I am a sixth-year PhD student working with Dr. Susan Solomon at Massachusetts Institute of Technology. My PhD thesis is on detecting and understanding processes affecting stratospheric ozone variability. I did my undergraduate study at the University of Wisconsin-Madison, double majored in Atmospheric & Oceanic Sciences and Applied Mathematics, and with a certificate in Computer Science. During my undergraduate, I was advised by Dr. Tracey Holloway using satellite and ground-based measurements to study atmospheric formaldehyde.

Man-made chlorofluorocarbons (CFCs) are the main reason for the Antarctic ozone “hole”. Under the Montreal Protocol, the production of these molecules has been regulated since 1990s. Hydrochlorofluorocarbons (HCFCs) and hydrofluorocarbons (HFCs) were later used to replace CFCs. They are potent greenhouse gases, and their productions are also under regulation. It is important to understand all the loss pathways of these molecules to better estimate anthropogenic emissions and assess the global compliance to the Montreal Protocol.

The ocean uptake of these molecules was long been thought as a minor loss pathway. As anthropogenic emissions went down, natural losses become more important. We used a hierarchy of models to study the ocean uptake of CFCs, HCFCs, and HFCs and assess the impact on emission estimations.

Related work: Wang et al. (2021); Wang et al. (2023)

Heterogeneous chlorine activation is a major driver for stratospheric ozone depletion and is understood to happen mainly on polar stratospheric clouds (PSCs) at temperatures below about 195 K. The 2020 Australian wildfire released large amounts of organic aerosols, whose chemical properties under stratospheric conditions are virtually unknown.

We combined the past 30 years of satellite data to estimate chlorine activation after a series of volcanic eruptions and wildfires of different magnitudes. We found that chlorine activation after major wildfires can happen at warm mid-latitude temperatures even above 220 K. Model incorporating such mechanism also shows remarkable agreement with the observations.

Related work: Wang et al. (2023); Solomon et al. (2023); Zhang et al. (2024); Wang and Solomon (2024);

Layer-Wise Relevance Propagation (LRP) offers a way to interpret the neural networks, which is generally thought as a “black box” process. We applied the LRP technique to understand the physical processes in subgrid convection parameterization (Wang et al., 2022) and in Atlantic Multidecadal Variability predictions (Liu et al., 2023).