CORDIS - EU research results

Flood zoning in Southeast Attica using gauge calibrated radar rainfall and advanced modeling techniques

Final Report Summary - FLADAR (Flood zoning in Southeast Attica using gauge calibrated radar rainfall and advanced modeling techniques)

Project context and objectives

Floods are natural hazards with the potential for many fatalities and significant economic losses. However, the degree of vulnerability to any disaster is also a function of human action and behaviour. The main objective of the FLADAR project is to follow the steps suggested by the 2007/60/EC Directive on the assessment and management of flood risks in order to initially conduct a preliminary flood risk assessment and then draw flood hazard zones for a partially developed watershed in East Attica (130 km2). Until recently, land use in the area was solely rural. However, during the last 15 years, an unorganised urban expansion, stimulated by the infrastructure work for the 2004 Olympics and an increase in the capital city’s population, resulted in a mixed urban and rural environment. Even though there were concerns about several new flood-prone locations, no comprehensive studies leading to flood hazard maps exist.

Radar rainfall data was provided by the National Observatory of Athens (NOA) for four storm events and were calibrated at selected stations in a rain-gauge network operated by the National Technical University of Athens (NTUA). A radar-systematic error ranging from 34 % to 45 % at four rain gauges was removed using a coefficient. The Geographic Information Systems’ (GIS) pre-processed files were imported in to the fully distributed Vflo hydrological model to capture the spatial variability of the basin’s rainfall run-off. The Rafina stage gauge marks the furthest point downstream where observed data is measured and the study area is described using 2 755 cells of 200 m2 in size.

The next steps of the analysis involved setting up a hydraulic model and deciding on different ways to express the rainfall variability of rare storm events in the basin. The creation of one or more random rain components (OpenMI compliant) provided the opportunity to introduce storm movement to the watershed and statistically identify the range of flows (median, maximum, minimum) that may be expected at selected cross sections for events up to a specific volume and rainfall intensity. The 100-year 6-hour design storm raised the water level at the Pikermi cross section so that the banks overflowed; there was an even greater impact downstream towards Rafina. However, the banks downstream of Pikermi station are prone to erosion so special attention should be provided by the relevant authorities before a rare event takes place.

The final phase of the study, the GIS floodplain delineation, is actually an iterative process because it may include, at a local level, erroneous results caused by an inaccurate depiction of cross sections, an omission of significant ineffective/storage areas and coarse grid analysis. Thus, the exported hydraulic data should be studied in more detail at a cross-sectional level, especially when dealing with locations such as potential evacuation routes for the communities. It is interesting to note that the preliminary study analysis, conducted at the beginning of the FLADAR project that used only topography, land use and settlement information, pinpointed the vulnerable areas in the watershed where urban planners and engineers should pay the most attention. Such knowledge could actually benefit many communities since it shows that in ungauged watersheds or sub-basins, where measured values are almost non-existent, there are simplified but still reliable ways to identify several defenceless areas so as to protect them in order to minimise future losses from flood events.

Project website: