Next Article in Journal
A Review of Biomass Estimation Methods for Forest Ecosystems in Kenya: Techniques, Challenges, and Future Perspectives
Previous Article in Journal
Establishing Reference Models for Ecological Restoration—Case Study from Colorado National Monument, USA
Previous Article in Special Issue
Long-Term Trajectory Analysis of Avocado Orchards in the Avocado Belt, Mexico
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Land
Volume 14
Issue 9
10.3390/land14091872
land-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Mapping Long-Term Wildfire Dynamics in Portugal Using Trajectory Analysis (1975–2024)

by
Bruno Barbosa
1,2,
Ana Gonçalves
1,2,
Sandra Oliveira
1,2 and
Cláudia M. Viana
1,2,*
1
Centre of Geographical Studies, Institute of Geography and Spatial Planning, University of Lisbon, 1600-276 Lisbon, Portugal
2
Associate Laboratory TERRA, Instituto Superior de Agronomia, Tapada da Ajuda, 1349-017 Lisbon, Portugal
*
Author to whom correspondence should be addressed.
Land 2025, 14(9), 1872; https://doi.org/10.3390/land14091872
Submission received: 4 August 2025 / Revised: 4 September 2025 / Accepted: 9 September 2025 / Published: 13 September 2025
(This article belongs to the Special Issue Applications of Trajectory Analysis to Describe Changes in a Time Series of Maps)
Browse Figures
Versions Notes

Abstract

Wildfire regimes in Mediterranean landscapes are becoming increasingly unpredictable, driven by the combined effects of climate change, land-use transitions, and socio-economic pressures. Traditional metrics such as burned area or ignition points often fail to capture the complexity of the temporal and spatial recurrence of fire events. To address this gap, we apply, for the first time, a trajectory analysis framework to wildfire occurrence data across mainland Portugal (1975–2024), using pixel-level binary time series at 100 m resolution. Originally developed for land cover change detection, this method classifies each pixel into sequences representing distinct temporal patterns (e.g., stability, gains, losses, or alternations) over defined periods. Results reveal a predominance of stable absence and alternation-type trajectories, particularly “All alternation gain first”, which points to recurrent yet irregular fire activity. Regional differences further highlight the influence of divergent socio-ecological contexts. The findings suggest that fire regimes in Portugal are not only recurrent but structurally dynamic, and that trajectory-based classification offers a novel and valuable tool for long-term monitoring and regionally adapted fire management. Applying this method to wildfire data required specific adjustments to account for the unique temporal and thematic characteristics of fire regimes, ensuring a meaningful interpretation of the results.

1. Introduction

Wildfire regimes are shifting globally, where climate change, land-use transitions, and socio-economic pressures interact to amplify fire frequency, intensity, and ecological impacts [1,2,3]. Across fire-prone landscapes worldwide, the duration of the fire season is expanding, extreme fire events are becoming more frequent, and fire behaviour is increasingly unpredictable due to compounded drivers such as drought, heatwaves, and changing vegetation patterns [4,5,6,7]. These trends are not only linked to climate change but also amplified by socio-economic dynamics, including rural depopulation, land abandonment and subsequent fuel accumulation, and ineffective adaptive land management strategies [8,9,10,11]. Recently, the United Nations Environment Programme has also stressed that a lack of effective investment in wildfire prevention remains one of the most pressing challenges at the global scale [12]. Mediterranean areas are particularly sensitive to such changes; although historically shaped by fire as a natural disturbance, the current frequency and intensity of fire activity far exceeds previous levels, resulting in the loss of biodiversity, ecological degradation and ongoing threats to human life and livelihoods [1,13,14]. Recent studies indicate that traditional fire regimes across Mediterranean ecosystems are undergoing a fundamental regime shift (a structural transformation in the spatial, temporal, and ecological dimensions of wildfire behavior) [2,15]. This transition extends beyond simple increases in fire frequency or burned area, representing instead a systemic alteration in fire dynamics characterized by, for example, shortened fire-return intervals and increased resistance to suppression efforts due to more extreme fire weather conditions [16,17,18].
Portugal has been recurrently affected by large wildfires, with some events resulting in tragic human losses and long-term ecological disruption [10,13,19]. Over the past five decades, fire regimes in mainland Portugal have evolved significantly, challenging existing strategies for fire mitigation, land management and policy intervention [14,20,21]. Traditional practices regarding the use of fire by rural communities have changed remarkably in the last century [22], linked to farmland abandonment and the subsequent fuel loads increase seen throughout most of the country. In Central Portugal specifically, Moreira et al. [23] analysed fire regimes at the civil parish level, for two periods spanning 1975 to 2018, and uncovered a general trend for fire incidence and burn frequency to increase across time, coupled with a decrease in population density and farmland coverage, and instead an increase in tree plantations.
In this context, understanding long-term wildfire dynamics at multiple spatial and temporal scales has become increasingly important. Earlier studies, such as Catry et al. [24], mapped ignition risk based on population density, land cover, and accessibility, demonstrating the feasibility of national-scale models. Kanevski & Pereira [25] used fractal analysis to reveal spatial heterogeneity and clustering patterns in fire occurrence (1975–2013). Nunes et al. [26] expanded this perspective by analysing spatial fire patterns over more than three decades, identifying persistent high-risk zones shaped by both environmental and anthropogenic factors. Bergonse et al. [27] analysed fire patterns and their biophysical drivers for the civil parishes of the Central region in Portugal, using fire data for 44 years. Silva et al. [28] explored changes in the length of the fire season using data on fire activity between 1980 and 2018.
Despite their valuable contributions, these studies rely mainly on aggregated metrics of fire occurrence and burned area across broad spatial and temporal scales, which may mask the inherent spatiotemporal complexity of fire regimes. In particular, such approaches often overlook sequential dynamics, such as the repeated burning of specific areas or the transitions between burned and unburned states, which can instead be captured through pixel-level, time-series analyses [6,15]. Addressing this gap requires methodological frameworks capable of tracking the sequential trajectories of fire activity at fine spatial and temporal resolutions.
To respond to this challenge, we propose, for the first time, the application of trajectory analysis to long-term wildfire data in Portugal. Originally developed for land use and land cover (LULC) change studies [29], this methodology has not yet been explored in wildfire research. Its potential lies in uncovering spatial patterns and sequential dynamics at fine spatio-temporal scales, which remain obscured by more aggregated approaches. In this paper, we apply a Trajectory Analysis framework across mainland Portugal, using wildfire occurrence data from 1975 to 2024, at a 100 m resolution. Our objectives are threefold: (i) to test the application of this methodology to uncover long-term trends of fire activity in mainland Portugal; (ii) to classify pixel-level fire histories and identify dominant trajectories in different Portuguese regions; (iii) to detect spatial patterns of persistence, emergence, and alternation in fire regimes. By transferring trajectory analysis to wildfire research, this study opens new methodological pathways for analysing long-term fire dynamics. In doing so, we contribute to a more detailed understanding of wildfire regimes and provide tools to support more adaptive and spatially differentiated fire management strategies.

2. Materials and Methods

2.1. Study Area and Dataset

The study focuses on mainland Portugal (Figure 1a), a country characterized by a Mediterranean-type climate, with Atlantic influences in the North. It covers an area of about 89.000 km2, and over a third of the territory is occupied by forests and shrubland. The main tree species are maritime pine (Pinus pinaster), eucalyptus (Eucalyptus globulus) and cork oak (Quercus suber), with an irregular distribution throughout the territory; tree plantations of maritime pine and eucalyptus dominate the Central region, whereas cork oak and holm oak (Quercus ilex) prevail in Alentejo, in the south. Other oak species are more present in the North region (e.g., Castanea sativa) and shrubland is abundant in central mountainous inland areas [30]. Portugal has a recurrent wildfire activity, being one of the most affected countries of Southern Europe, recording an annual average burned area of about 113,000 ha between 1980 and 2023 [31]. The most affected regions are Norte and Centro (Figure 1b), where the maximum number of times burned was 19 times over 50 years, although the pixels that burned 10 times or more correspond to less than 1% in both regions.
The analysis covers 50 years of data, ranging from 1975 to 2024. Fire occurrence data, in the form of fire perimeters representing the area burned, were obtained from the Portuguese Forest Services [32] and converted into annual binary rasters (burned/not burned) at 100 m spatial resolution. The 100 m resolution was adopted to harmonize the official fire perimeter dataset (1975–2024) and to minimize inconsistencies over time, obtaining a balance between the spatial detail required for a model applied at national and regional levels and computational feasibility., Due to the raster resolution, fires smaller than 1 hectare were excluded. This process resulted in a final dataset of 55,014 fire perimeters, burning approximately 56,070 km2 since 1975. which corresponds to about 63% of the territory of mainland Portugal.

2.2. Trajectory Analysis

Trajectory analysis, as presented by Bilintoh et al. [29] has emerged as a powerful method for detecting landscape change over time, particularly when applied to categorical time series at the pixel level. By classifying each pixel according to its sequence of change or stability (e.g., gain, loss, persistence, alternation), this method enables a clearer understanding of both the magnitude and direction of landscape transformations. In this study, we adapted the methodology proposed [29] to the context of wildfires in Portugal, analysing change through a time series of binary fire presence/absence maps. This approach enables each pixel to be classified according to its complete temporal fire history, allowing for the detection of cyclical reburning, dynamic hotspots, and evolving fire regimes. These insights complement existing spatial and socio-environmental models by capturing the temporal depth of fire disturbance patterns. Furthermore, the decomposition of gross change into quantity, exchange, and alternation components, as adopted in Bilintoh et al. [29], builds on the earlier distinction between quantity and allocation disagreement proposed by Pontius & Millones [33], and provides a structured framework for disentangling complex patterns of change and identifying areas undergoing either gradual shifts or more volatile dynamics.
The analysis was implemented using the R package ‘timeseriesTrajectories’ version 1.0.3 [34], and while this approach typically yields three standard outputs, we adapted the results to focus specifically on temporal wildfire patterns in our dataset. Our implementation proceeds through three analytical stages:
(i)
annual burned area quantification: the primary output calculates the spatial extent of the burned area (the binary variable of interest) for each year in the time series;
(ii)
Pixel-level fire frequency analysis: we calculated the total number of burning events and the frequency of state transitions (burned/unburned) for each pixel over the study period;
(iii)
trajectory classification: the algorithm assigned each pixel to a predefined trajectory category based on its complete fire history.
The trajectory classification distinguished eight categories (Figure 2). Categories 1–6 correspond to pixels where fire occurred at least once during the study period (loss or gain with/without alternation), whereas categories 7 and 8 represent stable conditions, with persistent fire presence or stable absence of fire, respectively.
Land 14 01872 g002
Figure 2. Example of binary variable trajectories between 5 years (times t1 to t5). “0” indicates fire absence and “1” indicates fire occurrence. The code column presents the color applied to each trajectory and if it expresses absence (0) or presence (1) in the final year of the period. Adapted from Bilintoh et al. [29]. The colors correspond to the legend of the trajectory maps (e.g. Figure 3).
Figure 2. Example of binary variable trajectories between 5 years (times t1 to t5). “0” indicates fire absence and “1” indicates fire occurrence. The code column presents the color applied to each trajectory and if it expresses absence (0) or presence (1) in the final year of the period. Adapted from Bilintoh et al. [29]. The colors correspond to the legend of the trajectory maps (e.g. Figure 3).
Land 14 01872 g002
Land 14 01872 g003
Figure 3. Fire trajectories per decade, between 1975 and 2024.
Figure 3. Fire trajectories per decade, between 1975 and 2024.
Land 14 01872 g003
The resulting trajectories exhibit different patterns of evolution over time, varying between gains, losses, alternations, and stable conditions. Gain trajectories are represented by areas that shifted from unburned to burned during the analysis period. These can occur without alternation (trajectory 3), involving a single transition, or with alternation (trajectory 4), where multiple transitions occur but fire is still present in the final year. Loss trajectories (trajectories 1 and 2) show the opposite pattern behaviour, representing areas that shifted from burned to unburned. Trajectories 5 and 6 also involve multiple transitions over time, but both the initial and final stages are the same (either burned or unburned). Lastly, stable occurrence trajectories represent pixels that remained either burned or always unburned consistently over the entire period. In the particular case of the wildfire process, a stable burned category is typically unexpected, as the recurrence of fire at the same location in consecutive years is unlikely due to the dependency on vegetation regrowth and fuel accumulation.
To explore how the definition of time intervals may influence the results on wildfire trajectories, we applied two different temporal aggregations schemes: (a) five sequential 10-year intervals: 1975–1985, 1985–1995, 1995–2005, 2005–2015, and 2015–2024; and (b) two sequential 25-year intervals—1975–2000 and 2000–2024 (Figure 4). Each period overlaps with the following in the last/first year, to ensure continuous temporal coverage and to include the identification of transitional patterns for the last year of each period (which corresponds to the first year of the following time interval).
To understand how spatial variation influences wildfire trajectories, we divided the study area into three distinct regions. First, we considered the entire mainland territory of Portugal to capture the overall behaviour of wildfire trajectories. Subsequently, we conducted a more detailed regional analysis (at the NUTS-2 level), focusing on two regions that have been significantly affected by wildfires throughout the study period—namely, the Norte and the Algarve regions (Figure 1).

3. Results

3.1. Trajectories for Five Decades of Wildfire Occurrence in Mainland Portugal

Until 1995, the maximum fire frequency recorded was five times per decade. Between 1995 and 2015, this number increased to six times, observed in restricted areas of the north and central-east regions. In the most recent decade (2015–2024), the maximum number of fire events decreased to four, with large extents of the territory burning only once in a decade, especially in central mainland Portugal (Figure 5).
The spatial patterns of the trajectories show that stable fire presence does not occur in any of the five decades analysed. In contrast, stable absence dominates across all periods, starting with 91.3% of the territory between 1975 and 1985 and gradually decreasing thereafter (Table 1), reaching its lowest value (83.6%) in the decade 1995–2005. This drop suggests that nearly 9% more of the territory has burned between 1995 and 2005. In the following decade (2005–2015), the area affected by wildfires decreased approximately by 6%, followed by a 2% increase in the most recent decade.
Besides stable absence, the second most prevalent trajectory throughout the decades is “All alternation gain first”, affecting between 4.6% and 11.4% of the territory, depending on the timeframe (Figure 3). This trajectory depicts the occurrence of multiple changes transitions in pixel status (burned/unburned) over time, but both the first and the last years of the decade have not burned, changing to burned pixels in-between intermediate years in that period. The third most frequent trajectory varies across decades, following an alternating pattern: “Gain without alternation” is more prominent in the first and third periods (1975–1985 and 1995–2005), affecting over 2% of the territory; “Loss without alternation” appears more strongly in the second and fourth decades (1985–1995 and 2005–2015); in the final decade (2015–2024), “Gain without alternation” re-emerges as the third most common trajectory, despite covering a smaller area.

3.2. Trajectories for 25-Year Intervals of Wildfire Occurrence in Mainland Portugal

When considering 25-year time intervals, the maximum fire frequency reached 11 events in both periods (1975–2000 and 2000–2024) (Figure 6). A small proportion of pixels have changed fire status, from burned to unburned (0.93% in 1975–2000; 1.51% in 2000–2024) or vice versa (1.53% and 1.47%, respectively). In 1975–2000, 247 pixels transitioned between fire states burned and unburned more than 15 times over the 25-year period, a number that increased to 456 pixels in the period 2000–2024, indicating highly dynamic fire regimes in these areas.
Similar to what was found for the decadal analysis, stable absence remained the dominant trajectory, although with slightly lower values. In the first interval (1975–2000) this trajectory occupied approximately 82% of the mainland Portugal, decreasing by around 7% in the second period (2000–2024). In both periods, the second most prevalent trajectory was “All alternation gain first”, which increased from 15% to over 21% of the territory between the two periods (see Table 1). This pattern indicates that these pixels were unburned in both the first and last years of the interval but experienced fire in at least one intermediate year.
The third most frequent trajectory shows contrasting results in relation to the decadal analysis. In the first period, trajectory 4 (“Gain with Alternation”) covered 1% of the area, and in the second period it maintained a similar extent but was slightly surpassed by trajectory 2 (“Loss with Alternation”) (Table 1). Trajectory 4 shows areas that transitioned from unburned to burned over time, with fluctuations between states; while Trajectory 2 indicates the opposite trend, i.e., areas that began the period burned and ended unburned, also with alternating fire status.
The spatial distribution of the trajectories exhibits distinct patterns between the two periods (Figure 7). Since 2000, trajectory 6 (“All alternation gain first”) expanded considerably, particularly in the southern Portugal. Trajectories 3 and 4 (“Gain without alternation and Gain with alternation”, respectively) are visibly concentrated in upper-west sections of the central region, whereas in the first period the pattern where was more dispersed across the northeast of the country. In contrast, Trajectories 1 and 2 (“Loss without alternation and Loss with alternation”, respectively) maintained a dispersed pattern in both periods but shifted their predominant location from the northwest to the northeast of the mainland (Figure 7).

3.3. Trajectories of Wildfire Occurrence at Regional Level

The regional level analysis shows different trends between the Norte and Algarve regions (Table 2). The trajectory related with stable absence consistently occupies a larger proportion of the territory in the Algarve across all timeframes. In terms of fire frequency, the maximum number of times a pixel burned in a decade ranged from five to seven in the Norte region, and between two and three in Algarve (Figure 8 and Figure 9). In the southern region, wildfire occurrence increased most significantly during the third decade (1995–2005). In the Norte region, fire occurrence also increased during the third and fourth decades (1995–2005–2015), before decreasing by approximately 5% in the most recent decade (2015–2024).
In both regions, trajectory 6 (“All alternation gain first”) consistently ranked as the second most frequent in all timeframes (Table 2). However, the third most prevalent trajectory varied across regions and timeframes. In the first decade, trajectory 3 (“Gain without alternation”) comes third place in both regions, covering 4% of the Norte region, particularly concentrated in the western sector and affecting smaller areas in the Algarve (Figure 7). In this southern region, trajectory 3 became more prominent in the second decade (1985–1995), where it covered around 2% of the territory, primarily in the northwest (Figure 8). In this region, the remaining trajectories were either marginal or absent across most periods. In contrast, the Norte region exhibited greater diversity: the third dominant trajectory alternated between trajectory 3 (“Gain without alternation”) and trajectory 1 (“Loss without alternation”), depending on the decade. For example, between 2005 and 2015, trajectory 1 covered approximately 4.4% of the Norte region, with a dispersed pattern across western and central sub-regions (Figure 7).
From 1995 onward, wildfire affected areas began expanding eastward, with fire concentration peaking in the decade 2005–2015, a spatial pattern not observed in the other decades. When comparing the two 25-year periods, the difference is that the Algarve region is notorious. The proportion of stable absence decreased by 18% between 1975–2000 and 2000–2024. This reduction was largely offset by a sharp increase in trajectory 6 (“All alternation gain first”), which came to occupy nearly 28% of the region in the most recent period (Table 2).

4. Discussion

Wildfire occurrence is a major issue in Portugal, with records since 1975 showing the persistence of this environmental process across several regions, particularly in Norte, Centro and Algarve [26,35]. These regional patterns are partly explained by differences in forest ecosystems. For instance, the dominance of dense planted forests with fire-adapted species such as maritime pine (Pinus pinaster) and eucalyptus (Eucalyptus globulus) is linked to the occurrence of intense fires under extreme weather conditions, as observed in the year 2017 in Centro region [36]. The Alentejo region, despite being the driest and hottest of the mainland, is fuel-limited and agroforests predominate, including managed forests of cork and holm oak, therefore showing much lower records of fire activity [9,35].
This general long-term dynamic, however, conceals differences in time and space which require a more detailed examination. In this study, we applied a trajectory analysis framework to explore potential variations in fire occurrence trends in defined timeframes, and analysed how shifts in trajectories over time are expressed spatially at national and regional scales. By using a pixel-level classification at 100 m resolution, we reconstructed detailed fire occurrence history across mainland Portugal, identifying dominant trajectories for each period analysed.
The predominance of stable absence and alternation trajectories emerges as a relevant finding. This pattern is consistent with the high spatiotemporal variability of wildfire occurrence in Mediterranean landscapes. The absence of stable fire presence supports the assumption that wildfire is an episodic phenomenon, driven by fuel accumulation, ignition sources and climatic variability. These results reinforce the idea that fire trajectories differ substantially from land cover change dynamics, which are more likely to show persistent trends (e.g., urban expansion or forest loss) [37,38]. The trajectory “All Alternation gain first” consistently ranking second in all timeframes, reflects the cyclical re-burning patterns previous observed in Portuguese fire-prone landscapes [39,40]. This pattern may indicate areas where suppression and fuel management strategies are insufficient to prevent repeated ignitions, which in Portugal are predominantly of human origin, similarly to what is observed in other countries in Southern Europe [31]. Recurrent ignitions are often associated with land-use practices and use of fire, such as pasture renewal, burning of agricultural residues, and shrub clearing. These activities, while historically linked to rural management, have dramatically changed over the last century [22] and may contribute to the persistence of fire-prone landscapes, suggesting loops between vegetation recovery and renewed susceptibility to burning. Our results are in line with studies that identified long-term persistence and spatial clustering of fire activity in Portugal [26,41]. However, unlike conventional approaches focused on annual burned area or ignition points, our method reveals intra-pixel transitions that highlight temporal complexity, which is often obscured by aggregated metrics such as total burned area or ignition density.
The regional diversity of trajectories across Portuguese regions has strong implications for fire governance. In the Norte region, where more diverse trajectories emerged (including gain, loss, and alternation), fire regimes appear to reflect more complex socio-ecological dynamics [42]. In contrast, the Algarve region shows signs of a more recent shift, with a clear expansion of alternation-type trajectories towards the eastern part of the region since 1995. This pattern may reflect changing land use practices (e.g., agricultural abandonment), increased exposure of wildland-urban interface (WUI), or evolving climatic conditions [43,44,45]. These differences suggest that current fire mitigation strategies are likely insufficient to tackle challenges that are region-specific. Instead, strategies should be tailored to the dominant trajectory types observed. For example, in areas characterized by “Gain with alternation”, where fire occurrence is intermittent but recurring, priority could be given to long-term fuel management (such as prescribed burning, mechanical clearing or agroforestry mosaics), as part of a holistic fire management program, combined with proactive suppression capacity, given the likelihood of reburning after vegetation recovery [46]. These landscapes may be caught in cycle loops where fuel regeneration and insufficient intervention foster repeated ignitions [47]. In contrast, areas showing recent gains without alternation may reflect emerging fire regimes, possibly linked to land abandonment, new ignition pressures, or changes in vegetation structure [11]. In such areas, early detection systems, community-based surveillance, and ignition prevention measures (such as infrastructure maintenance or outreach in wildland-urban interfaces) could be more effective. Differentiating these patterns based on temporal trajectories supports a shift from reactive to anticipatory governance, by aligning intervention strategies with the underlying disturbance dynamics [48].
While trajectory analysis provided novel and relevant insights, its application to wildfire occurrence (a process with high interannual variability), required methodological adjustments. The results are sensitive to the choice of time intervals and starting years, which condition the classification outcomes. The definition of two different sets of timeframes (five 10-year intervals and two 25-year intervals) allowed testing for potential differences by comparing outcomes across scales. While most frequent trajectories remained relatively consistent, their spatial expression and proportional coverage varied across temporal scales, highlighting the sensitivity of trajectory outcomes to the configuration of temporal windows. Another critical aspect is the predominance of alternation type trajectories regardless of the length of the timeframes. Indeed, our results show that, besides stable absence, the trajectory representing “All Alternation gain first” comes second in relative proportion in all timeframes. This specific trajectory depicts multiple transitions over time, with the first and the last year of the period classified as unburned and shifting to burn status several times along the period. Compared to land use applications of trajectory analysis (e.g., [29]), our adjustment for fire occurrence analysis required thematic reinterpretation, particularly regarding the plausibility of stable presence and the predominance of the trajectories representing alternation trends in relation to those depicting gains and losses. These differences illustrate the need for process-specific adaptations when applying trajectory methods to phenomena with inherently different dynamics, to ensure an adequate interpretation of the results obtained.
Future research could explore the integration of climatic, environmental and socioeconomic variables to further explain the observed fire trajectories patterns. Variables representing population aging and declines in agricultural activity, for example, have been linked to land abandonment and fuel accumulation, which in turn increase wildfire hazard [26,49]. Similarly, incorporating drought indices or temperature anomalies would help distinguish whether certain trajectories result from structural landscape changes or from extreme fire weather conditions [6,50]). This multidimensional integration would support more robust, spatially explicit interpretation of fire regimes, better aligned with the social and environmental conditions in which they operate.

5. Conclusions

This study demonstrates that trajectory analysis, when adapted to wildfire occurrence, offers more than a novel classification tool, it represents a conceptual shift in how we understand the temporality of fire regimes. By encoding the history of fire presence at the pixel level over five decades, the approach moves beyond cumulative metrics and reveals the structural volatility and spatial memory embedded in fire-prone landscapes. The findings in this study show that fire regimes in Portugal are not simply characterized by recurrence, but by temporal heterogeneity and intermittent disturbances. These dynamics are often obscured by traditional burned-area metrics. The predominance of alternation type trajectories, particularly “All Alternation gain first”, underscores the dynamic and spatially diffuse nature of fire occurrence. This reinforces the importance of interpreting wildfire not merely as an annual hazard but as a temporally embedded environmental process shaped by land-use legacies, fuel dynamics, and climatic variability. Adapting the method from Bilintoh et al. [29] to wildfire data proved not only feasible, but methodologically generative. It enabled new ways of identifying where and when fire transitions occur, knowledge that is critical for the design of more anticipatory and differentiated management strategies. Rather than reacting to individual fire events, policymakers and land managers can use trajectory classifications to recognize areas of persistent vulnerability, shifting risk, and potential resilience. This research contributes to advancing methods for long-term analysis of wildfire regimes, demonstrating that trajectory analysis can offer valuable insights into complex disturbance dynamics, especially when applied at high spatial and temporal resolution. Its significance lies not only in the novel application to wildfire data, but also in supporting more adaptive, spatially targeted, and historically informed fire governance.

Author Contributions

Conceptualization, B.B., A.G., S.O. and C.M.V.; methodology, B.B., A.G., S.O. and C.M.V.; software, B.B. and S.O.; validation, B.B. and S.O.; formal analysis, B.B., A.G., S.O. and C.M.V.; investigation, B.B., A.G. and S.O.; data curation, B.B.; writing—original draft preparation, B.B., A.G., S.O. and C.M.V.; writing—review and editing, S.O. and C.M.V.; visualization, B.B. and S.O.; supervision, S:O.; funding acquisition, C.M.V. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by Foundation for Science and Technology (FCT) and CEG Research Unit, grant number UIDB/00295/2020 and UIDP/00295/2020; B. B. was supported by the Ph.D. fellowship funded by FCT [2022.12095.BD]; A.G. was supported by the Ph.D. fellowship funded by FCT [2020.07651.BD]; S. O. was funded by FCT through a CEEC contract [2020.03873.CEECIND]; and C.M.V. was funded by FCT through a CEEC contract [2022.08734.CEECIND].

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors on request.

Acknowledgments

We acknowledge the GEOMODLAB (Laboratory for Remote Sensing, Geographical Analysis and Modelling) of the Centre of Geographical Studies/Institute of Geography and Spatial Planning for providing computational infrastructure support for this research. The authors used OpenAI’s ChatGPT (GPT-4) to assist with language editing, reference formatting, and consistency checking during the preparation of the manuscript. All content was reviewed and verified by the authors, who take full responsibility for the scientific accuracy, originality, and integrity of the submitted work.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Dupuy, J.L.; Fargeon, H.; Martin-StPaul, N.; Pimont, F.; Ruffault, J.; Guijarro, M.; Hernando, C.; Madrigal, J.; Fernandes, P. Climate Change Impact on Future Wildfire Danger and Activity in Southern Europe: A Review. Ann. For. Sci. 2020, 77, 35. [Google Scholar] [CrossRef]
  2. Moreira, F.; Ascoli, D.; Safford, H.; Adams, M.A.; Moreno, J.M.; Pereira, J.M.C.; Catry, F.X.; Armesto, J.; Bond, W.; González, M.E.; et al. Wildfire Management in Mediterranean-Type Regions: Paradigm Change Needed. Environ. Res. Lett. 2020, 15, 011001. [Google Scholar] [CrossRef]
  3. Pausas, J.G.; Fernández-Muñoz, S. Fire Regime Changes in the Western Mediterranean Basin: From Fuel-Limited to Drought-Driven Fire Regime. Clim. Change 2012, 110, 215–226. [Google Scholar] [CrossRef]
  4. Aragoneses, E.; Chuvieco, E. Generation and Mapping of Fuel Types for Fire Risk Assessment. Fire 2021, 4, 59. [Google Scholar] [CrossRef]
  5. Bowman, D.M.J.S.; Kolden, C.A.; Abatzoglou, J.T.; Johnston, F.H.; van der Werf, G.R.; Flannigan, M. Vegetation Fires in the Anthropocene. Nat. Rev. Earth Environ. 2020, 1, 500–515. [Google Scholar] [CrossRef]
  6. Jones, M.W.; Abatzoglou, J.T.; Veraverbeke, S.; Andela, N.; Lasslop, G.; Forkel, M.; Smith, A.J.P.; Burton, C.; Betts, R.A.; van der Werf, G.R.; et al. Global and Regional Trends and Drivers of Fire Under Climate Change. Rev. Geophys. 2022, 60, e2020RG000726. [Google Scholar] [CrossRef]
  7. McGee, T.K. Public Engagement in Neighbourhood Level Wildfire Mitigation and Preparedness: Case Studies from Canada, the US and Australia. J. Environ. Manag. 2011, 92, 2524–2532. [Google Scholar] [CrossRef]
  8. Chergui, B.; Fahd, S.; Santos, X.; Pausas, J.G. Socioeconomic Factors Drive Fire-Regime Variability in the Mediterranean Basin. Ecosystems 2018, 21, 619–628. [Google Scholar] [CrossRef]
  9. Oliveira, S.; Gonçalves, A.; Zêzere, J.L. Reassessing Wildfire Susceptibility and Hazard for Mainland Portugal. Sci. Total Environ. 2021, 762, 143121. [Google Scholar] [CrossRef]
  10. Tedim, F.; Leone, V.; Amraoui, M.; Bouillon, C.; Coughlan, M.; Delogu, G.; Fernandes, P.; Ferreira, C.; McCaffrey, S.; McGee, T.; et al. Defining Extreme Wildfire Events: Difficulties, Challenges, and Impacts. Fire 2018, 1, 9. [Google Scholar] [CrossRef]
  11. Tonini, M.; Parente, J.; Pereira, M. Global Assessment of Land Cover Changes and Rural-Urban Interface in Portugal. Nat. Hazards Earth Syst. Sci. 2018, 18, 1647–1664. [Google Scholar] [CrossRef]
  12. United Nations Environment Programme. Spreading like Wildfire—The Rising Threat of Extraordinary Landscape Fires; A UNEP Rapid Response Assessment: Nairobi, Kenya, 2022. [Google Scholar]
  13. Oliveira, S.; Gonçalves, A.; Benali, A.; Sá, A.; Zêzere, J.L.; Pereira, J.M. Assessing Risk and Prioritizing Safety Interventions in Human Settlements Affected by Large Wildfires. Forests 2020, 11, 859. [Google Scholar] [CrossRef]
  14. San-Miguel-Ayanz, J.; Durrant, T.; Boca, R.; Maianti, P.; Liberta’, G.; Felix, J.; Oom, D.; Branco, A.; De Rigo, D.; Suarez-Moreno, M.; et al. Forest Fires in Europe, Middle East and North Africa 2022; Publications Office of the European Union: Luxembourg, 2023. [Google Scholar]
  15. Ruffault, J.; Curt, T.; Moron, V.; Trigo, R.M.; Mouillot, F.; Koutsias, N.; Pimont, F.; Martin-StPaul, N.; Barbero, R.; Dupuy, J.L.; et al. Increased Likelihood of Heat-Induced Large Wildfires in the Mediterranean Basin. Sci. Rep. 2020, 10, 13790. [Google Scholar] [CrossRef]
  16. Beighley, M.; Hyde, A.C. Portugal Wildfire Management in a New Era: Assessing Fire Risks, Resources and Reforms; Centro de Estudos Florestais–Instituto Superior de Agronomia, Universidade de Lisboa: Lisboa, Portugal, 2018. [Google Scholar]
  17. Peris-Llopis, M.; Vastaranta, M.; Saarinen, N.; González-Olabarria, J.R.; García-Gonzalo, J.; Mola-Yudego, B. Post-Fire Vegetation Dynamics and Location as Main Drivers of Fire Recurrence in Mediterranean Forests. For. Ecol. Manag. 2024, 568, 122126. [Google Scholar] [CrossRef]
  18. Turco, M.; Rosa-Cánovas, J.J.; Bedia, J.; Jerez, S.; Montávez, J.P.; Llasat, M.C.; Provenzale, A. Exacerbated Fires in Mediterranean Europe Due to Anthropogenic Warming Projected with Non-Stationary Climate-Fire Models. Nat. Commun. 2018, 9, 3821. [Google Scholar] [CrossRef]
  19. Lozano, O.M.; Salis, M.; Ager, A.A.; Arca, B.; Alcasena, F.J.; Monteiro, A.T.; Finney, M.A.; Del Giudice, L.; Scoccimarro, E.; Spano, D. Assessing Climate Change Impacts on Wildfire Exposure in Mediterranean Areas. Risk Anal. 2017, 37, 1898–1916. [Google Scholar] [CrossRef] [PubMed]
  20. Benali, A.; Aparício, B.A.; Gonçalves, A.; Oliveira, S. Defining Priorities for Wildfire Mitigation Actions at the Local Scale: Insights from a Novel Risk Analysis Method Applied in Portugal. Front. For. Glob. Change 2023, 6, 1270210. [Google Scholar] [CrossRef]
  21. Gonçalves, A.; Oliveira, S.; Zêzere, J.L. Assessing the Implementation of Wildfire Mitigation Initiatives for the Protection of Villages in Portugal. Trees For. People 2025, 21, 100935. [Google Scholar] [CrossRef]
  22. De Oliveira, E.; Colaço, C.M.; Fernandes, P.M.; Sequeira, A.C. Remains of Traditional Fire Use in Portugal: A acho Historical Analysis. Trees For. People 2023, 14, 100458. [Google Scholar] [CrossRef]
  23. Moreira, F.; Leal, M.; Bergonse, R.; Canadas, M.J.; Novais, A.; Oliveira, S.; Ribeiro, P.F.; Zêzere, J.L.; Santos, J.L. Recent Trends in Fire Regimes and Associated Territorial Features in a Fire-Prone Mediterranean Region. Fire 2023, 6, 60. [Google Scholar] [CrossRef]
  24. Catry, F.X.; Rego, F.C.; Bação, F.L.; Moreira, F. Modeling and Mapping Wildfire Ignition Risk in Portugal. Int. J. Wildland Fire 2009, 18, 921–931. [Google Scholar] [CrossRef]
  25. Kanevski, M.; Pereira, M.G. Local Fractality: The Case of Forest Fires in Portugal. Physica A 2017, 479, 400–410. [Google Scholar] [CrossRef]
  26. Nunes, A.N.; Lourenço, L.; Castro Meira, A.C. Exploring Spatial Patterns and Drivers of Forest Fires in Portugal (1980–2014). Sci. Total Environ. 2016, 573, 1190–1202. [Google Scholar] [CrossRef]
  27. Bergonse, R.; Oliveira, S.; Zêzere, J.L.; Moreira, F.; Ribeiro, P.F.; Leal, M.; Lima e Santos, J.M. Biophysical Controls over Fire Regime Properties in Central Portugal. Sci. Total Environ. 2022, 810, 152314. [Google Scholar] [CrossRef] [PubMed]
  28. Silva, P.; Carmo, M.; Rio, J.; Novo, I. Changes in the Seasonality of Fire Activity and Fire Weather in Portugal: Is the Wildfire Season Really Longer? Meteorology 2023, 2, 74–86. [Google Scholar] [CrossRef]
  29. Bilintoh, T.M.; Pontius, R.G.; Zhang, A. Methods to Compare Sites Concerning a Category’s Change during Various Time Intervals. GIScience Remote Sens. 2024, 61, 2409484. [Google Scholar] [CrossRef]
  30. DGT (Direção-Geral do Território). Carta de Uso e Ocupação do Solo de Portugal Continental Para 2018 (COS2018); DGT: Lisboa, Portugal, 2018. Available online: https://snig.dgterritorio.gov.pt (accessed on 5 May 2025).
  31. San-Miguel-Ayanz, J.; Durrant, T.; Boca, R.; Maianti, P.; Liberta’, G.; Felix, J.; Oom, D.; Branco, A.; De Rigo, D.; Suarez-Moreno, M.; et al. Forest Fires in Europe, Middle East and North Africa 2023; Publications Office of the European Union: Luxembourg, 2024. [Google Scholar]
  32. ICNF Instituto. De Conservação da Natureza. e Florestas. Territórios Ardidos. 2025. Available online: https://geocatalogo.icnf.pt/metadados/area_ardida.html (accessed on 5 May 2025).
  33. Pontius, R.G.; Millones, M. Death to Kappa: Birth of Quantity Disagreement and Allocation Disagreement for Accuracy Assessment. Int. J. Remote Sens. 2011, 32, 4407–4429. [Google Scholar] [CrossRef]
  34. Bilintoh, T.; Pontius, J.R. TimeseriesTrajectories: Analyzes The Trajectories of a Variable During a Time Series, R Package Version 1.0.3. 2025. Available online: https://github.com/bilintoh/timeseriesTrajectories (accessed on 5 May 2025).
  35. Pereira, M.G.; Gonçalves, N.; Amraoui, M. The Influence of Wildfire Climate on Wildfire Incidence: The Case of Portugal. Fire 2024, 7, 234. [Google Scholar] [CrossRef]
  36. Gómez-González, S.; Ojeda, F.; Fernandes, P.M. Portugal and Chile: Longing for Sustainable Forestry While Rising from the Ashes. Environ. Sci. Policy 2018, 81, 104–107. [Google Scholar] [CrossRef]
  37. Moritz, M.A.; Batllori, E.; Bradstock, R.A.; Gill, A.M.; Handmer, J.; Hessburg, P.F.; Leonard, J.; McCaffrey, S.; Odion, D.C.; Schoennagel, T.; et al. Learning to Coexist with Wildfire. Nature 2014, 515, 58–66. [Google Scholar] [CrossRef]
  38. Pausas, J.G.; Keeley, J.E. A Burning Story: The Role of Fire in the History of Life. Bioscience 2009, 59, 593–601. [Google Scholar] [CrossRef]
  39. Bergonse, R.; Oliveira, S.; Zêzere, J.L.; Moreira, F.; Ribeiro, P.F.; Leal, M.; Santos, J.M.L. Differentiating Fire Regimes and Their Biophysical Drivers in Central Portugal. Fire 2023, 6, 112. [Google Scholar] [CrossRef]
  40. Calheiros, T.; Benali, A.; Pereira, M.; Silva, J.; Nunes, J. Drivers of Extreme Burnt Area in Portugal: Fire Weather and Vegetation. Nat. Hazards Earth Syst. Sci. 2022, 22, 4019–4037. [Google Scholar] [CrossRef]
  41. Benali, A.; Guiomar, N.; Gonçalves, H.; Mota, B.; Silva, F.; Fernandes, P.M.; Mota, C.; Penha, A.; Santos, J.; Pereira, J.M.C.; et al. The Portuguese Large Wildfire Spread Database (PT-FireSprd). Earth Syst. Sci. Data 2023, 15, 3791–3818. [Google Scholar] [CrossRef]
  42. Catry, F.X.; Rego, F.C.; Silva, J.S.; Moreira, F.; Camia, A.; Ricotta, C.; Conedera, M. Fire Starts and Human Activities. In Towards Integrated Fire Management Outcomes of the European Project Fire Paradox; European Forest Institute: Joensuu, Finland, 2010; pp. 9–21. [Google Scholar]
  43. Moreira, F.; Vaz, P.; Catry, F.; Silva, J.S. Regional Variations in Wildfire Susceptibility of Land-Cover Types in Portugal: Implications for Landscape Management to Minimize Fire Hazard. Int. J. Wildland Fire 2009, 18, 563–574. [Google Scholar] [CrossRef]
  44. Oliveira, S.; Moreira, F.; Boca, R.; San-Miguel-Ayanz, J.; Pereira, J.M.C. Assessment of Fire Selectivity in Relation to Land Cover and Topography: A Comparison between Southern European Countries. Int. J. Wildland Fire 2014, 23, 620–630. [Google Scholar] [CrossRef]
  45. Calheiros, T.; Nunes, J.P.; Pereira, M.G. Recent Evolution of Spatial and Temporal Patterns of Burnt Areas and Fire Weather Risk in the Iberian Peninsula. Agric. For. Meteorol. 2020, 287, 107923. [Google Scholar] [CrossRef]
  46. Pais, S.; Aquilué, N.; Honrado, J.P.; Fernandes, P.M.; Regos, A. Optimizing Wildfire Prevention through the Integration of Prescribed Burning into ‘Fire-Smart’ Land-Use Policies. Fire 2023, 6, 457. [Google Scholar] [CrossRef]
  47. Fernandes, P.M. Empirical Support for the Use of Prescribed Burning as a Fuel Treatment. Curr. For. Rep. 2015, 1, 118–127. [Google Scholar] [CrossRef]
  48. Tedim, F.; Leone, V.; Xanthopoulos, G. A Wildfire Risk Management Concept Based on a Social-Ecological Approach in the European Union: Fire Smart Territory. Int. J. Disaster Risk Reduct. 2016, 18, 138–153. [Google Scholar] [CrossRef]
  49. Oliveira, S.; Zêzere, J.L. Assessing the Biophysical and Social Drivers of Burned Area Distribution at the Local Scale. J. Environ. Manag. 2020, 264, 110449. [Google Scholar] [CrossRef] [PubMed]
  50. Duane, A.; Aquilué, N.; Gil-Tena, A.; Brotons, L. Integrating Fire Spread Patterns in Fire Modelling at Landscape Scale. Environ. Model. Softw. 2016, 86, 219–231. [Google Scholar] [CrossRef]
Land 14 01872 g001
Figure 1. (a) Study area—mainland Portugal and regional administrative division. (b) Number of times burned (100 m) between 1975 and 2024.
Figure 1. (a) Study area—mainland Portugal and regional administrative division. (b) Number of times burned (100 m) between 1975 and 2024.
Land 14 01872 g001
Land 14 01872 g004
Figure 4. Annual records of number of fires and burned area, between 1975 and 2024. The reference years used to define the analysis timeframes are highlighted in red (decadal intervals) and green (25-year intervals).
Figure 4. Annual records of number of fires and burned area, between 1975 and 2024. The reference years used to define the analysis timeframes are highlighted in red (decadal intervals) and green (25-year intervals).
Land 14 01872 g004
Land 14 01872 g005
Figure 5. Number of times burned (fire frequency) per decade, between 1975 and 2024.
Figure 5. Number of times burned (fire frequency) per decade, between 1975 and 2024.
Land 14 01872 g005
Land 14 01872 g006
Figure 6. Number of times burned (fire frequency) per 25-years intervals: (a) 1975–2000 and (b) 2000–2024.
Figure 6. Number of times burned (fire frequency) per 25-years intervals: (a) 1975–2000 and (b) 2000–2024.
Land 14 01872 g006
Land 14 01872 g007
Figure 7. Fire trajectories trajectory classification per 25-year interval: (a) 1975–2000, and (b) 2000–2024.
Figure 7. Fire trajectories trajectory classification per 25-year interval: (a) 1975–2000, and (b) 2000–2024.
Land 14 01872 g007
Land 14 01872 g008
Figure 8. Results obtained for Norte region: Number of times burned (left); Fire trajectory classification (right) per decade between 1975 and 2024.
Figure 8. Results obtained for Norte region: Number of times burned (left); Fire trajectory classification (right) per decade between 1975 and 2024.
Land 14 01872 g008
Land 14 01872 g009
Figure 9. Results obtained for Algarve region: Number of times burned (left); Fire trajectory classification (right) per decade between 1975 and 2024.
Figure 9. Results obtained for Algarve region: Number of times burned (left); Fire trajectory classification (right) per decade between 1975 and 2024.
Land 14 01872 g009
Table 1. Percentage of the area in Portugal’s mainland covered by each trajectory across the five decadal intervals (red markers) and two 25-years periods (green markers). The three highest values per timeframe are presented in bold. The Trajectory colors are the same as the legend of fire trajectory showed in Figure 3.
Table 1. Percentage of the area in Portugal’s mainland covered by each trajectory across the five decadal intervals (red markers) and two 25-years periods (green markers). The three highest values per timeframe are presented in bold. The Trajectory colors are the same as the legend of fire trajectory showed in Figure 3.
RegionTrajectory1975–1985 (%)1985–1995 (%)1995–2005 (%)2005–2015 (%)2015–2024 (%)1975–2000 (%)2000–2024 (%)
Portugal10.542.040.772.950.440.240.54
20.231.030.440.810.120.680.97
32.390.912.620.340.970.530.51
40.660.390.970.160.521.000.95
50.230.210.300.130.070.080.10
64.617.7511.266.0310.4315.0121.24
891.3487.6783.6489.5987.4582.4575.69
Table 2. Percentage of land area covered by each fire trajectory in the Norte and Algarve regions, across five decades (red intervals) and two 25-years’ timeframes (green intervals). The three most frequent trajectories per region and timeframe are highlighted in bold. The Trajectory colors are the same as the legend of fire trajectory shown in Figure 8 and Figure 9 (left).
Table 2. Percentage of land area covered by each fire trajectory in the Norte and Algarve regions, across five decades (red intervals) and two 25-years’ timeframes (green intervals). The three most frequent trajectories per region and timeframe are highlighted in bold. The Trajectory colors are the same as the legend of fire trajectory shown in Figure 8 and Figure 9 (left).
RegionTrajectory1975–1985 (%)1985–1995 (%)1995–2005 (%)2005–2015 (%)2015–2024 (%)1975–2000 (%)2000–2024 (%)
Norte11.103.320.784.390.960.460.96
20.612.030.762.290.241.491.95
34.001.013.840.551.760.910.75
41.260.732.740.371.262.102.14
50.410.320.510.420.140.170.27
66.7511.1813.9613.2411.6222.6826.20
885.8781.4177.4078.7584.0272.1967.73
Algarve10.040.500.690.300.050.010.01
20.040.011.470.000.000.070.05
30.512.100.250.050.010.050.01
40.010.060.05--0.010.00
5-0.000.00----
63.443.5517.945.619.999.8327.93
895.9693.7879.6094.0489.9690.0272.00
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Barbosa, B.; Gonçalves, A.; Oliveira, S.; Viana, C.M. Mapping Long-Term Wildfire Dynamics in Portugal Using Trajectory Analysis (1975–2024). Land 2025, 14, 1872. https://doi.org/10.3390/land14091872

AMA Style

Barbosa B, Gonçalves A, Oliveira S, Viana CM. Mapping Long-Term Wildfire Dynamics in Portugal Using Trajectory Analysis (1975–2024). Land. 2025; 14(9):1872. https://doi.org/10.3390/land14091872

Chicago/Turabian Style

Barbosa, Bruno, Ana Gonçalves, Sandra Oliveira, and Cláudia M. Viana. 2025. "Mapping Long-Term Wildfire Dynamics in Portugal Using Trajectory Analysis (1975–2024)" Land 14, no. 9: 1872. https://doi.org/10.3390/land14091872

APA Style

Barbosa, B., Gonçalves, A., Oliveira, S., & Viana, C. M. (2025). Mapping Long-Term Wildfire Dynamics in Portugal Using Trajectory Analysis (1975–2024). Land, 14(9), 1872. https://doi.org/10.3390/land14091872

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop