Liam Thompson

Liam Thompson, University of Oklahoma/University of Utah

A Generalized framework for the evaluation and comparison of atmospheric chemistry models with observations

Recorded Talk

Our ability to observe Earth system processes, ranging from the production of ozone to the overturning of oceanic circulations is ever-increasing. Yet, our ability to integrate different types of observation systems to further understand Earth system model biases and weaknesses remain fragmented. Atmospheric chemists can observe pollutants through satellite observations, surface-based sensors, and aircraft measurements. These observations have different temporal and spatial strengths that range from limited density of surface observations to satellite retrieval uncertainties due to the interference of clouds. This complexity makes the unification of these observation types difficult when conducting model evaluations. Regardless, leveraging and integrating these observations in Earth system model evaluations is crucial to fully understanding model biases.

Thus we showcase an open-source, model- and observation-agnostic, Python-based model evaluation toolkit developed by the National Center for Atmospheric Research and National Oceanic and Atmospheric Administration known as MELODIES-MONET (Model EvaLuation using Observations, DIagnostics and Experiments Software (MELODIES) Model and ObservatioN Evaluation Toolkit (MONET)). MELODIES-MONET can play a crucial role in unifying model evaluation workflows that leverage surface, satellite, and aircraft observations. Previously, MELODIES-MONET only supported fixed latitude longitude model grids such as Weather Research and Forecasting with Chemistry and Community Earth System Model (CESM) Finite Volume. With UXarray, MELODIES-MONET now supports the evaluation of unstructured model grid output. Such examples include the Model and Prediction Across Scales (MPAS) project and CESM Spectral Element. Finally, we highlight planned future applications of MELODIES-MONET in evaluating model output temporally aligned to periods of NSF airborne field studies.

Mentors: Ben Gaubert, Orhan Eroglu, Jerrold Acdan

Slides and poster