Simplified Forecast-Model-Run-Collection Use and Analysis Using PyFerret

The development and analysis of predictive geophysical models typically involve repeated execution of the models with updated current conditions, then comparing past forecasts with current conditions and previous forecasts. Often, predicted values from each run of each model are saved in separated files, with the date of the forecast saved as offsets from the date the model was run. Aggregating and analyzing all these forecasts of multiple models can be quite cumbersome using ordinary data analysis packages.

PyFerret is the application, and Python module, that has evolved from Ferret, a widely used and recognized program, particularly in the oceanographic community, for interactive access, analysis, and visualization of data. PyFerret, and Ferret, easily handle the automatic aggregation and alignment of such forecast model run collections (FMRC), allowing simple analysis of behavior within a model as well as comparison of behaviors in different models. This presentation will explain and demonstrate PyFerret’s easy-to-use methods for performing an FMRC aggregation, analyzing and comparing the predictive skill of the models, as well as exporting the results as objects used in other Python packages such as netCDF-Python or xarray.