Alternative Assumptions about Topographic Height Improve Global Climate Model Simulation Fidelity

A key boundary condition in global climate models (GCMs) is topography. Typically, a high-resolution observed topographic dataset is placed on a GCM grid by averaging the observed topography over each model grid cell. For example, in a GCM with a 50 km horizontal resolution, observed topography is averaged over the relevant 50 x 50 km area and the topographic height of the entire grid cell is set to that averaged value. This process has the unfortunate effect of smoothing out topographic peaks, such that the model wind is steered far less by topography than in reality. Prior work has highlighted the important mechanical effects on atmospheric circulation of thin but high mountains, such as the Himalayas, Andes, and Sierra Madres. We hypothesize that the systematic smoothing of topography contributes to pervasive large-scale biases in GCMs by underrepresenting these mechanical effects. To test this, we perform a suite of experiments with the GFDL GCM CM2.5-FLOR, which has a 50 km atmosphere/land resolution, and with the relatively lower resolution CESM 1.0.5 GCM, in which we alter the model topography to allow for a more accurate representation of the effective maximum elevations of mountain ranges. With this alternative topographic assumption, improved atmospheric circulation leads to more accurate spatial distributions of precipitation, especially over North America and the Intertropical Convergence Zones. In this presentation, we will document these improvements and their dynamical causes, and discuss how this alternative and arguably improved topography might influence projected future climate.