Datameteo Hail forecasting index model

 
Datameteo will be at the Meteorological Technology World Expo 2014 in Bruxelles on October 2014 21th.

The must-attend event for anyone looking for improved weather forecasting ability, technologies and services with a professional audience coming from various different sectors like: aviation, agro, energy, mobility  or enviroment.

In this innovative and competitive contest Datameteo at 1500 CEST weather will present a conference into the break out session area titled:

"A new composite hail forecast index: The Italian weather scenario"

In the latest years climate change has created serious difficulties who must make choices in the medium to long term. After a long testing work we have created a scalable hail storm index (Hail, composite index+Significant hail parameter) in house developed with the integration of various multiphysical parameterization and model schemes. A complete index validation forecast vs observation on the italian area has been made in order to better understand the hail season trend and improve the nowcasting and short term forecast accuracy.

A complete report will be put online the same day of the conference,  in order to keep a clear view on this new frontier of the meteorologhy: the hail index risk

At this stage i would like to thank to Ms Simona Busacca that in her final thesis for the Master in Climate Change at  the University of Exeter has applied and innovative analysis approach comparing the hail index forecast vs observations for the past season 2013. Three different evaluation indexes  : sensitivity, sensibility,  accuracy could help us to improve the fine tuning of the predictions

A huge analysis work how summarized by the image below IBIMET source that shows the hail, strong winds and flooding rain occurences density areas in Italy for the latest 10 years.. It's no easy to find complete and accurate data Archives for all the zones. Due to extreme events reports not available with the same consistency for all the country the maps below has been developed integrating reanalysis techniques too

Maps showing hail, flooding rains, and strong winds more than 120 Km/h occurences density area  IBIMET source

The first source keep in account was the observed data  archives from official Italian Military Air Force's weather stations (MeteoAM) and Italian Agency for Air Navigation ( ENAV). There are about ninety stations and they are spread all over the country, allowing both ground and atmospheric monitoring.

 

Each weather station covers a radius of 12 km, and send an amonut of data/reports during the operational hours. The dataset was limited to the period of time in question, and several stations in the mountains and marginal lands were excluded, as they are not sufficiently representative. The final selection results in a small number of observations.

Collecting observation data about hailstorms is often a difficult task  The main reason is the scale of the event. Due to its formation process, hail is a highly local phenomenon, and develops over a relatively short time span. This makes the recording quite complicated and strongly dependant on the resolution of the monitoring.

Observation data for the study were obtained from the datasets from private and corporate users and Condifesa’s archives. The number of observations collected  was different for each region in the considered period i. The sources differ greatly from each other, both in the type of institution and the aims and objectives of the research. As a consequence, the range of data obtained by these sources varies due to the collecting method and the purpose of the datasets.

Private and corporate users are specialized in collecting hail observation reports. Its dataset collects examinations spread over the whole country. However, the amount of data provided by them is rather small and locally biased due to the collecting method and the purpose. These users receive hail warnings from many sources ( insurances etc..)that send warnings from anywhere in Italy. The observations received are then compared with radar meteorological images, which allow forecasters to find proof of the hailstorm occurrence and validate the data. The resulting dataset is mainly referred to highly populated urban areas.

On the other hand, Condifesa (through ASNACODI, Associazione Nazionale Condifesa) is a public company with various bases in each Italian province. It supports farmers’ interests when opening and contracting insurance policies. It follows all the interactions between farmers and insurance agents and therefore participates in hail damage validation. Condifesa’s data consist in warnings given by farmers when hailstones harm their crops. Condifesa confirms the actual occurrence of the event through a visit and evaluation of an expert

MeteoBrowser hail moinitoring shwowing the hail storm event in Florence Tuscany 19 September 2014

The present study evaluates the efficiency of a hail forecast model, which operates in Italy and surrounding area but is ready to operate in the whole Europe.. The research covers the hail season in 2013 ( 26th of March, 15 of November, 234 days), when the model was first activated after some seasons of trial period. Each forecast is compared with observations in order to validate the prediction. Various aspects of the forecast output are taken into account, such as the hail probability risk and the hailstone size estimations

In the image above taken from our GIS Platform Meteobrowser you can see the hail forecast index marker ( grey=low , green=medium, orange=strong, red=very strong ) with overlayed in the area in the circle the radar image at the very same time with a color shading from light to dark blue. After midday of the 19th of September a strong hail storm ( more than 10 cm of hailstones on the ground in some areas of the city ) hit Florence in Tuscany. Wind blowing at more than 100 km/h , rain and hail was the result of this violent event caused by a super cellular storm cloud

Our team of meteorologist using Meteobrower hail monitoring can realtime make forecast evaluation vs realtime current weather data , trim the hail index forecast or use the data and forecast saved archivesto make analysis or fine tuning the model using the past archives

The forecast is performed at different levels of resolution (1 km, to 10 km) and with different initial and boundary conditions in order to simulate the development of various weather conditions. As a result, the model operates as an ensemble multi model system, so that the atmospheric conditions can be described with higher accuracy.
This also helps correct any bias that can be derived from ground weather stations. The outputs are then compared with convection indexes and the different scenarios given by the ensemble multi model system to define a risk level

A new exciting frontier to explore will be the full interaction of the weather hail index forecast model platform with the nowcasting or short term sources like radar and lightning devices to improve the model skill on the
very short term period

Model data ouputcan be delivered in various format and are quality checked and manually trimmed by our meteorologist staff.
-GRIB, NetCDF easy to be integrated in third part weather model chains or platforms
-CSV, API JSON /XML images the most diffused and easy to use informatic protocol with ready to use data