Publication Date

Summer 2023

Degree Type


Degree Name

Master of Science (MS)


Meteorology and Climate Science


Minghui Diao; Eugene Cordero; Xiaohong Liu


Predicting future climate change relies on accurate representations of the earth’s surface radiation budget. Complex interactions between large-scale dynamical conditions (e.g., low-pressure systems) and microscale processes (e.g., cloud properties) are key contributors to energy budget biases within global climate models. Proper estimation of cloud microscale processes and their responses to synoptic-scale dynamics will greatly improve the accuracy of the simulated energy balance within global climate models. High-latitudinal clouds have significant influences on the Earth’s radiative balance. We examined observations from two field campaigns at Macquarie Island in the Southern Ocean and McMurdo Station in Antarctica. Two global climate models are evaluated for their representations of cloud and radiative properties and the dynamical effects. A cyclone compositing method is used to define four quadrants relative to the extratropical cyclone centers over the Southern Ocean and quantify “dynamics-cloud-radiation” relationships. Observations at Macquarie Island show similar cloud fraction and liquid water path (LWP) between northeast quadrant (i.e., frontal warm sector) and northwest quadrant (i.e., post-frontal cold sector). In contrast, observations at McMurdo show higher cloud fraction and higher LWP in southwest quadrant than southeast quadrant, indicating larger dynamical influences on cloud properties at higher southern latitudes. Observed net surface radiation in shortwave and longwave (SW and LW) both show statistically significantly higher values on the westside of the extratropical cyclones compared with the eastside at two locations. Two climate models show more significant cloud and radiation biases at Macquarie Island compared with McMurdo, including overestimation of LWP and opposite correlations between net surface LW radiation and cyclone positions, indicating that climate models have larger radiative biases at lower southern latitudes than high southern latitudes. The net surface radiation biases are affected by multiple controlling factors, including cloud fraction, LWP, seasonal variability, and the relative positions to the extratropical cyclones, indicating that cloud microscale processes and their responses to synoptic-scale dynamics both contribute to the accuracy of simulated energy balance. Overall, this study will help to improve future development of cloud parameterizations and ultimately increase the fidelity of future climate prediction.

Included in

Meteorology Commons