An integrated framework for habitat restoration in fire-prone areas: part 1 – co-creation of land management scenarios

Introduction

Habitat conservation and restoration are central goals defining legislation across Europe, in particular the recently approved Nature Restoration Law (European Commission 2021). Under this new legislation, it is advocated that individual European Union (EU) countries restore at least 30% of land habitats in poor condition by 2030.

Apart from the urgency of complying with these new targets, Portugal faces major ecological, social and economic challenges related to wildfires (Oliveira et al. 2020; Girona-García et al. 2023). Fire is an inexorable part of landscape dynamics, given the existence of the Mediterranean climate, marked by long, warm and dry summers and highly flammable vegetation, making it very prone to wildfires (Fernandes 2013a; Santana et al. 2016; Häusler et al. 2019; Dupuy et al. 2020). Moreover, statistics and projections of climate change (IPCC 2022) point to an aggravation of conditions, including rising temperatures and long drought periods, leading to a plausible intensification of wildfire risk in Portugal (Dupuy et al. 2020).

Combining natural habitat conservation and fire hazard management can conflict in practice; fire management practices may disrupt or threaten natural habitats, or species that rely on specific ecological conditions. Conversely, prioritising habitat restoration without addressing fire hazard and risks can lead to increased vulnerability to wildfires, endangering both ecosystems and human communities (Brown et al. 2004; Foresta et al. 2016). A combined approach between habitat restoration objectives and fire hazard management is essential, integrating ecological science with fire management strategies to create resilient landscapes that support both biodiversity and fire safety.

Tackling the challenges of wildfires requires a multidisciplinary strategy based on a thorough understanding of the region, natural capital, vegetation characteristics and wildfire dynamics, supported by effective national and European policies (Santana et al. 2016; Colantoni et al. 2020). Policies such as the EU Forest Strategy and the European Green Deal provide critical frameworks for addressing wildfire risks and promoting global sustainable land management practices (Wulf 2003; European Commission 2021). The catastrophic wildfires of 2017 emphasised the already urgent need for developing strategies to cope with fire risk and mitigate negative impacts (Skulska et al. 2020; Andrade and Bugalho 2023). In this sense, Portugal has been developing and implementing national policies, such as those adopted in the Landscape Transformation Plan, to increase landscape resilience (Council of ministers 2021). These include a series of initiatives aimed at promoting integrated landscape management, with a particular focus on decreasing wildfire risk in vulnerable territories. The creation of Integrated Landscape Management Areas (ILMAs) in the target vulnerable territories is the focal management support tool of these policy initiatives, aiming to increase forest area under integrated management to increase forest resilience to fires, natural capital valorisation and rural economy promotion (Dec. Lei 82/2021). These efforts underscore the efforts made to align national and regional policies and initiatives with broader policy frameworks to support wildfire risk management (Valente et al. 2015a; Navalho et al. 2017; Wunder et al. 2021). These recent initiatives, still at an early stage of their implementation, offer outstanding opportunities for testing strategies that optimise the implementation of global policies and guidelines to the local scale.

Downscaling from global guidelines to local practices should be done following a coherent strategy with local stakeholders. Co-creating different forest and land management scenarios with the local stakeholders to explore various strategies and their potential outcomes can be a valuable approach to plan future actions and to increase possibilities for their effective implementation. Such hypothetical land management scenarios, aimed at increasing landscape resilience, can be designed to include valuable aspects of habitat and ecosystem restoration, and valorisation of natural resources, but also consider the conditions and needs of the region, from economic, social and cultural perspectives (Valente et al. 2015a; McBride et al. 2017).

In this sense, stakeholder participatory approaches could serve as valuable initiatives for co-creating forest management scenarios tailored to local needs and realities (Otero et al. 2018; Almeida 2022). By engaging governmental and non-governmental organisations, research institutions, local landowners and other actors, these approaches can facilitate collaboration and ensure that diverse perspectives are considered in the decision-making processes (Almeida 2022).

Subsequently, the integration of fire modelling in the land management scenarios explored could provide valuable insights into assessing wildfire risk and prioritising management practices in strategic locations across the landscape (McCaw et al. 2008; Syphard et al. 2011; Fernandes 2013b; Santana et al. 2016; Parente et al. 2022). Fuel models, like the Northern Forest Fire Laboratory (NFFL) fuel models or fire behaviour fuel models (Anderson 1982), characterise fuel beds for use with software tools incorporating Rothermel’s fire spread model (Rothermel 1972), thus allowing estimates of fire behaviour characteristics and fire growth across the landscape. The use of such models as a component of the designed land management scenarios and the further development of fire growth simulations can offer insights into the potential for mitigating wildfire impacts, depending on different combinations of management strategies. The output of fire simulations could feed post-fire effects modelling, thereby allowing accounting for post-fire risk assessment and evaluating the potential need for post-fire mitigation strategies (Parente et al. 2022).

The main objective of the present study is to propose a multi-factor framework for landscape planning aimed at increasing the area of natural habitats, through restoration (passive or active), and management of wildfire hazard. This aim is in line with the focus of the political instruments in place and intends to deliver a framework applicable to the management of the study area and other comparable areas. This framework is intended to combine ecological, social and management dimensions through a multi-step and multi-actor approach.

The present work (Part I) describes the overall concept of the framework, focusing on its implementation in the study case. It details the co-creation of the land management scenarios, the fuel models selected to describe the different forest types and the proposed management strategies for each of the scenarios developed, based on the stakeholders’ perceptions. Part 2 will focus on fire simulations performed for the different land management scenarios, describing the relevant steps, from the selection of future climate scenarios to the fire weather conditions of the simulated fire events.

Materials and methods

Study area background

The case study is located in northern Portugal, in the Bragança District, NUTS (Nomenclature of Units for Territorial Statistics) III level region of Terras de Trás-os-Montes, and is included in the Mediterranean Climate Zone, in the transition between dry and warm summer (Csb) to dry and hot summer (Csa) (Kottek et al. 2006) (Fig. 1). This area is located in a protected area, the Montesinho Natural Park, that is part of the Portuguese Natura 2000 national site of Montesinho/Nogueira, included in the Natura 2000 Network. This European network was created to ensure the long-term conservation of species and habitats in Europe, and is the predominant instrument for nature conservation in the European Union (European Comission, 2021).

Fig. 1.

Location of the study area, forest area of Baixa da Lombada, in mainland Portugal, and soil use cartography of the study area, according to COS 2018 (DGT 2019).

It is an area of approximately 1540 ha, with agricultural and forest cover, that was recently constituted as an ILMA (Integrated Landscape Management Area of Baixa da Lombada), as part of the recent implementation of the National Plan for Landscape Management. This political instrument envisages increasing the forest area under management in a way that balances the protection of natural capital, landscape resilience to wildfires and promotion of the local economy.

Framework definition

The framework design and its individual steps were defined by the multidisciplinary team in collaboration with the main stakeholder, the APATA association in Bragança, responsible for the management of the study area. The factors considered relevant for achieving the objective of the framework are presented in Fig. 2.

Fig. 2.

Overall view of the factors and criteria considered for developing the present framework (outer circles), with the aim to contribute towards landscape resilience (inner circles).

The framework was defined according to three main aspects: (a) the main factors identified as playing a role in the achievement of the objectives for the case study; (b) the combination of contributing factors to create the tasks leading to achievement of the target objectives; (c) the qualitative or quantitative data flow among the various tasks (Fig. 3). The tasks are hierarchically organised: intermediate tasks, producing results useful to pursue other tasks, and final tasks, which result directly in the evaluation of the proposed objectives (potential area of habitats to be restored or landscape area burnt for a given land management scenario) (Fig. 3).

Fig. 3.

Schematic representation of the proposed framework, with focus given to Part 1 of its implementation (present work): individual tasks that combine the important factors (solid rectangles), sequential implementation steps and qualitative and quantitative data flows.

The three main steps for implementing the proposed framework in the present work (Part 1 – co-creation of land management scenarios) were:

Step 1 – Definition of the Baseline and Alternative Land Use Maps – defined as current land use and the potential changes needed to maximise restoration of target habitats in the area.

Step 2 – Interaction with Stakeholders – where both land use maps were presented and discussed and the stakeholders’ suggestions in terms of opportunities and challenges of the presented land use maps were surveyed.

Step 3 – Creation of Landscape Management Scenarios – the stakeholders’ perceptions and suggestions were integrated into the land use maps to create scenarios of alternative forest management strategies that were translated into fuel model maps (Fig. 3).

Results

Framework development – Steps 1–3

Step 1 – definition of baseline and alternative land use

Current land use is dominated by woody vegetation (1080 ha); shrublands and heathlands are the most abundant land cover category (550 ha), followed by coniferous (300 ha) and broadleaf forests (230 ha) (Fig. 4, Table 3). Agriculture cover occupies a total of 452 ha in the area, dominated by temporary crops (250 ha), followed by orchards (105 ha) and farmland with natural and semi-natural areas (60 ha) (Fig. 4, Table 3). The output of the potential distribution maps for both Habitats 9230 and 9340 (the most suitable areas for restoration) are indicated in the areas shown in Fig. 4. The resulting area for the expansion of Habitat 9230, dominated by Pyrenean oak, was estimated as 392 ha; for Habitat 9340, dominated by holm oak, potential establishment of 30 ha was estimated. The 392 ha increase of Habitat 9230 would replace 104 ha of other woody cover types, namely 100 ha of shrublands and heathlands and 4 ha of pine forests. The highest impact would be, however, on agricultural cover, with a total of 288 ha replaced, between the various categories of farming, pasture and grasslands (236 ha), and orchards (52 ha) (Fig. 4, Table 4). Introducing holm oak forest would replace 30 ha of shrublands and heathlands (Fig. 4, Table 4).

Fig. 4.

Alternative soil use map, with the simulated introduction of the proposed two new habitats – Habitat 9230 and Habitat 9340, forests of Pyrenean oak and of holm oak, respectively.

Table 3.Land cover categories of the study area characterising the baseline land cover map.

Land use		Area (ha)	Total (ha)
Forest use	Broadleaf forests	230	1080
	Coniferous forests	300
	Shrublands and heathlands	550
Agriculture use	Spontaneous grasslands	5	452
	Pastures	12
	Complex cultivation patterns	20
	Agriculture with natural and seminatural areas	60
	Temporary crops and rice fields	250
	Orchards	105
Built areas		8	8
Total			1540

Table 4.Changes in land cover categories caused by the simulated establishment of new habitats in the study area – Habitat 9230 and Habitat 9340, forests of Pyrenean oak and of holm oak (in bold), respectively.

			Area (ha)	Total (ha)
Habitat 9230
Increase	Forest use	Pyrenean oak forest	392
Decrease	Forest use	Shrubland and heathlands	100	104
		Pine forest	4
	Agriculture use	Crops, pastures and grasslands	236	288
		Orchards	52
Habitat 9340
Increase	Forest use	Holm oak forest	30
Decrease	Forest use	Shrubland and heathlands	30	30

Step 2 – interaction with stakeholders: stakeholders’ perceptions survey

The opportunities identified by the stakeholders included the value of the new proposed habitat maps for increasing ecosystem services by providing multifunctional uses of the forest. These included potential mushroom harvesting within the new proposed habitat patches, the valorisation of firewood and biomass through silvicultural practices and the potential boost to the bioeconomy of the region through payment for ecosystem services (Fig. 5). In terms of challenges, a concern that such habitat areas should not be overvalued in the area was identified, and that they should be managed to include a variety of forest typologies to integrate other uses within the patches, such as grazing areas for livestock. Also, some stakeholders mentioned that the priority should be managing current land cover, particularly regarding fire risk, rather than changing it.

Fig. 5.

Interactions and indication of perceptions regarding the alternative land use map during the participative workshop.

Overall, the discussions highlighted the need to find an equilibrium between strict measures leading to habitat conservation and other socio-economic objectives, such as generating economic return, attracting and maintaining inhabitants, promoting traditional activities (e.g. hunting, fishing) and preventing wildfires and other risks. (Fig. 5). These (qualitative) challenges and opportunities identified by the stakeholder panel were incorporated into scenario building by defining a series of mid-and long-term management strategies, translated into (quantitative) fuel model maps (Fig. 6).

Fig. 6.

Fuel model maps of the study area according to the alternative forest management strategies, derived from the stakeholders’ perceived challenges and the corresponding target forest typologies.

Step 3 – landscape management scenarios: fuel model maps

We developed four management scenarios for forest in the study area following the participatory workshop where the inclusion of proposed areas for habitat restoration and conservation were presented to local, regional and national stakeholders from various sectors. The landscape management scenarios, derived from the land use maps, were represented by fuel model maps, each showing a combination of land cover categories and forest types resulting from a series of management strategies (Step 3, Figs 3, 6).

The baseline land use map was used to create two fuel model maps and the alternative land cover map was transformed into three alternative fuel model maps (Fig. 6). Each of the fuel model maps represents alternative forest management strategies derived from the stakeholders’ perceptions of opportunities and challenges. The base scenario represents the current situation, with no additional forest management apart from current practices. The base fuel model map is dominated by Fuel model 5, representing shrublands and heathlands, followed by Fuel model 1, which represents short grass characteristic of most agricultural uses and pastures (403 ha), and Fuel model 4, representing a dense, crown fire-prone pine forest, with no management since its plantation (approximately 40 years ago). The second scenario resulting from the baseline model (Scenario 1) represents a strategy focused on managing existing land use. In this scenario, the management of the existing pine forest is simulated in such a way that some parts of it are transformed into an open pine stand (120 ha), with Fuel model 2 assigned. Approximately 94 ha would be managed to decrease tree density and the understorey fuel load, resulting in a putative Fuel model 9 (Fig. 6, Table 5). The alternative land use map represents the potential changes in the cover of Habitats 9230 (Pyrenean oak forest) and 9340 (holm oak forest), obtained by active restoration measures. Scenario 2 intends to simulate the immediate establishment of 392 ha of Pyrenean oak and 30 ha of holm oak. In relation to the base scenario, this would produce a 201 ha increase in Fuel model 5 (shrubland), given the initial size of the planted oaks and the likelihood of regeneration of native shrubs in the planted areas. Scenario 3 increases the horizontal variability of the fuels by simulating mid-to-long term management of the established forests, creating different forest types. The taller Pyrenean oak and holm oak forests were simulated as open canopy woodlands (76 ha), with negligible surface fuel load (Fuel model 2). In the case of the Pyrenean oak forest, the shorter stands were simulated to grow into closed tall stands, similar to Fuel model 9 (61 ha) (Fig. 6, Table 4). Scenario 4 is the most complex in terms of management strategies, including the strategies defined for Scenarios 1 and 3 (Fig. 6, Table 5). These fuel model maps will be used to produce fire simulations (Part 2), and to assess which of the scenarios results in the best output in terms of reduced burnt area compared with the base scenario.

Table 5.Area occupied by each fuel model in the base scenario and resulting changes for each scenario build. Fuel models are ordered by increasing fuel depth. Coloured shading reflects the map legend on Fig. 6.

		Fuel model (ha)
		8	9	1	2	5	6	7	4
Scenario	Description/fuel depth (m)	0.061	0.061	0.305	0.305	0.61	0.762	0.762	1.829
Base	Current situation	0	159	402.75	0	126	511.75	26.5	311.75
1	Managed pre-existing pine	0	93.75	0	119.25	0	0	2.25	−213
2	Recently installed habitats	0	−18.75	−111.5	0	201	−68.5	−2.25	0
3	Managed new habitats	1.75	60.75	−111.5	75.75	44	−68.5	0	0
4	Scenarios 2 + 3	1.75	154.5	−111.5	195	44	−218.5	−2.25	−213

Discussion

The framework developed within this study provides a reliable and integrative approach to address forest management challenges in multiple-use forests in Portugal, in this specific case in the Montesinho Natural Park in Bragança. By integrating ecological modelling of species potential distribution, stakeholder perspectives and land cover scenarios, and assignment of appropriate fuel models, our framework provides a valuable tool for creating sustainable forest management guidelines, valuing the multifunctional uses of the landscape and multiple management objectives.

The baseline land cover analysis revealed that woody cover dominates the study area, with significant areas of shrublands, heathlands and coniferous forests. Agricultural land covered a smaller portion of the area, with temporary crops and orchards being the most common types. This analysis provided essential insights into current land use patterns and informed the basis for developing alternative scenarios. The results were consistent with the diverse array of vegetation types described for Montesinho Natural Park, involving mixed oak forests, dominated by Quercus rotundifolia Lam. and Quercus pyrenaica Willd., along with extensive conifer stands of Pinus pinaster Aiton and Pinus sylvestris L. (Aguiar 2000; Castro et al. 2010; Plieninger et al. 2010). Additionally, the park also includes areas of heathlands dominated by Erica australis L., with low shrublands formed by Genista tridentata L. (formerly Pterospartum tridentatum (L.) Willk.) and Erica umbellata L. (Aguiar 2000). These varied habitats contribute to the ecological value of the park, providing essential ecosystem services, supporting biodiversity and landscape diversity (Aguiar 2000; Castro et al. 2010; Plieninger et al. 2010; Fernandes 2013a, 2013b).

Through the participatory workshop, stakeholders played a central role in shaping the alternative forest management scenarios for the study area, guaranteeing their adjustment to the specific ecological and socio-economic contexts of the region. This approach is based on the premise that forest owners and other actors engaged in forest management can reject policy and behavioural changes that do not reflect their values, interests and motivations and will more easily adopt forest policies supporting their management activities. The methodology beneath the participatory workshop has already been used in similar studies, some related to wildfire management in Mediterranean ecosystems (Schaich and Plieninger 2013; López-Rodríguez et al. 2015; Valente et al. 2015b; Borges et al. 2017; Otero et al. 2018; Gonçalves et al. 2022). Fernandes (2013a) reinforced the relevance of involving diverse stakeholders in planning effective wildfire management strategies. Integration of the perceptions about opportunities and challenges from the stakeholders present in the workshop (public entities, research institutions and representatives of local communities) ensured that the management scenarios were socially and politically acceptable, and ecological and economically grounded within the study area ecosystem and Montesinho landscape. Such an approach increases collaboration between different actors, the legitimacy of management decisions and the chances of successful wildfire mitigation initiatives (Fernandes 2013a). It was possible to observe the trade-off between the best and the most relevant afforestation approaches and land use conflicts as perceived by stakeholders. In addition to reflecting reasonably balanced ecological, social and economic trade-offs, the generated scenarios must also integrate wildfire mitigation strategies.

Comparative analysis of the different scenarios resulted in different levels of potential wildfire risk within the Montesinho area when translated into fuel models, emphasising the significance of considering vegetation dynamics in wildfire management planning (Zhang et al. 2017). For instance, scenarios prioritising intensive habitat restoration (Scenario 2), which are dominated by Fuel model 5, may lead to increased vegetation density and fuel loads, potentially increasing wildfire hazard (Zhang et al. 2017). Conversely, scenarios that focus on the management of existing land cover, such as Scenarios 1 and 4, and management of installed habitats, such as Scenario 3, incorporating stakeholders’ perceptions and suggestions, include techniques for managing vegetation that lessen fuel loads and improve landscape resistance to fires. A comparative analysis of the proposed land management scenarios, through the relationship between the NFFL models used and the forest types proposed and analysed for the Portuguese mainland (Fernandes 2009; Fernandes and Loureiro 2022) in the main components that define fire hazard, offers some insight into the potential decrease in fire propagation risk attained by Scenario 1, compared with the baseline, and attained by Scenarios 3 and 4 compared with Scenario 2. Management of Quercus pyrenaica and Quercus rotundifolia habitats incorporated into Scenario 2, so as to develop Scenarios 3 and 4, partially replaces NFFL Fuel model 5 by Fuel models 2 and 9. In parallel, management of pre-existing land cover resulted in the evolution of base Scenario 2 to Scenario 1 and in the development of Scenarios 3–4. With this land management strategy, the replacement of NFFL Fuel model 5 (comparable with diverse low closed) and 4 (comparable with Pinus sylvestris, low closed) by Fuel models 2 (comparable with Quercus suber, open tall) and 9 (comparable with Pinus sylvestris, open tall, or Quercus suber, closed tall), decreases fire hazard scores for all the components of fire hazard (Fernandes 2009; Fernandes and Loureiro 2022).

Nevertheless, current fuel models may lack sensitivity to respond to all the nuances of the different forest management scenarios, indicating the need for further research and model refinement for specific ecological conditions, combination of forest species and management practices (Loureiro and Fernandes 2021; Fernandes and Loureiro 2022). However, current technical limitations related to the proposed fire simulation program (WRF SFire) required the use of the NFFL fuel models.

These limitations, though, should not hinder the general objective and applicability of the proposed framework, which is flexible enough to accommodate further refinements in all its implementation steps. Notwithstanding the arguably insufficient fuel models depicting local vegetation types, their use in the current framework brings an added dimension to the land use maps, appropriately illustrating the potential effects of different land management strategies not only in shaping a multifunctional landscape, but also in their putative relative effect in terms of fire hazard (Wilson et al. 2018). The assessment of the effectiveness of each scenario in terms of relative fire hazard will nonetheless benefit from further evaluation through fire simulations using WRF SFire (see Part 2). The potential of the designed management scenarios for promoting (or mitigating) fire spread under different weather conditions will be addressed in Part 2.

Conclusions and further work

This study demonstrates the relevance of participatory approaches in environmental management and the importance of interdisciplinary collaboration in addressing complex environmental challenges. A participatory and proactive approach was used to define the most interesting forest management scenarios, targeting habitat restoration goals framed in the regional context. The subsequent integration of those scenarios with fuel models demonstrated the fire hazard potential of each scenario, and the implications and impacts of management choices, allowing better forest management planning and supporting integrated management decisions. This comprehensive approach, scientifically based and supported by stakeholder engagement, is crucial for building landscape resistance and resilience to wildfires and ensuring sustainable land management practices in Portugal and beyond.

Supplementary material

Supplementary material is available online.