JEDI SkyLab can now conduct experiments with newly-published data, adding near-real-time (NRT) data assimilation to its suite of accomplishments.
For those running an individual experiment using NRT data, the system is highly configurable; you can choose only the observation types you want in the experiment and feed them in, a task that used to involve exponentially more labor. Some of the observation types currently available include AMSUA, GMI, GOES ABI, TEMPO, PrepBufr Aircraft, PrepBufr Radiosonde, GNSS-RO and TROPICS. More observation types are being added all the time, as the observation team adds new data converters to put files in our IODA format and the infrastructure team integrates those new observation types into the SkyLab workflow. Both GFS/UFS and MPAS backgrounds can be used for NRT experiments in SkyLab.
Perhaps the best example of this functionality is the recent integration of TEMPO NO2 products into the JEDI data assimilation system less than 24 hours after the official product release. This collaboration with NASA GMAO demonstrated a real-world application of NRT data assimilation with important new data–NO2 is a significant pollutant and can be used as a tracer for general air quality. The ability to integrate new sensors incredibly fast and run experiments with up-to-date data will be instrumental in improving accuracy for future weather and climate forecasting.
In addition, a Continuous Integration (CI) test runs every night using the previous day’s data stored in R2D2; the CI executes a full ingest experiment that can be seen in the github viewer while in progress and then on experiments.jcsda.org. The nightly CI test works to collect, convert, store, and publish observations from the previous day. It runs from a new JEDI build every time, and builds the observation database that can be used in SkyLab experiments. With this database and the NRT functionality, users can easily run experiments with NRT data.
This ability in JEDI SkyLab is an important step for continuous data assimilation capabilities, and boosts the usefulness of SkyLab for our partners and public users!
Photo by Kevin Quezada on Unsplash