Research on rural fire focus more on describing the specific processes and components of fire behaviour and regime, rather than integrating them into fire prevention models. This section attempts to bring together sectoral perspectives, to develop an integrative, fire resilient and sustainable, conceptual model, namely:
I) Landscape fire resilience
The concept of resilience is the capacity of systems to reorganize and recover from change and disturbance without changing to other states (Walker et al., 2004; Ahern, 2011). This concept linked with the comprehension of the landscape as a system (Magalhães et al., 2007) leads to the definition of landscape fire resilience as the capacity of landscape in absorbing the disturbance caused by rural fires without losing its function, structure, and identity and ultimately weakening fire frequency and intensity or magnitude. The landscape fire resilience is determined by several factors related to a) fire behaviour; b) flammability of tree species; c) landscape discontinuities; and d) wildland-urban interface.
a) Fire behaviour
Fire behaviour depends on climatic factors, e.g. temperature, precipitation, and wind, combined with other biophysical factors related to the river basin morphology, including slope, aspect and altitude (Rothermel, 1983; Moreira et al., 2007; Heyerdahl et al., 2010). Although changes in weather conditions can lead to unpredictable fire behaviour (Coen, 2015), the biophysical factors can contribute, to some extent, to the reduction of fire risk and, therefore, should be taken into account in land-use planning, namely:
North aspect hillslopes, with a slope higher than 25%, by receiving less radiation throughout the year, burn less than the other hillslope aspects (Oliveira et al., 2014);
The fire progression speed doubles for every 10º (about 17%) increase in slope, and it can rise continuously in steep hillslopes from bottom to ridge, by approximately 5–6 Km/h of fire speed (Viegas, 1989);
Above slopes higher than 30º (57%), the relationship between the slope and fire speed is almost exponential (Viegas, 2006);
When the fire reaches the top of the river basin if it does not progress to the opposite side due to the hillside breeze, it begins to plow along the contour lines losing speed.
b) Flammability of tree species
The post-fire observation shows that tree species do not burn equally. Specifically, native species burn less than Eg and Pp, and regenerate better even in severe fire circumstances. These assumptions are substantiated by several studies, namely Dickinson et al. (2016), and Calviño-Cancela et al. (2017). Also, regarding fire proneness, Silva et al. (2009) verified a tendency in different forest types, in decreasing order: maritime pine forests, eucalyptus forests, unspecified hardwood forests, unspecified coniferous forests, cork oak, chestnut, and holm oak forests. Calviño-Cancela et al. (2017) state that native species are more resistant to fire than those introduced. Moreover in the USA even trees, such as native oak trees, have been replaced by acer, poplar, and beech, as they are more resistant to fire with less flammability due to their higher rate of litter decomposition (Dickinson, et al, 2016). In the same way, Pereira et al. (2014) affirmed that there are various degrees of fire risk according to different land-uses e.g. agricultural fields, pastures, and, ultimately, spaces without shrubs or tree vegetation decrease fire risk. However, these last situations must be avoided, and covered with permanent herbaceous to prevent soil erosion.
Based on the several post-fire evidence, a diversified landscape is more resilient to fire than a monoculture landscape of fast-growing tree species.
c) Landscape discontinuities
Aiming for an improvement of fire prevention characteristics of landscapes, several authors (Carmo et al., 2011; Moreira et al., 2009, Povak et al., 2018) justify the need to have landscape discontinuities with several types of land-uses and fuel break networks, among others:
Agee et al. (2000) propose wide areas of shaded fuel-breaks networks covered with low-fuel vegetation areas, coupled with fuel control strips. Also, the Forestry Commission Practice Guide (2014) highlighted two types of fire-resistant networks in the landscape: the «fire-breaks» without vegetation and «fire-belts» with broadleaved trees, which can occur separately or combined.
According to Povak et al. (2018) the streams and valley bottoms play a fundamental role in establishing landscape discontinuities. From a river basin perspective, these two landscape components are more crucial to reduce the size and intensity of the fire than ridges or hilltops.
Furthermore, Heyerdahl et al. (2010) points to the necessity of introducing fire retardant strips along the contour lines when the hillside is too long to avoid top-down and down-up fire;
Swales or infiltration ditches, constructed along the contour lines with a berm downslope planted with native broadleaf trees and associated ponds (Mollison, 1988) can function as linear fire-belts. These structures also reduce soil loss by erosion and increase the water basin's total water flow through infiltration.
The mulching technique is a way to increase soil and water conservation, and reduce post-fire This technique helps to increase water infiltration and retention in the soil, leading to lower species combustibility, along with a reduction in soil erosion (Fernández et al., 2010; Keizer et al., 2018).
d) Wildland-Urban Interface
The wildland-urban interface (WUI) is an area of transition between wildland and urban agglomeration, in or adjacent to wildfire prone areas, with a high vulnerability and wildfire risk (Calviño-Cancela et al., 2016). Mapping and evaluating WUI areas are fundamental to develop strategies to reduce fire risk (Bento-Gonçalves and Vieira, 2020), as several authors (Gibbons et al., 2018; Badia et al., 2019) refer to the need to maintain a clean strip of combustible material in the surroundings of settlements (100 m) or near isolated buildings (up to 30 m).
II) Ecological sustainability
Ecological sustainability is a broad concept that ensures landscape ecological quality (Callicot and Mumford, 2002; Termorshuizen et al., 2007). This concept is closely related to the capacity of landscape adaptability and resilience. In this study work, ecological sustainability is regarded as a landscape quality dependent on the value and conservation of the natural resources. It is synthesized in an ecological network and dependent on the most adequate land uses, set by ecological land suitability assessment.
a) Ecological Network
The Ecological Network (EN) is recognized as a system of landscape structures or ecosystems (Forman, 1995; Magalhães, 2001) that provides the necessary physical and biological conditions for maintaining or restoring ecological sustainability, connectivity, and biodiversity. Therefore, the EN components correspond to highly valuable ecosystems, which represent specific ecological functions, directly influenced by hydrologic availability, soil genesis processes and fertility, plant biodiversity (species), habitat resources and climate (Cunha and Magalhães, 2019). The EN mapping is rooted in landscape ecology theory (McHarg, 1967; Ahern, 1995; Magalhães, 2001; Fabos, 2004; Lennon et al., 2015; Liquete et al., 2015). The EN was mapped for mainland Portugal (Magalhães et al., 2013b; Cunha and Magalhães, 2019), covering several planning scales.
In this context, the Ecological Network (EN) assumes a holistic view of land use planning and biodiversity conservation and constitutes the core of the broader Green Infrastructure (GI) framework (EC, 2013b; Civic and Jones-Walters, 2015) and a planning tool that supports the European Biodiversity Strategy 2030 (EC 2020).
b) Ecological land suitability
The Food and Agriculture Organization (FAO, 1976) defines land suitability as the fitness of a given type of land for a determined use. The earliest application of this concept was from the American landscape architects (e.g. Tyrwhitt, 1950; Lewis, 1964) in the late nineteenth and early twentieth-century using hand-drawn overlays to locate territories suitable for future construction (Collins et al., 2001). The land-use suitability analysis using this overlay technique has been applied by different authors, such as McHarg (1967), which used ecological inventory as criteria to map the best land use and recently applied with geographical information systems (GIS) tools in several contexts.
Based on the same technique, the ecological land suitability analysis considers the abiotic and biotic characteristics thresholds for each land use. This analysis allows more careful land-use planning under the resilience of the landscape to support it. By applying this analysis at the national level, Magalhães et al. (2016) concluded that 22% of the Portuguese territory is suitable for agriculture. Still almost half of it is covered with other land uses. Moreover, about one-third of the Portugal area is ideal for conservation forest but is occupied by exotic species. This analysis also identifies the potential for expanding different species, such as cork oak, holm oak, carob, and chestnut.
The proposed model considers land-use planning as a tool that integrates the above perspectives of landscape fire resilience and ecological sustainability, into a spatial framework.