Southern Ocean
• Online Access: https://hdl.handle.net/10092/100746; https://doi.org/10.26021/946

Clouds are poorly represented in climate models. This has been attributed to factors that include simulating too little cloud cover and under-representing the amount of supercooled liquid water in clouds. This leads to biases the cloud radiative effect, which in turn causes a positive shortwave radiation bias over the Southern Ocean, where too much sunlight is hitting the surface of the ocean. This thesis presents an analysis of the 2BCL5 and DAR- DAR satellite datasets, as well as comparison with the ground based AWARE dataset. This work was undertaken with the aim of collecting and comparing satellite and ground based observations of cloud to develop a representation of the vertical structure of clouds and their phase over the Southern Ocean. Comparisons between 2BCL5 and DARDAR found that the two display differences in the amount of cloud observed. 2BCL5 detects more clouds than DARDAR, except below 1 km where DARDAR shows a greater amount of cloud. The two also show differences in their partitioning between the ice, mixed and liquid cloud phases. 2BCL5 always detects more mixed phase cloud while DARDAR mostly classifies as either ice or liquid phase. This was found to be due to differences in the way the datasets classify cloud; 2BCL5 will generalise a whole cloud layer as mixed phase cloud if it detects both ice and supercooled liquid water while DARDAR will classify the parts of cloud that are ice, liquid, and mixed phase separately. Comparisons between 2BCL5/DARDAR and AWARE find that 2BCL5 matches better with AWARE than DARDAR does. Between 1.5 km and 4.5 km 2BCL5 agrees with AWARE while DARDAR shows a greater spread. These differences were found to be statistically significant. Below 1.5 km neither satellite dataset matches better with AWARE, even though DARDAR sees more cloud below 1km than 2BCL5. DARDAR seeing more cloud can likely be attributed to a greater amount of false positives where DARDAR is classifying noise in the radar signal incorrectly as cloud. Above 7 km neither 2BCL5 or DARDAR does a ...