Description
Hailstorms inflict significant economic damage, primarily through property and crop losses. This dataset tracks the historical baseline rate of hail storms across the globe for terrestrial regions, providing valuable insights for understanding and mitigating the impact of hail events. This information is crucial for assessing property risk, informing insurance pricing, guiding infrastructure design, and optimizing agricultural practices.
Technical Description
The global Hail Historical Baseline (v1) dataset estimates the annual frequency of severe hailstorm occurrence, as retrieved from satellite-borne passive microwave imagery. These data products can be useful for weather and climatological research related to storms, as well as applications involving risk management and emergency management. This data is derived from the multi-frequency (37 GHz and 19 GHz) passive microwave estimation of the probability of hail, accumulated over the TRMM and GPM domains and normalized for overpass counts (Bang and Cecil, 2019; Bang and Cecil, 2021).
GPM or TRMM Polarization Corrected Temperature (PCT) features were paired with ground hail reports to train a hail retrieval algorithm to estimate the probability of hail. Probability of hail is calculated for each GPM (or TRMM) feature using the 37 GHz and 19 GHz PCT retrieval (Bang and Cecil, 2019). "Features" are contiguous areas larger than 1 GPM GMI (or TRMM TMI) pixel with 89 (or 85) GHz PCT < 200 K. Features must pass the surface artifact/deep convection screen (see Bang and Cecil, 2021) to be counted. Features with a minimum estimated probability of hail > 20% are counted in 2-degree boxes. The total of these individual hail probabilities is tallied for each box, and normalized for overpass counts and box area (see Bang and Cecil, 2019, Equation 5).
Between 39 degrees north and south latitudes, the climatology is from combining both the TRMM and GPM hail probabilities and overpass counts, normalized for orbit and sampling. Beyond 39 degrees, the climatology is solely from the GPM satellite. 37-GHz and 19-GHz PCT values from the two satellites are adjusted for internal consistency before computing and combining the hail probabilities. The PCT adjustment is described in the Appendix of Bang and Cecil (2021).
The effectiveness and regional bias of this hail retrieval was tested using well-known spaceborne radar-based hail proxies (such as reflectivity at ?20°C) (Bang and Cecil, 2021). The hail retrieval exhibits little regional variability, even when examining regional reflectivity profiles at a fine scale. It shows improved performance in oceanic regimes (Bang and Cecil, 2021).
In this layer, values over oceans are masked, and the risk class layer is calculated using percentiles of terrestrial values only.
This dataset contains the following fields:
- norm_hail_density: Estimate of the number of severe hailstorms per 10,000 sq. km per year at pixel locations over the historical baseline used for calculation. For instance, a value of 1.0 would indicate one severe hailstorm per year within a 100 km square area at that location. This is a value between 0.0 and 6.32, the global maximum.
- hail_risk_class: The hail risk classification for each value. Values are integers from 1 to 5 representing the following classes: Very Low Risk (1), Low Risk (2), Moderate Risk (3), High Risk (4), and Very High Risk (5). Values are categorized using the 44th, 67th, 84th, and 94th percentiles of the norm_hail_density field for terrestrial values.