Applications of the New York State Mesonet with Urban High-Resolution Numerical Weather Predictions

To enable proactive, local decision making for diverse, weather-sensitive business applications, IBM has deployed numerical weather prediction models for over two decades. An important element has been the use of high-quality local observations from public and private sources for both data assimilation and to assess the model fidelity. The former has been conventional observations from NOAA's Meteorological Assimilation Data Ingest System. Private data have included weather stations operated by collaborators as well as our own efforts. However, there typically remains a significant gap at the intermediate mesoscale. The New York State Mesonet (NYSM) was designed specifically to address such an observational gap.

Therefore, we have enabled automated utilization of near-real-time conventional observations from the NYSM for our operational execution of a customized version of the community WRF-ARW weather model in New York State. This includes configurations for the New York City metropolitan area as well as upstate lake watersheds. We will illustrate the value of the NYSM data with a case study for New York City. There are approximately 20 sites within the current 667m WRF nest and the rest of the sites are included within the lower-resolution, outer nests of the WRF configuration.

On Wednesday, 1 September 2021, the remnants of Hurricane Ida led to record-breaking rainfall in the New York City metropolitan area, including over 200mm in only a few hours, resulting in subway platforms and streets looking like rivers, and over a dozen fatalities from drowning. For the first time, the National Weather Service issued a flash flood emergency for New York City: "This is a life-threatening situation." "Seek higher ground now!"

Despite such clear communications many of the impacts may have been avoided with longer prediction lead times. We examine this idea by discussing the data generated by the aforementioned deployment of WRF. Given that data from the NYSM illustrated the localized characteristics of the storm, we also use such data to validate this model.

We will present an overview of the case study event, the approach to the modelling and the results to illustrate how this scale of modelling leveraging the data from a dense, regional mesonet can provide improved guidance prior to such extreme events.