FLIRE foresees eight implementation actions, which aim to develop all the necessary tools and cover most of the main activities of the project.

Action B.1. Catchment Hydrological Modelling

focuses on the simulation of the spatially distributed hydrological processes that take place in the rural part of the study area. The expected outcome of this action is the development of a calibrated hydrological model that will take into account hydrometeorlogical and geomporphological features of the area, the dynamic nature of the catchment (e.g. alterations of its spatial characteristics) and factors that affect flooding. Thus, the hydrological response of the rural part will be quantified with the application of a rainfall-runoff model in the area, the outputs of which will serve as input to the urban modeling component (Action B.2).

Action B.2. Urban Flood Modelling

focuses on the efficient coupling of a surface flow model with a sewer flow model for the simulation and forecasting of flooding in the urbanized part of the study area. The expected outcome of this action is the development of a customized urban flood model of the area that may be implemented in cases where significant changes in land use, top soil infiltration and surface pathways morphology occur due to forest fires. The new urban model will be simplified for better management of computational time.

Action B.3. Flood Risk Assessment

focuses on the completion of flood risk assessment in the study area, making full use of the outcomes of the two precedent Actions B.1 (Catchment modelling) and B.2 (Urban flood modelling). The expected outcome of this action is the production of flood hazard and flood risk maps for a number of historic and synthetic rainfall time series (scenarios) and the development of a pattern recognition algorithm for the association between near real-time weather information and rainfall scenarios.

Action B.4. Forest fire risk assessment and mitigation planning

focus on the development of the fire risk assessment and the fire propagation components that will be integrated in the framework of FLIRE. The expected outcome of this action is the development, adaptation and testing of a fire risk module using the KBDI index and a fire propagation module that will be based on the GFMIS forest fire simulator. (Fig. B4a,B4b, B4c).



Action B.5. Short-term weather forecasting

focuses on the development of a short-term weather forecasting system. This system will receive a daily rainfall forecast, as well as observed real-time data from raingauge stations in and around the case study area. The expected outcome of this action is the production of a short-term forecast for the case study area based on real-time datasets. (Fig. B5).



Action B.6. DSS Tools

focuses on the development of a web-based decision Support System (DSS) which will combine information from different model outputs (fire and flood information) and create early warning information like fire warnings and flood warnings for the local authorities. The DSS will integrate the various FLIRE tools and will enable an effective approach of flood and fire risk management.(Fig. B6).



Action B.7. Planning tool for flood management

focuses on the development of a Planning tool for flood risk management, i.e. a tool that will assist flood risk management in the study area at a planning level. The expected outcomes of this action are an improved understanding of the possible intervention options for managing floods in the case study area and their cost-benefits, as well as the development of an action plan (set of measures) in collaboration with stakeholders with specific interventions identified and their impact on reducing flood risk assessed.

Action B.8. Application of the tools

focuses on the application of all the tools of the DSS Platform within the estimated time schedule. The expected outcomes of this action are the integration of the DSS Tool (output of Action B.6) and the Planning Tool for Flood Management (output of Action B.7) in a common platform and the on-line application of the entire system.

The hydrological modelling of the catchment implies a good understanding of the hydrology of the area and the main parameters that govern its flooding. Consideration of the dynamic nature of the catchment during the hydrological modelling (in view of events, such as fires and floods but also urbanization), using “live” timeseries together with up-to-date underlying maps and assessment of the hydrological behaviour of peri-urban areas considering hydromorphological issues (initial soil moisture conditions, initial abstraction from the ground etc.), alterations in the spatial characteristics of the catchment is crucial. Thus, the quantification of the hydrological response of the rural catchment with the application of a rainfall-runoff model in the area will take into account a complete and continuously updated dataset of hydromorphological parameters. In order to achieve the above, the following tasks will be undertaken: Collection of hydrometeorological and topographic datasets, collection of satellite data for land cover after the occurrence of flood or fire events, collection of soil moisture datasets from satellites and sensitivity analysis of the impact that alterations in the spatial characteristics of the catchment may have on its hydrological behavior. The result of this action will be a calibrated hydrological model able to link to urban flood modelling tools and its outputs (which will correspond to responses of the rural catchment to different rainfall events, under different hydrological conditions) will be tailored to inputs required from the urban flood model.

During extreme rainfall events urban sewerage systems are incapable of conveying all surface runoff and evidence shows that a considerable volume of water is conveyed on the surface. This leads to urban pluvial flooding which causes problems to human life and the environment. To minimize the risk from flood events, accurate representation of a drainage system and calculation of its response to different conditions is needed. This can be achieved by modelling the complexity of urban flooding and computer software has advanced significantly in this sector over the last few years. The most important recent development is the ability to accurately represent the interaction between the underground sewer networks (minor system) and the surface flow (major system). Thus the urban flood modeling action implies the proper selection of existing urban flood models (sewer flow – surface flow, 1D/1D, 1D/2D approach) and their efficient coupling and integration. Moreover, this model will be upgraded and customized so that it can be also implemented in cases where significant changes in land use, top soil infiltration and surface pathways morphology occur due to forest fires.

The next step forward is the integration of the catchment modelling and urban modelling components and their combined running for different rainfall scenarios. These scenarios will be both historic and synthetic rainfall timeseries. The historic rainfall datasets will be retrieved from the HOA (http://hoa.ntua.gr) and the meteorological stations of NOA (www.noa.gr). Synthetic rainfall timeseries (covering events that range between recurrent events [T=2 years] and exceptional events [T=200 years] and considering climate change scenarios) will be generated from the historical timeseries using appropriate algorithm. Then the off-line hydrological and hydraulic analysis of the study area for the above rainfall timeseries will be performed. Flood hazard maps (maps depicting water levels and velocities in inundated areas) and flood risk maps for each rainfall scenario will be generated and stored in a scenario database. These maps will be produced based on socioeconomic datasets for the inundated areas (including inter alia information on loss of human lives, land use data and information on damage to residences and other buildings, property values, information on loss of agricultural production, reduced harvest, population densities and socioeconomic status of residents (residents’ GPD etc)), taking into consideration the ecological impact of floods as well.

The fire risk assessment and fire propagation will be delivered in a similar way. The forest fire risk assessment module, based on the KBDI index will allow the daily monitoring of the forest fire risk in the study area and the forest fire simulation web service for assessing fire propagation on-line will be integrated in the DSS (described below), in order to allow early warnings in case of significant forecasted events. Moreover, the forest fire simulation module will be used for running various potential scenarios in order to assess interactions between flood and fire events and support prevention planning.
Parallel to that, a short-term weather forecasting tool will be developed. This tool will receive a daily meteorological forecast with 1hr temporal resolution and a 2kmx2km spatial resolution and real-time datasets from relevant NOA and HOA raingauge stations in and around the case study area. A short term forecast will be produced by integrating the two datasets and will be available a couple of hours ahead of the predicted event with a resolution of 10min. Particularly for floods, a pattern recognition/CBR algorithm will match this short-term forecasting with one of existing rainfall scenarios that were used in the off-line runs of the flood models. This algorithm will be developed by using of a set of fuzzy similarity metrics for the association between near real-time weather information and rainfall scenarios and thus computed flood risk, in order to save considerable time for the issuing of warnings. Similarly for fires, the weather forecast will be used for the production of fire risk maps.

A Weather Information Management Tool (WIMT) will include weather information and assist the assessment of flood and fire risk. Particularly for floods, when a significant rainfall event is predicted to occur based on the daily rainfall forecast, a high- level warning will be posted in the website by the system. Closer to the predicted event, the pattern recognition algorithm will result in the matching of the short –term forecasting report with one (or several) flood events, given that each rainfall scenario is coupled with a specific flood scenario. If the flood event forecasted is significant enough, an early warning will be sent by the system to selected stakeholders via a number of standard methods (e.g. sms, email, phone service etc.). Similarly for fires, warnings will be sent when necessary to stakeholders involved.

In order to combine all the above information and tools and make them available and valuable to local authorities, decision makers and citizens a web-based decision Support System (DSS) will be developed. This platform will integrate the various FLIRE tools to be used for combined real-time flood and forest fire risk management. It will produce early warnings for both floods and fires, communicate them to the local authorities and help the decision makers to timely plan their responses when necessary.

One additional tool will be developed as part of the project, the planning tool for flood risk management. This tool will assess and suggest possible intervention options for managing floods in the case study area on the long run and perform cost-benefit analyses. The on-line application of the entire system (all the tools of the DSS Platform) will be tested for 9 months so that any (unforeseen) problems that may arise during the implementation of the action will be successfully addressed.