Seasonal prediction of climate-driven fire risk for decision-making and operational applications in a Mediterranean region
In one sentence. Builds a data-driven model linking SPEI drought indices to Catalan burned areas, forecasting summer burned area seasonally with stakeholder co-design and cross-validated skill.
Abstract
In this paper, we assess and develop a climate service focused on the production of seasonal predictions for summer wildfires in a Mediterranean region through a participatory approach with end-users. We start by building a data-driven model that links a drought indicator (Standardised Precipitation Evapotranspiration Index; SPEI) with a series of burned areas in Catalonia (northeastern Spain). Afterwards, we feed this model with SPEI forecasts obtained through a combination of the antecedent observed conditions and climatology. Finally, we assess the forecasting skill of the system by using cross-validation to evaluate the predictions as if they had been made operationally. Our fire forecasting system reveals an untapped and useful burned area predictive ability. We argue that this source of predictability is mostly attributable to the effect of observed initial conditions on summer drought conditions. This system was conceived with the stakeholders, merging climate-driven predictions with information that is of interests to the users, including the identification of climate variables, thresholds and models. The co-production of this customized system allows fire-risk outlooks to be translated into usable information for fire management. This fire forecasting ability plays a crucial role in developing proactive fire management practices such as long-term fuel assessment and other fire-risk planning, thus minimising the impact of adverse climate conditions on summer burned area.
Labels
- Topic
- Extremes & hazards, Sub-seasonal to seasonal
- Type
- Application study
- Method
- Classical ML (trees, GP, linear, kernels)
- Time scale
- Seasonal to annual
- Domain
- Regional
- Data
- Other
- Status
- Research
- Identifiers
- arXiv:1905.01110 DOI:10.1016/j.scitotenv.2019.04.296 OpenAlex:W2940533603
- Added
- 2026-10-22
Labels and summary are generated by Claude and reviewed weekly. Wrong? Edit this entry.