AR-CHaMo Severe Weather Prediction

Experimental automated forecasts from ESSL based on three NWP models

This website is made possible by the project PreCAST, funded by the Austrian FWF Wissenschaftsfonds, carried out in partnership with ZAMG, and supported by ECMWF.
AR-CHaMo was originally developed in the German BMBF projects STEPCLIM, ARCS (co-funded by Munich Re) and CHECC (ClimXtreme) by Francesco Battaglioli, Pieter Groenemeijer and Anja Rädler with support of Mateusz Taszarek and Tomáš Púčik.
* for 24-hour forecasts in the past

Move the mouse over the times on the right to see forecasts for different times. To change the forecast start time, type it into the box above (in yyyymmddhh format) and press enter.

The forecast probabilities come from the Additive Regression Convective Hazard Models, or AR-CHaMo, applied to the latest deterministic runs of the ECMWF IFS, DWD ICON-EU and NCEP/NOAA GFS models. AR-CHaMo has developed using the ERA5 reanalysis, ATDnet (courtesy of the Met Office) and EUCLID lightning detection data, and ESWD reports, collected by ESSL and its volunteer partners. The mean of the three uncalibrated probabilities of each NWP model is displayed.

Disclaimer:
These forecasts are experimental and are provided without any warranty. They should be used with care. For official weather forecasts and warnings, consult your national hydrometeorological service or visit meteoalarm.org.
Forecast Times: