This Python application is designed to process and animate hail data based on the NEXRAD Level 3 hydrometeor classification product. It uses a mix of libraries like MetPy, Matplot, Xarray, etc., to manipulate data available from NOAA's near-realtime radar data feeds.
Example total hail map generated from several hours of radar data around Omaha, NE, US:
The hail index represents a combination of hail severity and hail duration at a given point.
The NEXRAD hydrometeor classification product specifies 3 levels of hail:
- Hail (<1 in.)
- Large Hail (1-2 in.)
- Giant Hail (≥ 2in)
(For more info, see Hydrometeor Classification Algorithm 2 (HCA2) Overview and RDA/RPG Build 17.0 Training.)
This application simply uses a values of 1, 2, or 3 to represent the corresponding hail severity in each radar scan, and sums news value when subsequent scans indicate hail at the same location.
A future version may map these three categories separately to show a more granular report.
The index is not very determinstic in that, for example, an index of 9 could indicate small hail occured at a location in 9 radar frames, or giant hail occured 3 times. Nevertheless, it is a useful comparison of relative hail activity in an area.
Additionally, hydrometeors are only sampled at given radar scan intervals (about 5 minutes). Unfortunately, gaps or stripes can be observed in depictions of small or fast-moving storms. The patterns are usually obvious at least, and you can infer that hail was continuous even between the stripes. (Eventually maybe I'll try to implement some sort of frame interpolation.)