RiskFP is a geospatial web-based modeling platform developed to support both forest managers and forest insurance actors in managing the vulnerability of their assets/portfolios to fire risk. Its development started thanks to the support of Climate-KIC programs (OASIS-FIRE, OASIS+, RiskFP projects). In 2016, interactions with 2 brokers allowed us to identify specific needs from the insurance sector. One of the main needs expressed was targeting extreme forest fire situations, those which strongly damage their insurance portfolio. The Forest Demonstrator developed in the wake of the H2020-OASIS-insurance project aims at developing a prototype of a decision-making tool for climate risk and forest fire assessment, management and planning.
In the course of the project, RiskFP has been tested by several users (mainly forest managers, insurance companies and fire-fighters) in several regions worldwide: Brazil, Chile, South Europe, and Australia.
RISKFP has been specifically designed to meet the needs expressed by the insurance sector. ARIA started working with a commercial forestry insurer in Oregon, USA, and a property insurer based in Australia. Our insurance partner, ForestRe, identified a larger concern from the insurance sector for this secondary peril.
RISKFP has been designed as a fire hazard modelling platform accessible via a web-GIS user interface. The platform includes Software as a Service (SaaS) functionalities to compute, visualise and extract all the data directly from the web-GIS user interface. It includes 4 main services/modules that are available for bushfire zones of about 50km x 50km in size.
Realistic Disaster Scenarios and CO2 releases
The generator of Realistic Disaster Scenario is based on three components:
- The “critical landscape fire weather patterns”, an empirical fire weather index that identifies severe-extreme fire weather days during which weather and fuel conditions could lead to extreme fires.
- A fire propagation model based on the well-known Wildfire Analyst® that estimates fire growth and spread patterns. The simulated scenarios could be used to complete the information from historical fires in damage and loss modelling.
- CO2 emission calculation of the release based on the scenario modelled.
RiskFP allows to map precisely on a periodical basis (e.g. daily) the local forest fire risk for a given day through the graphical representation of several hazard levels. This feature provides the user with an overview of the most critical locations regarding the potential behaviour of the fire in case of a hypothetical ignition in this same area for a certain day.
Seasonal and climate forecast for critical days
The forecasting service focuses on offering the forestry sector with an operational forecast tool to predict their potential exposure to fire risk and feed decision-making processes at both a seasonal horizon and a long-term horizon. It provides information on large fires frequency.
Fire Wild-Urban-Interface (FireWUI)
FireWUI allows the estimation of thermal damages on vegetation and building materials at the wild-urban-interface (WUI). It can be useful to evaluate the effectiveness of risk reduction options like fuel treatment strategies (forest cleaning within the vicinity of the buildings or of the urban interface, forest cleaning near isolated assets in forests) on wildfire exposure at the WUI scale, depending on the local conditions of wind, topography and vegetation.
Demonstration Zones and Validation
During the project, RiskFP has been demonstrated for Australia and South of France.
- for scenarios of risk reduction at the wildland-urban interface (with Aix-Marseilles university)
- for demonstrating and evaluating the Realistic Disaster Scenario functionality against real observations of fire spread in Australia.
The comparison between modelled data and observations is satisfying considering the fire spread calculation for the 2 major events that were selected. The results for the Aberfeldy-Donnellys event are almost perfect. The exercise is a bit more challenging for the Bunyip ridge Track fire. Indeed, this is due to the spotting effect which is a quite random and erratic phenomenon. Though spotting is treated in the Wildfire Analyst model, it only refers to short-range spotting of a few 100-meters near the fire front. Extreme fires with a well-developed convection column can spot quite far.
Integration with Oasis Platform
Finally, the OASIS platform will be used for RiskFP-insurance as a business channel. Indeed, RiskFP is being implemented and will be available soon on the Oasis Hub. An action has also already been undertaken to couple the RiskFP outputs with the Oasis-LMF model (Oasis Loss Modelling Framework). The purpose is to provide an information suitable for cat models to estimate potential losses, given some risk value input data.