The IRFMS prioritises evidence-based decision-making. By mandating probabilistic risk assessments, establishing detailed hazard and risk maps, and developing performance indicators tied to project and strategic outcomes, the NAP ensures actions are guided by robust data. Through regular progress reports, barometers, and lessons learned processes, the NAP promotes iterative learning and continuous improvement, solidifying the role of evidence-based practices.
Towards an integrated rural fire management framework in Portugal
Portugal reviewed its national fire hazard map in 2022 to inform prevention measures. Produced by the national forest office (ICNF), the map provides an overview of wildfire likelihood and potential severity across mainland Portugal. It combines data on rural land cover, slope, elevation, and historical wildfire occurrences to classify the national territory into five hazard levels (Presidency of the Council of Ministers, 2021[27]) (Figure 6).
Portugal’s recent methodological updates in wildfire hazard mapping enhances its predictive accuracy. An earlier hazard map classified mainland Portugal using arbitrary quantiles, assigning equal land portions to each hazard class without reflecting actual fire occurrence (Ministry of Agriculture, 2012[64]). The 2022 hazard map adopts a new methodology, considering terrain and vegetation, including fuel accumulation and combining it with fire probability derived from decades of historical records. Hazard levels are now defined using performance-based thresholds, ensuring that the “high” and “very high” classes cover most historically burned areas while representing a smaller share of the territory (ICNF, 2020[65]). This shift makes the classification more aligned with actual wildfire patterns, increasing its value for prevention and planning (Oliveira S, 2021[66]).
The other key tool to assess wildfire hazard in Portugal is the rural fire danger map. This map uses the same danger categories of the national hazard map and has a resolution of one kilometre. By integrating near-real-time meteorological information with other hazard information, it provides a snapshot of likely daily hazard levels and is the main tool to inform short-term risk reduction and emergency preparedness and response decisions. For example, it is used to determine daily bans and restrictions on potentially hazardous activities in rural areas – such as for example waste burning, the use of barbecues, the ignition of bonfires, and the use of forest machinery – and to inform the allocation of resources for fire monitoring and firefighting (Presidency of the Council of Ministers, 2022[67]) (IPMA, 2025[68]) (Figure 6).
Beyond these maps, Portugal has developed a wide set of advanced technologies and analytical systems to strengthen fire pre-suppression and suppression. These include the establishment of a dedicated decision-support unit to analyse rural fire behaviour, forecast its evolution, and guide tactical priorities (Núcleo de Apoio à Decisão – Análise de Incêndios Rurais, NAD-AIR); a monitoring platform providing real-time geographic and operational data; the expansion of the video-surveillance network and improved satellite capability; the systematic use of Copernicus and other satellite data for fire mapping and tracking; and a National Fire Progression Database to support evidence-based operations. Together, these developments significantly improve situational awareness, operational readiness, and decision-making during fire events (AGIF, 2025[10]).
While hazard assessment is essential to assess fire risks, relevant data are limited. Beyond appreciating the hazard, risk assessment allows to integrate information of exposure and vulnerability of social, economic and environmental assets. Despite progress in developing a digital land registry, Portugal’s national cadastral system remains largely incomplete,1 limiting the creation of a comprehensive inventory of assets exposed to wildfire risk. Public authorities also face challenges obtaining information on asset values from private owners, hindering accurate risk and potential loss estimations. While detailed demographic data on population density, age, and other vulnerability indicators exist, they are not fully used to feed into wildfire risk assessments. Some intermunicipal communities have developed partial risk maps combining the national hazard map with local data on vegetation types or infrastructure locations. However, these maps are rarely comparable, and infrequent updates limit their reliability. The national forest agency (ICNF) is working to create a nationwide wildfire risk map that integrates asset exposure, economic value, and vulnerability into existing hazard assessments (OECD, 2023[5]).
Note: The National Wildfire Hazard Map elaborated in 2022 is shown on the left, while the Rural Fire Danger Map for 12 November 2025 is shown on the right. These maps represent different dimensions of the same hazard: the left-hand map reflects structural risk, while the right-hand one provides a daily snapshot of expected fire danger levels.
Source: Adapted from (ICNF, 2022[69]) and (IPMA, 2025[70]).
Currently, climate change projections are not integrated in national wildfire hazard maps. Data on projected wildfire hazard levels under different climate scenarios are available, including projected number of days with a significant fire hazard index (OECD, 2023[5]) (IPMA and University of Lisbon, 2025[71]). The Portuguese Environmental Agency has already conducted work to project forest fire hazards under the national roadmap for adaptation 2100. This includes tools like the enhanced Fire Weather Index and downscaled climate data that can inform fire hazard maps and help assess projected fire risks when combined with data on exposure and vulnerability (Portuguese Environment Agency, 2024[72]).
The national hazard map is intended to guide most fire prevention regulations and activities, but it has been criticised. This map helps identify priority areas for fire risk reduction efforts, including fuel management obligations or building permitting requirements, besides being used to inform inspections, surveillance activities, and awareness-raising campaigns. However, local stakeholders note that the map lacks sufficient detail to effectively prioritise local actions (Tedim et al., 2023[33]). This limitation has contributed to delays in the adoption of municipal fire management plans, which must be based on the national map for planning and prioritisation purposes (AGIF, 2023[30]) (AGIF, 2025[10]). Although subnational authorities in Portugal have some flexibility to assess local fire risks and adapt regulations accordingly, this flexibility is rarely used in practice. In contrast, other countries have developed tools to enable municipalities to assess wildfire risk at a finer spatial scale, as seen in Chile (Box 9). Collaborative approaches have also proven effective, such as local natural risk prevention plans in France (Box 3) and in Victoria, Australia, where local risk maps are updated twice yearly to reflect changes in vulnerability.
While municipalities were initially required to integrate the national wildfire hazard map into their spatial planning instruments regardless of their wildfire hazard level (OECD, 2023[6]), a 2023 amendment to Decree-Law No. 82/2021 removed this obligation. In practice, today, the integration of wildfire hazard considerations into local spatial planning now takes place through the designation of priority prevention and safety areas (Áreas Prioritárias de Prevenção e Segurança, APPS). These areas are approved at the sub-regional level, with the direct involvement of municipalities.
Chile’s Climate Risk Atlas (ARClim) is a national platform that maps current and projected climate threats and assesses 80 climate risks at the communal level under the IPCC’s worst-case scenario. It is being expanded into a comprehensive adaptation information system, integrating indicators on subnational and sectoral adaptation progress, as well as links to research and citizen engagement platforms to support informed decision-making. Chile is developing subnational indicators for use in regional adaptation plans and environmental assessments. ARClim includes impact indicators – such as effects on transmission lines or crop productivity – and aims to build a cross-sectoral climate risk framework to better understand interdependencies and prevent maladaptation. It identifies 36 different climate hazards, for a baseline period of 1980-2010 and future hazard projections for 2035-2065, under a RCP8.5 scenario. Fires are not considered as projected climate hazard but as risks that will be suffered by specific sectors. The projected risks suggest a strong correlation between an increase in temperature and fire spread. Other drivers such as vegetation cover or latitude are used to forecast the sensitivity to fires, backed by machine learning algorithms.
Note: The extract of ARClim shows the hazard level, exposure and vulnerability assessment for various Chilean cities (forward-looking risk index on the right), taking into account dimensions such as distance to a road, forest cover, population density or elevation.
The implementation of the NAP has been monitored since its inception. The objective is to monitor progress in fund allocation as well as in the implementation of measures defined in the NAP. Monitoring and evaluation are carried out quarterly, with results consolidated annually by AGIF in collaboration with all relevant stakeholders (Government of Portugal, 2018[73]). An interim evaluation is scheduled for 2025 to assess the NAPs overall relevance, effectiveness, efficiency, and sustainability, allowing for adjustments to be made before 2030. The NAP monitoring system is based on a coordinated approach, with a dedicated online reporting platform, clear indicator tracking, annual reporting and scheduled evaluations. This shift creates a transparent, interoperable and evidence-based monitoring culture, enabling better planning, prioritisation and accountability across all levels of fire management.
In addition, opinion surveys and lessons learned processes help improve performance management. AGIF conducted two national barometers in 2019 and 2021 to assess public perceptions of wildfire management and the effectiveness of awareness campaigns (GfK, 2021[74]). Overall, citizens acknowledged the importance of these campaigns, particularly for understanding preventive measures and improving wildfire management (60%), and for feeling more involved in fire management compared to previous years (50%). While these surveys help monitor awareness, they do not necessarily provide evidence on which methods are most effective to communicate about fire risk and conducive behaviour to reduce risks (AGIF, 2021[75]). In parallel, AGIF’s collaborative lessons-learned process allows agencies to share information, analyse activities, and implement improvements. Since 2018, two such exercises were conducted after major fires, highlighting operational gaps in resource allocation, fuel management, and local response capacity. These resulted in 11 lessons and 29 corrective actions, each with clear responsibilities and timelines.
Some NAP indicators lack clear definitions and ambition, limiting their usefulness to evaluating progress. Some NAP indicators being collected look at the share of “strategic decisions” that are risk-based (target: 100%) or the share of the population adopting best practices (target: 70), without there being a definition of what constitutes either “strategic decisions” or “best practices.” This makes it difficult to ensure consistent measurement across the country and raises questions about the reliability of the evidence produced to demonstrate improved risk management. Moreover, not all national targets established to achieve wildfire resilience represent an ambitious shift such as limiting the share of fires exceeding 500 hectares to 0.3% of total fires. Historical data shows that only two years between 2006 and 2018 exceeded this level, with an average of 0.18% between 2012 and 2023 (AGIF, 2025[10]). In contrast, the State of Victoria (Australia) sets a more outcome-oriented target: maintaining fuel-driven bushfire risk at or below 70% of maximum potential levels – defined by worst-case fuel build-up and extreme fire conditions (Forest Fire Danger Index of 130). Tracking this indicator can provide a more meaningful measure of wildfire risk reduction (Victoria State Government, 2022[76]).
NAP indicators fall short of assessing impact. Several NAP’s outcome indicators – such as “managing 1.2 million hectares of fuel until 2030” – reflect outputs more than outcomes in terms of actual wildfire risk reduction. Assessing the effectiveness of fuel management requires ex post analysis of how interventions influence fire behaviour, as done in California’s fuels treatment effectiveness dashboard (Box 10). Similarly, AGIF tracks outreach of awareness raising campaigns (e.g. number of people or municipalities reached) but not how this translates into behavioural change (AGIF, 2023[30]). Although it reports that 35% of the population has adopted best practices, the underlying methodology is unclear (AGIF, 2025[10]). Finally, some indicators aim to measure project impacts that cannot be directly attributed to the projects themselves – for example, rural development initiatives often neglect key social drivers like access to services. As a result, observed outcomes may not reflect the actual effectiveness of the projects, and thus cannot provide robust evidence for continuous improvement.
Finally, evidence produced by the lessons learned process could more strongly foster continuous improvement, beyond for operational management of fires. The lessons-learned process remains largely focused on post-fire response, with limited engagement from local authorities and few evaluations of prevention measures. Additionally, implementing corrective actions has proven challenging, as they often require reallocating resources or changing established practices, which can face institutional resistance. Consequently, although the process is useful, it has not driven systemic change.
California’s Fuels Treatment Effectiveness Reporting (FTER) Dashboard is an interactive tool developed by CAL FIRE to evaluate and visualise the impact of vegetation management treatments on wildfire behaviour across California. It compiles data from Fuels Treatment Effectiveness Reports, which assess how various fuel reduction activities – such as prescribed burns, mechanical thinning, and fuel breaks – affect fire dynamics, including spread rate and intensity ex post. These reports also document how treatments assist in fire containment, provide access for firefighting personnel, and protect life, property, and natural resources. The dashboard integrates information from CAL FIRE’s incident database (CAL FIRS) and the CalMAPPER geospatial system to identify wildfires that occur within a quarter mile of fuel treatment areas. When such proximity is detected, local CAL FIRE units are required to complete a survey within seven days of the fire being controlled, detailing the treatment's effectiveness. This systematic approach ensures that the benefits of fuel treatments are documented and analysed, providing valuable insights for future wildfire mitigation strategies.
To enhance wildfire management, additional data are needed. This includes a comprehensive inventory of exposed assets, detailed vulnerability assessments using up-to-date spatial and socio-economic data, infrastructure characteristics, critical facilities, and climate projections to evaluate future risks. Improved public access to data and digital tools can strengthen risk assessment. Collaborative, multi-level approaches engaging local authorities, scientists, citizens, and forest associations in co-developing risk maps and land-use planning foster shared understanding and local buy-in.
In addition, progress could be better monitored using SMART (specific, measurable, achievable, relevant, and time-bound) indicators under the NAP, with clear responsibilities for data collection and attention to data reliability and comparability. Finally, establishing a more regular lessons-learned process, led by AGIF, would evaluate the Integrated Rural Fire Management System, identify gaps, and guide continuous improvement.
← 1. In 2024, the Digital Land Registry platform covered approximately 58% of mainland Portugal (AGIF, 2025[10]).