A Simplified-Physics Atmosphere General Circulation Model for Idealized Climate Dynamics Studies

Abstract Simplified global atmospheric general circulation models are important tools for understanding basic climate dynamics and diagnosing more complete modeling systems. We have developed a simple, single-model framework with nonlinear primitive equations that can be run via an interactive Python Jupyter notebook. This framework was developed as an educational tool for students to experiment with the model, as it is straightforward to impose forcing, postprocess data, and visualize results in one notebook. However, experienced scientists can also use the model for original research. We have implemented three distinct innovations. First, the underlying dynamic core is entirely based in Python (as opposed to FORTRAN with a Python wrapper), making the model more accessible to both undergraduate and graduate students. The model can easily be run on a laptop without concerns such as compiler options or operating systems, thus removing common technical barriers to access. Second, the background state can be specified within the same framework as “strong,” where the model equations have been modified into an anomaly model, or “weak,” where the original full-field model is relaxed to the background state. Third, “forcing” is described in terms of prescribed latent heat release in the troposphere, the topography, and the background state. Forcing and resolution changes are easily implemented using one preprocessing script that supports a wide range of hypothesis-driven experimentation options. This paper includes an exploration of the model’s mean state, some sensitivity experiments, and an example diagnosis of El Niño–Southern Oscillation (ENSO) teleconnections. Significance Statement Simplified atmospheric computer models are important tools for basic understanding of how the climate system works. We have developed a simple modeling framework that was developed as an educational tool for students to design and implement hypothesis-testing experiments and, importantly, is entirely written as a Python Jupyter notebook so that it is easily accessible to students and others new to modeling. However, the model can also be used by experienced scientists for original research. This manuscript describes the modeling framework, including how to obtain the model and some examples of the model’s capabilities.