A new era in weather forecasting is approaching as the National Observatory of Athens is creating a pioneering early warning system using AI for severe weather events and natural disasters
Four high-risk areas have already been placed under the microscope of the Observatory’s operational unit, METEO: Thessaly, Rethymno, Penteli, and the Municipality of Athens — with the first results proving particularly encouraging.
Also underway is a collaboration between the National Observatory of Athens and HEDNO (the Hellenic Electricity Distribution Network Operator) to create a digital platform — updated every six hours — that will include a risk analysis for all of HEDNO’s infrastructure across the country, based on real data and weather scenario modeling. This will allow for timely interventions such as maintenance work or underground cabling of sections of the network.
As the Director of Research at the National Observatory of Athens, Vaso Kotroni, noted at a conference yesterday, Greece has recorded 645 severe weather events since 2000, with a significant increase from 2013 onwards: during that period, events rose by 58% and human casualties by 75%. Notably, 37% of Attica’s forested land has burned in the last nine years.
In this context, early warning systems that refine forecasts both in time and space — down to the level of a city block — and that predict in detail the consequences of an event before it occurs, are deemed essential.
A pioneering AI system for short-term flood forecasting in Thessaly has already been completed. As Kostas Lagouvardos, Director of Research at the National Observatory and coordinator of METEO, explained: “By combining data from the last five years from an extensive network of meteorological and hydrological stations operating in the region, a system was developed for the first time in Greece that allows for real-time, near-instantaneous operational assessment of river level changes at selected points in Thessaly’s hydrographic network.”
The system, developed in collaboration with the METEO unit and the Inland Waters division of HCMR (the Hellenic Centre for Marine Research), and with the support of the Region of Thessaly, is based on modern AI techniques, with models trained to accurately capture the relationship between recorded rainfall and river level response.
In this way, it directly converts rainfall data into water level forecasts, while simultaneously incorporating real-time water level measurements from hydrological stations, continuously improving forecast accuracy. This approach allows for river level estimation over a time horizon of up to 12 hours, providing vital reaction time for the relevant authorities.
A striking real-world test came in early April: when the Pineios river was at just 20 cm, the system predicted with high accuracy that the level would rise rapidly to 5 metres the very next day, due to rainfall in Pertouli. It also provided a reliable real-time forecast of level changes for the next 12 hours.
The conference also presented a fully operational early warning system running in Rethymno. Seven new meteorological stations and additional measurement instruments were installed in the area; their data is processed in real time, enabling immediate alerts about river levels and local air quality.
As Rethymno’s Deputy Mayor Giorgos Skordilis noted, the system is crucial in an area where “huge flood damage occurred in 2019–2020, damage that has still not been fully repaired,” adding that its value was also demonstrated during a recent Saharan dust episode: “When the threshold is 40–50 micrograms per cubic metre of air and we exceeded 1,000 micrograms, the Municipality of Rethymno was the first to issue a public warning, thanks to this system — even before the problem had been officially recorded.”
An early warning system for Penteli, focused on forest fires and using high technology, is also operational. Meanwhile, the construction of a forecasting system for the Municipality of Athens is now underway; it will include specialized forecasts for heat stress at specific city points, and will identify precisely which areas are at risk of flooding.
New meteorological stations will be installed to cover all city neighbourhoods, producing tailored forecasts. As Nikos Chrysogelos, Climate and Circular Economy Adviser to the Municipality of Athens, stressed: “It is important that these systems be transformed into decision-making tools, so that the consequences of extreme events can be avoided. They are specialized instruments that can give us information down to the level of a single apartment building. At the same time, alongside adaptation to new conditions, we must also act with immediate measures to curb the climate crisis.”
