Michele de Rosa

Satellites give us the possibility to monitor and warn against extreme weather conditions in a way unknown until a few years ago.

Their 'eyes' are able to look at inaccessible sites and regions which are not currently covered by any ground measurement recordings, like deserts and oceans.

In 2013, the company Geo-K, a spin-off of the Tor Vergata University of Rome, created a group of people whose study was focused on the development of pre-operational nowcasting systems, based on Meteosat Second Generation (MSG) satellite data.

My passion for the subject of weather led to me to joining this group. I left my job, after 11 years working as a project manager on defense and civilian systems for a private company. Now, my role at Geo-K is researcher and meteorological systems analyst.

Before joining Geo-K I already had an active interest in both weather and MSG data. In 2009 I became the co-founder of the Mondometeo website, for monitoring and providing short-term weather predictions using MSG data.

My area of research concentrated on the nowcasting of extreme weather events using MSG satellite data. I have presented results at three EUMETSAT conferences (2010, 2013 and 2014).

After I joined, our group started to study and improve my model — which I originally developed during my PhD, together with Prof. Frank Silvio Marzano — in order to fully understand its potential.

The main aim of the model, later named MeteoCAST, is to predict some weather parameters, like land surface temperature or cloudiness and rainfall, through the sole use of the synthetic MSG infrared channels, produced by the algorithm on a pixel basis. At the end of this stage a pre-operational system, based on the enhanced model, covering Italy, Switzerland, Austria and UK, has been set up at the Geo-K site.

Recently, together with my research group and in collaboration with the Meteorological Service of the Italian Air Force, a new storm detection system, named StormTrack, has been studied and developed. The model uses a combination of several MSG channels to detect and track convective objects and warn about their behaviour for civil purposes.

The aim of the model is to use the high temporal and spatial MSG resolution to monitor and (in the near future) predict the evolution of convective objects, especially over areas that are not currently covered by ground measurements.

A pre-operational system has been set up at the Geo-K site in order to monitor the convection activity over Italy, Switzerland Austria and South Africa in near real-time.

Together with the previous tasks, a validation activity has been started to demonstrate and certify the positive effects of the outputs of both systems. Some pre-operational web services have also been developed, to share model outputs worldwide.

These free services are hosted on the Mondometeo website and are available 24 hours a day. Currently Italy, Switzerland, Austria, UK and South Africa are covered.

We believe the web is a unique and faster way to share alerts about weather anomalies. The internet makes it possible to warn people everywhere (potentially). We think that it could improve the life of a lot of people around the world: one of the goals of atmospheric monitoring.

In order to give an idea about the possibilities created by our remote services, a mobile app was developed, which has been active since September 2014.

The app, called InstantWeather, is also free and is available in the Android store. It provides very short-term weather predictions for the area around the user's position, via our web services.

Future steps will be improvement of our models using the validation phase results; the application of our services to real-world fields like aviation, outdoor activities and civilian protection, and the extension of their coverage to other countries and to other satellites such as Meteosat Third Generation and the GOES satellite family.