Kay Suselj

Data Scientist
398K - Science Data Modeling and Computing Group
Jet Propulsion Laboratory, California Institute of Technology

  • Physical parameterization in atmospheric models
  • Stochastic parameterizations and its impact on the evolution of model fields
  • Understanding of turbulence and convection in the geophysical fluids
  • Interaction between physical processes and dynamics in atmospheric models

Atmospheric models are based on solid physical principles. Nevertheless, for computers to solve them efficiently, for example to make a weather forecast or predict the future climate, certain simplifications are inevitable. These simplifications neglect smaller scale motions, which are instead represented by parameterizations. How accurate and complex do these parameterizations have to be? Can we develop better predictive models and theories for unresolved motions? How do unresolved motions interact with other physical processes, such as condensation and evaporation, atmospheric radiation and microphysics? These are some of the overarching questions that my research tries to address. For this, our team utilizes data from turbulence and cloud resolving models and observations. We are developing a parameterization, based on the Eddy-Diffusivity/Mass-Flux approach, which can be used in any climate or weather prediction model. I collaborate with weather prediction and climate centers to test the skill of our parameterization in their models.