SIParCS 2021 - Heather R. Craker

Climatology Calculation Support in the GeoCAT Ecosystem

Recorded Talk

As NCAR progresses in its efforts to transition away from using the NCAR Command Language (NCL) to Python for scientific programming, much needs to be done to ensure that Python has all of the same visual and computational functionalities that exist in NCL. The Geoscience Community Analysis Toolkit (GeoCAT) team has been working on recreating NCL functions in pure Python in their geocat-comp package. This summer, the geocat-comp function climatology_average() was developed to replicate many of NCL’s climate average calculations. This function works by grouping data along the time dimension and then taking the average of each group. The groups are created across all years in the given dataset to yield long term climate averages. For example, a multiyear dataset can be grouped and averaged by month resulting in the number of timestamps being reduced to just 12 with each timestamp corresponding to the average of the data points in January, February, March, and so on regardless of which year those data were in. This function can group data by hour, day, month, or season and is flexible enough to work with data that uses non-standard calendars, which is common with atmospheric model output. Functions to calculate climate anomalies, remove annual cycles from data, and more still need to be added to the GeoCAT ecosystem, but the hope is that these easy-to-use functions will make doing climate model diagnostics work in Python easier.

Mentors: Alea Kootz & Orhan Eroglu

Slides and poster