My research focuses primarily on understanding and modeling atmospheric processes that are too small to be resolved by climate models, specifically atmospheric boundary layer processes and moist convection. These processes are one major source of uncertainty in climate model projections, specifically for lead times of about 30-40 years. I am utilizing machine learning combined with physics to improve the modeling of various processes such as the vertical flux of heat and pollutants. I believe incorporating physics into machine learning models is essential to improving their robustness and discovering new physics. By combining machine learning with a deep understanding of the underlying physics, I develop more accurate and robust models that can lead to new discoveries and advancements in the field of atmospheric science. Click here to read more on this topic.

Moreover, I investigate the role of sub-grid-scale (scales smaller than 100km that are not resolved by a climate model) on the emergent behavior of a system or process. This part of my research is particularly exciting because it has the potential to make significant contributions to our understanding of how to better model unresolved processes with important influence on the Earth’s climate and weather patterns. Click here to read more.

In addition to my work on the boundary layer and structure, I have also studied the cloud responses to a coupled ocean-atmosphere and the impact of ocean temperature anomalies on atmosphere circulation and transport. I believe that understanding the interactions between the ocean and atmosphere is crucial to predicting weather patterns. See the last link for more.

I am also interested in extreme events and occasionally work on side projects related to extremes. The topic of extreme events becomes even more fascinating when machine learning is applied to detect, predict, and explain them..

Atmospheric Boundary layer modelling

The Emergent Impact of Mesoscale Patterns

Cloud Responses to Ocean-Atmosphere Coupling