SUMMARY

In the last 20 years, Portugal has been severely affected
by large wildfires with dramatic consequences.

The main objective of this project is to combine the expertise of several institutions of the scientific community to provide sound and efficient tools capable of improving decision making during wildfires crisis to minimize its negative consequences.
A key issue is the lack of decision support mechanisms for operational interventions.

Due to the complexity of large wildfires, their effective suppression requires suitable and well coordinated resources, up-to-date knowledge of the landscape, and accurate prediction of fire behavior.

A Decision Support System (DSS) that can integrate the panoply of required information in a simple and efficient platform is the main scientific challenge of foRESTER. The main goal is to provide fire managers with useful and sound information to improve fire suppression strategy and decisions.

To accomplish this, foRESTER proposes a fast, reliable and informative DSS based on advanced computational intelligence and visualization techniques, that integrates innovative technologies from multi-sensor systems, cutting edge satellite image processing, and near real-time (NRT) fire spread predictions (FSP).

Hardware Design, Satellite Image Processing
and Multicriteria Decision

Fire Spread Modelling

Cartography and IPSentinel Platform

Land Cover Land Use Classification

Image Processing

Weather Forecast

end-user: Câmara Municipal de Mação (MacFIRE)

Project Description

Scientific Domain

Main Area

Sistemas de sensorização, de informação e de comunicações de emergência e sua integração nos processos de decisão

Secondary area

Gestão do fogo e comportamento de fogos extremos

Acronym

foRESTER

Project title

Data fusion of sensor networks and fire spread modelling for decision support in forest fire suppression.

Requested funding

399.903,75€

Keyword 1

Low cost multi-sensor WSN

Keyword 2

Land cover land use mapping from satellite data

Keyword 3

Fire spread Modelling

Keyword 4

Multi-criteria Decision Support System

In order to have a more precise and granular information from the field a low cost and power efficient, flexible, reconfigurable, multifunctional smart sensing system will be developed. This system combine fixed and portable sensors in a flexible wireless sensor network (WSN), which will acquire key weather data, together with visible and infrared images for local classification of the fire front. This information will be processed locally using a real-time signal-processing algorithm that will run in a hardware platform (FPGA based), decreasing the amount of data to be transmitted to the DSS.

The project aims to provide a low-cost framework based on a WSN and new information system tools to produce a pilot demonstrator of the effectiveness of DSS for supporting decision-making in fire suppression context. It will be built on past experience from MacFIRE, together with new layers of information, providing a more informative DSS for the fire managers. Moreover, we will extend the pilot (institutional support letters in appendix) to the Médio-Tejo region and demonstrated it as a scalable platform for the entire national territory.

Fire spread models have a large potential to support fire management decisions. We aim at improving FSP by refining input data quality, performing regional calibration, and by integrating uncertainty into model predictions. More accurate weather forecasts will benefit from the WSN field data and annual fuel maps will be derived from the detailed and updated LCLU maps. Simulation outputs will identify the likelihood of a given area burning and the associated fire intensity, thus contributing to anticipate fire behavior and support suppression decisions.

The information from the WSN, LCLU maps, and FSP will be integrated in the DSS. The combination and fusion of this information, using computational intelligence and advanced visualization techniques, will support decision-making by fire managers. Beyond a layered visualization of information (resource positioning, FSP and WSN data) over cartographic maps and satellite images, DSS will provide tools to generate early warnings of changes or extreme meteorological conditions based on field WSN data and assist planning of resources allocation (personnel and machines).

To improve the quality of the landscape characterisation we will take advantage of the satellite imagery provided by the Sentinel-1 and 2 missions. Innovative algorithms will be developed to explore multi-sensor and multi-temporal satellite images to produce updated high-resolution Land Cover Land Use (LCLU) maps, to characterize thematic uncertainty and to enhance the spatial resolution of Sentinel-2 imagery. These updated and improved maps are crucial to provide a better characterisation of the landscape and improve FSP.