Mediterranean fire regime effects on pine-oak forest landscape mosaics under global change in NE Spain


Afforestation after land abandonment and the occurrence of large fires have significantly altered the composition of pine-oak ecosystems in the Mediterranean since 1950s, the latter favouring the prevalence of oak forests and shrublands to that of pine forests. Nevertheless, our ability to integrate the processes driving these changes in modelling tools and to project them under future global change scenarios is scarce. This study aims at investigating how Mediterranean forest landscape composition and seral stages may be affected by mid-term changes in fire regime and climate. Taking Catalonia (NE Spain) as study area, we predicted yearly changes in forest landscape composition using the MEDFIRE model which allows assessing the effects of different fire regimes on landscape dynamics such as post-fire regeneration and afforestation. We considered three climatic treatments based on observed and projected climate, two fire regimes largely differing in the amount of area burnt and the number of large fires, and two fire suppression strategies. While projected afforestation continued to increase forest cover in the 2050 horizon, a climate-related harsher fire regime (higher amounts of area burnt) accelerated a shift towards landscapes progressively dominated by oaks and shrublands, thus precluding general forest maturation. Fire-sensitive pine species contributed to net forest cover loss in the worst scenarios. An active fire suppression strategy partially compensated the effects of a climate-related harsher fire regime on pine forest loss and rejuvenation, whereas variability in climate projections weakly affected spatial fire allocation and afforestation. Our results highlight the need to explicitly incorporate fire suppression strategies in forest landscape composition forecasts in the Mediterranean. At mid-term, large-scale afforestation, post-fire forest rejuvenation and landscape composition changes may alter forest ecosystem functioning and potentially interact with fire suppression planning.


The Mediterranean basin has been under a long-lasting human activity, which transformed forest ecosystems for centuries, reducing their distribution and altering their species composition and age (Blondel and Aronson 1999). Mediterranean pine-oak ecosystems have evolved in the Mediterranean basin as either pure forests or mixed pine-oak associations and represent a significant proportion of forested areas (Sheffer 2012). The current distribution of pine-oak ecosystems and their species composition are a consequence of the interplay between human activity, natural disturbances, and the main limiting factors acting at different spatial scales (e.g., water availability limits oaks at the regional scale due to the species drought sensitivity, whereas shade tolerance mostly determines the forest composition at the stand level) (Zavala et al. 2000). Nevertheless, the distribution of pine-oak ecosystems has been also changing since the second half of the twentieth century due to broad-scale abandonment of traditional human activities, which fostered forest recovery and densification (Debussche et al. 1999; Améztegui et al. 2010). Many Mediterranean pine species are pioneer species, and this characteristic allowed them colonizing former agricultural lands after rural abandonment and subsequent shrub encroachment (Sheffer 2012). In addition, afforestation programs in former cultivated lands have also increased considerably pine distribution in the twentieth century due to their better adaptation to disturbed environments (Sheffer 2012). In contrast, Quercus spp. are mostly animal-dispersed and shade tolerant species and may establish well below Pinus spp. overstory and remain there for decades until disturbances create large-scale gaps (e.g., large fires) that allow Quercus spp. better regenerate in contrast to some Pinus spp. (Retana et al. 2002; Rodrigo et al. 2004; Puerta-Piñero et al. 2012).

Global change processes leading to large-scale afforestation and densification have also impacted on Mediterranean fire regime (Lloret et al. 2002; Vega-García and Chuvieco 2006) together with increased temperatures and drought seasons (Piñol et al. 1998; Pausas and Fernández-Muñoz 2011). During the twentieth century, fire regime shifted from a fuel-limited to a drought-driven regime and forest area burnt by large fires increased dramatically in the Mediterranean Europe (Pausas and Fernández-Muñoz 2011). Although the firefighting system has been shown to have a strong influence on fire regimes in Mediterranean Europe (Brotons et al. 2013; Moreno et al. 2014), increasingly frequent large fires can overwhelm firefighting system capacity due to their unpredictable character and high fire intensity (i.e. convective wildfires; Duane et al. 2015). Fire regime changes have therefore induced strong influences on pine-oak forest landscape composition in the last decades of the twentieth century with detrimental effects on pine forest dominance (Retana et al. 2002; Rodrigo et al. 2004; Puerta-Piñero et al. 2012). Pines and oaks in the Mediterranean present different strategies to cope with fire impacts, whereas Quercus species resprout after fire, pines are obligate seeders and only some of them present different degrees of serotiny. Fire sensitive pines without serotiny may be considered as losers under global change aggravating fire regimes since distances between burned and unburned pine areas in large fires are often too large (Retana et al. 2002) compared to the species maximum seed dispersal distance. Large-scale availability of pine propagules due to specific afforestation programs conducted in the twentieth century, and secondary succession may also have the potential to offset net pine surface loss (Pausas 2006; Sheffer 2012). As a consequence, significant impacts on ecosystem services may be expected due to forest landscape composition changes under harsher fire regimes as previously projected in other fire-prone forest ecosystems (Rocca et al. 2014). In addition, the shift from forest to alternative cover types (e.g., shrublands) because of post-fire regeneration failure may imply persistent consequences on ecosystem services provided by forests (e.g., significant carbon stock reduction) and forest biodiversity habitat loss (Stephens et al. 2013).

While we now have a good knowledge of the processes driving these changes, our ability to integrate this information in modelling tools that allow the projection of these systems under future global change scenarios is scarcer. Future large-scale dynamics of the pine-oak system in Mediterranean forest landscapes and the subsequent consequences on ecosystem functioning will depend on the interplay of several factors directly affecting the amount of burnt area such as fire regime adversity and firefighting activity but also because of spatio-temporal changes in ignition patterns due to fuel availability and climate variability. A higher fire probability has been predicted for Mediterranean biomes at mid- and long-term in the twenty-first century once accounting for climate projections uncertainties (Moritz et al. 2012), whereas fire risk may rise substantially in the last 30 years of the twenty-first century in Mediterranean Europe according to IPCC-SRES scenarios (Moriondo et al. 2006). Potential future large-scale forest loss has been modelled in the Mediterranean as a function of fire regime scenarios (Pausas 2006; De Cáceres et al. 2013), and composition changes have been projected in relation to predicted climate warming in detriment of the most mesic species (Benito Garzón et al. 2009; Ruiz-Labourdette et al. 2012 but see Keenan et al. 2011). However, we lack a comprehensive and quantitative understanding on how further changes in climate, fire regimes and afforestation rates may affect the future relative dominance of main tree species in Mediterranean forests (Zavala et al. 2000; Loepfe et al. 2012). In this work, we aimed at projecting how Mediterranean pine-oak forest landscape mosaics may be affected in terms of dominant species composition and seral stage evolution at mid-term (2050) by different fire regimes, while considering the impacts of firefighting capacity to offset fire regime severity. We used the MEDFIRE spatially explicit landscape fire-succession model that allows examining the spatial interactions between wildfires, landscape vegetation dynamics, and fire suppression strategies in Mediterranean ecosystems under different climatic scenarios in the region of Catalonia (NE Spain) (Brotons et al. 2013).

Materials and methods

Study area

Catalonia (NE Spain; 32,107 km2; Fig. 1) is a heterogeneous topographic region comprising mountainous areas like the Pyrenees (up to 3143 m), extensive interior and predominantly agricultural plains, and a long coastline along the Mediterranean Sea. The climate is mainly Mediterranean temperate, with maritime influence on the coast and colder temperatures in the Pyrenees.

Fig. 1

Location of the study area (a) and land cover map with pine-oak forests’ detail (b). A schematic view of MEDFIRE is also provided (c)

Forests represent about 38 % of the total area of Catalonia and have a wide diversity of forest types (Piqué et al. 2011b) and tree species (about a hundred). However, 90 % of the tree species are from the 14 most common species, which are Pinus halepensis (20 % of the total forest area), Pinus sylvestris (18 %), Quercus ilex (15 %), Pinus nigra (11 %), Quercus humilis, Quercus suber, Pinus uncinata, and Pinus pinea, among others. Mixed forests represent a significant proportion of the Catalan forest surface (ca. 40 %), and conifers represent 60 % of forest area either as dominant or mixed stands (Piqué et al. 2011b). As a result of the past harvesting treatments and the fire occurrences that traditionally affected Catalan forests, the average stand age of most forest typologies was under 50 years in 2000.

Despite the fact that fires had burned approximately 300,000 ha between 1980 and 2010 in Catalonia (datum gathered from the Fire Prevention Service), widespread afforestation have counteracted forest cover loss in the region (see Table 2.1 in the Online Resource 2), whereas P. nigra, P. sylvestris and P. pinea do not have life traits conferring resistance to high intensity crown fires (i.e. serotiny), direct post-fire regeneration of serotinous species such as P. halepensis has been shown to be compromised with increasing fire recurrence and shorter periods than the age of cone production (Lloret et al. 2003; Rodrigo et al. 2004). Therefore, in many areas affected by large fires, Quercus spp. dominate post-fire regeneration (see Fig. 2.1 in Online Resource 2).

The MEDFIRE model

MEDFIRE is a spatially explicit landscape dynamics model at 100-m resolution that allows to examine the spatial interactions between wildfires, landscape vegetation dynamics, climate change, and fire suppression strategies in Mediterranean ecosystems, and it is parameterized for Catalonia (Brotons et al. 2013). MEDFIRE allows predicting yearly changes of forest landscape composition as a function of fire regime, post-fire regeneration, and afforestation. A short description of the MEDFIRE model is provided below, particularly focusing on the model changes with respect to Brotons et al. (2013) conducted to specifically assess mid-term changes in forest landscape composition, seral stage evolution and to improve ignition probability modelling as a function of climate. More details and mathematical formulation of the modelling framework can be obtained from the MEDFIRE model seminal work (see Brotons et al. 2013) and the Online Resource 1.

Two state variables characterize the landscape and the vegetation in the model. Land Cover Type (LCT) is a categorical variable describing the dominant tree species in forested areas, other vegetated areas and main covers. Forest composition at the landscape scale was based on the species spatial distribution according to the tessellation of the Forest Map of Spain (up to three main species per forest patch were reported, with a minimum mapping unit of 2.25 ha) and forest stand data gathered from the Third National Forest Inventory (NFI). Mixed forests at the landscape scale were thus considered as spatial heterogeneous combinations of dominant forest tree species at 1 ha randomly spread within the forest patch depending on the species proportions per forest patch that were based on basal area interpolations (NFI data). Four main pine species (P. halepensis, P. nigra, P. pinea, P. sylvestris) and two oak species (Q. suber, Q. ilex) in the region are separately considered, while other oaks (e.g., Q. faginea and Q. humilis) are aggregated apart from other secondary conifers and deciduous species. Forest, shrubland, alpine grass, and agricultural lands are burnable covers, but only forest and shrubland are dynamic (i.e., the static cover composition remains unaltered over the time despite being burnable). Fuel Age is an integer variable that indicates the years since the last fire for recently burnt forests and shrublands and the age of unburnt forests obtained from the Third NFI (species top height) through the use of reference site index curves in Catalonia and interpolation techniques such as ordinary kriging.

The MEDFIRE is structured in the Fire and the Vegetation Dynamics sub-models (Fig. 1). The former includes a top-down fire regime that depends on annual climate severity since a higher number of fire weather risk days occur in adverse years and, therefore, larger amounts of area burnt are expected than in normal years due to the greater occurrence of large fires. Water deficit balance records and fire statistics in the period 1980–2000 were used to calibrate the fire regime. To classify climatically adverse and normal years, we used the cumulative soil water deficit (CSWD) as surrogate of an aridity index, which reflects water deficit in summer and has been shown to correlate well with large amounts of annual burnt area in NE Spain (Pausas and Paula 2012). Then, fire statistics were used to estimate both the annual burnt area and fire size distributions for normal and climatically adverse years (see Online Resource 1). For this study, an updated probability of fire ignition was calibrated as function of climate, neighbouring land covers including interfaces among them, and human infrastructures, from ignition points of forest fires larger than 50 ha that occurred in the period 1987–2011 (Table 1). This fire ignition probability layer affects the spatial distribution of ignitions under a dynamic landscape-climate framework. The MEDFIRE model replicates different fire suppression strategies from moderate to strong active fire suppression which mimic firefighting capacity under increasingly fire spread conditions (0–100 fire spread rate).

Table 1 Fitted models for probability of ignition (Formula 1) and probability of afforestation (Formula 2). Annual probability of afforestation was rescaled from Formula 2. See more details in Online Resource 1

The Vegetation Dynamics sub-model includes two ecological processes at the landscape scale: the post-fire regeneration and the afforestation from shrubland to forest. Forest stand dynamics and long-term forest succession are not handled due to the model scope and the temporal simulation horizon, and transitions between forest types only occur after fire. Post-fire direct regeneration depends on the species post-fire regeneration response [transition probabilities based on Rodrigo et al. (2004)], while spatially autocorrelated post-fire transition occurs at 40 % rate. Unlike Brotons et al. (2013) that considered different transition matrices according to bioclimatic regions, in this study we used only one transition matrix for the entire study area (Online Resource 1) based on Rodrigo et al. (2004) since post-fire regeneration in MEDFIRE is now constrained by the presence before the fire of the tree species within 1 km radius. Annual probability of afforestation was modelled as a function of shrubland age according to fuel age, amount of neighbouring forest in reproductive age, climate variables, and topography (Table 1).

Scenario definition

The effects of wildfires, climate, and fire suppression on forest landscape composition and seral stages were investigated in a set of scenarios (Table 2): three climate treatments which directly influence the probability of afforestation and the ignition process and which were defined according to observed climate data and two future climate projections; two fire regimes were proposed, widely differing in the percentage of climatically adverse years within the simulated period; and finally, an active fire suppression strategy was contrasted with no firefighting activity.

Table 2 Definition of the scenarios as a combination of the 3 × 2 × 2 treatments based on climate change, fire regime and firefighting strategies

In the climate treatments, the modelled processes depending on climate data (afforestation and ignition probability) were updated each decade from 2011 to 2050 according to averaged climate projections of the A2 and B2 IPCC-SRES scenarios following the general circulation model CGCM2 (Table 3). In the reference climate scenario (C0), future climate was not updated and was set constant according to the averaged 2000–2010 climate. The two IPCC storylines describe regionally oriented economic development with continuously increasing global population at different rates (smaller in the B2 scenarios) and with different degrees of sustainability and environmental protection (greater in the B2 scenario) (IPCC 2000). Climate data were provided at the spatial resolution of 1 km from climatic records and projections elaborated by the Spanish Meteorological Agency (AEMET).

Table 3 Observed and projected decadal average evolution of the climate in the study area

The fire regime scenarios were defined by the percentage of climatically adverse years which were those years with a CSWD regional average greater than 270 mm (see the Online Resource 1). The reference fire regime scenario (FR0) was based on the observed percentage of climatically adverse years in the 1980–2000 period (40 %). Although we also computed the percentage of climatically adverse years during the period 2010–2050 according to the climate projections for the A2 and B2 IPCC-SRES scenarios, we did not obtain a significant difference from the reference (see Online Resource 1). Nevertheless, we considered a harsher fire regime (FR1) in which we doubled the percentage of adverse years due to climate uncertainties associated with increasing occurrence of extreme weather events in Mediterranean Europe (Fischer and Schär 2010) that may therefore affect large wildfire episodes (Montserrat Aguadé 1998).

In this study we considered a moderate active fire suppression that leads to an effective fire size smaller than the potential fire size (Brotons et al. 2013). Active firefighting indicates the difficulty for fire extinction and was set at a fire spread rate of 70 as an indicator of medium active fire suppression which includes the extinction of fires burning agricultural covers, back fire fronts in sclerophyllous forests and sclerophyllous forests in flat conditions (Brotons et al. 2013).

Model simulations and output variables analysed

Model simulations were run for 50 years and 100 replicates per scenario were generated. We started running MEDFIRE from 2000 and not 2010 because of forest characterization data availability. For the first decade of the simulated period, the fire regime (number of climatically adverse years and annual burnt area) and the averaged decadal climate were the ones observed in 2001–2010. The temporal average evolution of forests (pines and oaks) and shrublands were assessed at the regional level. Forest seral stage evolution was assessed according to the time required for canopy closure after fire disturbance in Mediterranean forests which has been set at 30 years (Retana et al. 2002; Broncano et al. 2005). In addition, we also evaluated the effects of the scenarios on fire regime and, particularly, how climate treatments affected fire ignition spatial patterns.


Fire regime impacts

Firefighting considerably reduced total burnt area (Fig. 2), with ca. 56 % less burnt area in scenario FR0-C0-Act compared to scenario FR0-C0-noFF (579,746 ha burnt). However, a climate-related harsher fire regime implies a bigger target area to burn, a greater occurrence of large fires (Fig. 2.2 in Online Resource 2), and a bigger percentage of area overwhelming firefighting capacity. Thus, firefighting suppression scenarios in a harsher fire regime (FR1-Act) resulted in about 30 % more burnt area than FR0-Act. The area of burnt pine forests was the greatest in all the scenarios compared to the other burnable land cover types (Fig. 3). All species were projected to burn a lot more under harsher fire regime scenarios (FR1) and without firefighting, but the most prevalent species in the region (P. halepensis) was the species contributing the most to the amount of burnt pine forests in all the scenarios (Fig. 3).

Fig. 2

Statistical distributions for the total area burnt (k ha), obtained after 100 fifty-year simulations of the MEDFIRE model under different fire regime scenarios. See scenario definition in Table 2. Lower and upper whiskers indicate the 5 and 95 % quartiles, lower and upper hinges indicate the first and third quartile and the central black line indicates the median value. Black squares indicate the mean values

Fig. 3

Projected forest area burnt (mean ± SD, in k ha) of pine and oak species in each scenario (100 replicas). See scenario definition in Table 2

Landscape composition and forest seral stage changes

Afforestation showed a higher recruitment of pine forests over oak forests at the beginning of the simulated period in all the simulated scenarios (Fig. 4). Afforestation rate steadily decreased until 2020 and then continued to decrease for pines in all scenarios and to increase for oaks without firefighting and especially in harsher fire regimes.

Fig. 4

Averaged projected change in the amount of pine (a) and oak (b) afforestation under the 12 simulated scenarios. Thin lines are FR0 scenario, and thick lines are FR1; black lines are noFF scenarios and grey lines are Act, continuous lines are climate reference scenarios (C0), pointed lines are A2 climate scenarios and dot dashed lines are B2 climate scenarios. See scenario definition in Table 2

Changes in pine forest cover over the simulated period varied markedly between the different scenarios (Fig. 5). Only scenarios including firefighting (Act) showed continuous increases in pine forest cover. For the remaining scenarios (noFF), different pine cover trends were observed from 2010 onwards with a decreasing trend under scenarios including a harsher fire regime (FR1) or rather stability in the reference fire regime scenarios (FR0). Until nearly 2030 the cover of young forest seral stages increased and then reached a plateau up to the end of the simulated period in all the scenarios (Fig. 6). From 2010 onwards, late seral forest stage evolution was negative in the reference fire regime scenarios and no firefighting, but the trend was clearly more negative under scenarios including a climate-related harsher fire regime and no firefighting (ca. −10 and −20 % in 2050 compared to 2000, respectively; Fig. 5). Scenarios including firefighting strategies showed relative stability of late pine forest seral stage area (3 % more in FR0-C0-Act; Fig. 5). During the simulated period, P. halepensis was the species that consistently increased its surface in all the scenarios but in a lesser extent in the worst scenarios (e.g., FR1-noFF scenarios; Fig. 7). The trends for the remaining species were rather negative without firefighting and particularly with a harsher fire regime scenario.

Fig. 5

Averaged projected change in the amount of area covered of forests and shrublands (k ha) in Catalonia under the 12 simulated scenarios. Thin lines are FR0 scenario, and thick lines are FR1; black lines are noFF scenarios and grey lines are Act, continuous lines are climate reference scenarios (C0), pointed lines are A2 climate scenarios and dot dashed lines are B2 climate scenarios. See scenario definition in Table 2

Fig. 6

Averaged projected change in the amount of area (k ha) covered of young and late forest seral stage of the main plant functional types under the 12 simulated scenarios. Thin lines are FR0 scenario, and thick lines are FR1; black lines are noFF scenarios and grey lines are Act, continuous lines are climate reference scenarios (C0), pointed lines are A2 climate scenarios and dot dashed lines are B2 climate scenarios. See scenario definition in Table 2

Fig. 7

Projected proportion of change (mean ± SD) of the initial pine and oak forest area (in brackets). See scenario definition in Table 2

Oak forests tendency to expand (ca. 30 % more area at the end of the simulated period; Fig. 5) was favoured by harsher fire regimes (FR1) although it was smaller under scenarios including firefighting (Act). Increases of young oak forest seral stages were higher in the FR1 scenarios and lower in the Act scenarios, whereas the contrary was true in the case of late forest stages (Fig. 6). Consequently, all the oak species increased their forest area, but particularly other Quercus spp. (marcescents oaks such as Q. faginea and Q. humilis) in the scenario with a harsher fire regime and without firefighting (Fig. 7). Without land abandonment processes explicitly implemented in the MEDFIRE model, shrubland evolution was negative (ca. −35 %) due to afforestation and decreased more steadily when firefighting was operative (Act) and in reference fire regime scenarios (FR0) (Fig. 5).

Other climate-related impacts

Climate treatments (C0, A2 and B2) affected the probability of ignition and therefore the spatial patterns of fire occurrence (Fig. 8), which translated in a greater amount of area burnt for the most xeric species (e.g., P. halepensis, P. pinea) regarding the other species (e.g., P. nigra and P. sylvestris) in the scenarios including climate projections (Fig. 3) and a greater fire recurrence near the coast (Fig. 2.3 in Online Resource 2). On the other hand, afforestation was also impacted by future climate since the rate was lower at the end of the simulated period for the scenarios considering climate projections (Fig. 4).

Fig. 8

Maps of the frequency of burning at least one time over the simulated period under different climate scenarios with the reference fire regime (FR0) and without firefighting (noFF) and 100 replicas per scenario. See scenario definition in Table 2


Without moderate fire suppression capacity, ongoing impacts of current fire regime on forests in Catalonia due to large amounts of burnt area, frequent large fires and high fire recurrence may be expected in forthcoming decades which may therefore continue affecting pine-oak forest landscape composition and have exacerbated effects under climate-related harsher fire regimes. Fire regime impacts on pine-oak forest landscape mosaics were buffered when firefighting was considered, but harsher fire regimes resulted in more burnt area than reference scenarios because of a greater target area to be burnt beyond firefighting capacity. Under climate-related harsher fire regimes, we can neither exclude that the moderate firefighting effectiveness considered here (i.e. 70 % fire spread rate) may be overestimated because the extinction capacity has been over operational control since 1980 several times in Mediterranean Europe and especially in Catalonia (Costa et al. 2011). This may be the case of the so-called convective (or plume-driven or fuel-driven) fires, which represent a small proportion of total fires in the region but affect most of the burnt area (see Duane et al. 2015). Climate change may affect fire suppression effectiveness in the future influencing initial attack success (Fried et al. 2007) and also increase fire intensity outside effective fire suppression thresholds (Podur and Wotton 2010). Importantly, projected ignition probabilities varied regarding the observed climate when climate change projections (IPCC-SRES scenarios) were taken into account. This influenced the spatial patterns of fire occurrence and weakly increased the amount of burnt area of the most xeric species and decreased it for the most mesic species, with no major divergences among climate projections (A2 and B2 IPCC-SRES scenarios). Therefore, future climate variability may result in a spatial reallocation of fire occurrence due to spatial changes of probability of ignition which may also influence fire regime impacts on pine-oak systems. Previous simulations in the study area projected an increasing amount of burnt area and large fires frequency according to the A2 and B1 IPCC-SRES scenarios during the twenty-first century, but differences among scenarios were mainly perceptible at the end of the twenty-first century in terms of fire weather severity indices (Loepfe et al. 2012). This agrees with the lack of substantial differences observed among A2 and B2 climate treatments in our 2050 horizon simulations and the need to force a climate-related harsher fire regime in order to account for climate uncertainties associated with extreme weather events (see “Scenario definition” subsection in “Materials and methods”).

Our results reflected that recent increases in forest cover started since the 1950s will likely continue in the near future. Nevertheless, fire regime scenarios without firefighting showed a deceleration in pine forest increase and a disruption of pine and oak forest late seral forest stages, which have been more pronounced in harsher fire regime scenarios. In our simulations, the species most affected in terms of effective area loss were the pine species less adapted to large fires or fire sensitive, that is to say, the seeders without enough seedlings after disturbance (e.g., P. pinea, P. nigra and P. sylvestris; Rodrigo et al. 2004). Despite being the species most affected by fire, afforestation may continue favouring P. halepensis in line with past afforestation trends in Catalonia (Table 2.1 in the Online Resource 2). Nevertheless, the increased fire recurrence observed in climate-related harsher fire regime scenarios and under drier climate near the coast that fairly matched the species range may not be discarded to hamper post-fire direct regeneration of P. halepensis (Lloret et al. 2003; Rodrigo et al. 2004). In the case of P. nigra and P. sylvestris afforestation compensated the area loss due to wildfires in reference fire regimes including a moderate firefighting capacity, but no significant forest cover increases were observed in any case. Therefore, in the worst scenarios, the most likely trajectories of post-fire regeneration may be stronger towards resprouter regeneration (see the increasing cover of Other Quercus spp. in Fig. 7) or towards shrubland or grassland land cover change. Water availability may eventually influence post-fire recruitment such in the case of P. halepensis (Carnicer et al. 2014), particularly in mixed forests with Q. ilex (Broncano et al. 2005; see also Fig. 2.1 in Online Resource 2). In this sense, we acknowledge that our results regarding post-fire regeneration are conditioned by the transition matrix gathered from Rodrigo et al. (2004) which has been set from previous large fires occurred in the central and northern part of Catalonia and stochastic model simulations to predict the medium-term forest dynamics. Therefore, the post-fire transition probabilities may not necessarily extrapolate to all the simulated area due to climate heterogeneity in Catalonia and land use legacy divergences (Puerta-Piñeiro et al. 2012), or not be necessarily true in time because of climate change effects on vegetation dynamics.

In line with the widespread pine afforestation recently occurred in Catalonia (Table 2.1 in the Online Resource 2), pine afforestation contributed to more net gains of forest than oak recruitment at the beginning of the simulated period. During the simulation, afforestation declined since land abandonment processes are not implemented in the MEDFIRE model. Agricultural land abandonment is predicted to persist in the near future (Verburg et al. 2010) and, therefore, it could affect the results regarding the steady decrease of shrubland area in all the scenarios. Nevertheless, at the temporal scale considered (50 years), we do not estimate a high influence of unconsidered land abandonment since semi-natural regeneration is a scattered process occurring at slow rates encompassing several decades (ca. 30 years) for a recovery of shrublands and/or sparse forest cover (Bielsa et al. 2005). The projected pine-oak afforestation rate was not constant and depended on fire regime and firefighting in the last 25 years of simulation. The rising oak colonization rates at the end of the simulation period in harsher fire regimes without firefighting probably indicated post-fire forest composition changes of burnt forests of the fire sensitive pine species. The observed reduction of afforestation due to climate change (A2 and B2 climate projections) has been previously hypothesized by Loepfe et al. (2012) in Catalonia due to reduced actual evapotranspiration during the twenty-first century.

The rate of disruption of forest seral stage evolution depended on the tested plausible alternatives, principally in the case of pine forest and in the scenarios without firefighting and a harsher fire regime in which drastic reductions of late seral stages were observed compared to young seral stage trends. Previous large fire events (i.e. 1986, 1994 and 1998) weakly affected seral stage evolution after the time needed for canopy closure (30 years) (Fig. 6). Serotinous P. halepensis may be the species with comparative more burnt area and, therefore, a rejuvenation of the late seral stages may occur in the most favourable cases with successful direct regeneration. Nevertheless, ecosystem recovery is not always assured after fire and post-fire management may be envisaged in order to assist it (Moya et al. 2014). In addition, late seral stages do not necessarily imply adequate forest stand structures resistant or resilient to large fires as acknowledged in the region in recent management guidelines explicitly incorporating fire risk reduction purposes (Piqué et al. 2011a, b). For instance, in the region, the average forest age was above late seral stage age (i.e. 30 years) in 2000 and this had not prevented the occurrence of different large fire events in the 1980s and 1990s.

Firefighting in combination with landscape matrix heterogeneity management has been shown as the best management option for decreasing burnt area and large fire occurrence (Loepfe et al. 2012). Landscapes increasingly dominated by oaks may also influence future fire regime creating fire suppression opportunities that may help reduce the amount of burnt area by large fires (Regos et al. 2014) or under increasing fire risk (Moriondo et al. 2006). Pine forests and shrublands do not only burnt more because of their prevalence but because of fire selectivity to these land cover types regarding oaks (Fernandes 2009; Barros and Pereira 2014), which has been already considered in the MEDFIRE model (Brotons et al. 2013). Nevertheless, large fire size avoidance has been found for oak forests but not for pines and shrublands (Barros and Pereira 2014) which is probably related to the stand structure of oak forests that prevents fire spread (Fernandes 2009). A large-scale fuel type change and rejuvenation as here forecasted in all the simulated scenarios with increasing proportions of oak forests and by fire-prone pine cover (i.e. Pinus halepensis; Pique et al. 2011a) may be a non-negligible factor to account for in fire suppression and fire prevention planning. Although young seral stages of either pine forests or oak forests could eventually be similar to shrublands in terms of fire behaviour (Anderson 1982), at the landscape scale, differences among species or type of land cover could be indicating environmental prevalence such as moisture conditions that may still trigger different fire behaviour or fire flammability (Broncano et al. 2005). Opportunities may also come from climate warming owing to limited fuel production and subsequent fire risk reduction (Loepfe et al. 2012; Moritz et al. 2012; Rocca et al. 2014). At mid-long term, landscapes increasingly dominated by less fire-prone forest types such as oaks and with a greater shrubby vegetation prevalence may turn back fire regime more dependent on fuel availability (see also Loepfe et al. 2012) and less on climate severity, thus reversing the trends observed since the second half of the twenty-first century (Pausas and Fernández-Muñoz 2011).

Further research

In order to expand the simulated time window beyond the 2050 horizon and improve our current projections, we acknowledge that our modelling approach needs to explicitly incorporate the response of the main forest tree species in the study area to climate and/or water deficit. The consideration of climate change into the landscape vegetation dynamics will therefore refine post-fire transition probabilities and species climatic habitat suitability since projected climate change in the twenty-first century has been previously shown to be potentially detrimental for mesic species such as P. sylvestris but less for more xeric species such as Q. ilex and P. halepensis due to changes in the species distribution (Benito Garzón et al. 2009; Ruiz-Labourdette et al. 2012 but see Keenan et al. 2011). In addition, the future coupling of a land use change model with the MEDFIRE forest landscape dynamics model will allow to improve afforestation and deforestation modelling as a function of the main socioeconomic drivers strongly influencing forest landscape dynamics in the Mediterranean (e.g., rural land abandonment, urbanization and agricultural conversion). Further research should also focus on explicitly considering the influence of extreme weather events on fire regime and the influence of young forests increasing proportion on potential reductions of annual target burnt area. Finally, when necessary, we will revisit the parameterized processes following the state-of-the-art in ecological modelling.


Global change processes have influenced Mediterranean forest landscape composition and extent since the second half of the twentieth century. Ongoing afforestation has been forecasted in our simulations, but pine-oak forest landscape composition dynamics in a near future may be strongly influenced by a harsher fire regime that may accelerate a shift towards landscapes less dominated by pines but with more oaks and shrubby vegetation. Forest ageing may be also hampered and in the worst scenarios significant net forest losses may occur. Fire sensitive species (P. pinea, P. sylvestris and P. nigra) may be the most affected, whereas P. halepensis may also be impacted by fire regimes with increasing fire recurrence and climate change affecting the probability of ignition and afforestation. An active fire suppression capacity may potentially offset fire regime harshness and should therefore be considered in forest landscape composition forecasts to handle future uncertainty. To conclude, although forest ecosystem functioning may be locally altered as a consequence of fire regime, ongoing afforestation may counteract negative impacts at the regional scale and suppression opportunities may arise from forest landscape composition dynamics favouring less fire-prone species (oaks) or fuel limitations (shrublands) due to climate change.


  1. Améztegui A, Brotons L, Coll L (2010) Land-use changes as major drivers of mountain pine (Pinus uncinata Ram.) expansion in the Pyrenees. Glob Ecol Biogeogr 19:632–641

    Google Scholar 

  2. Anderson HE (1982) Aids to determining fuel models for estimating fire behavior. General Technical Report INT-122. Intermountain Forest and Range Experiment Station Ogden. USDA Forest Service

  3. Barros A, Pereira J (2014) Wildfire selectivity for land cover type: Does size matter? PLoS ONE 9:e84760

    Article  PubMed  PubMed Central  Google Scholar 

  4. Benito Garzón M, Sánchez de Dios R, Sainz Ollero H (2009) Effects of climate change on the distribution of Iberian tree species. Appl Veg Sci 11:169–178

    Article  Google Scholar 

  5. Bielsa I, Pons X, Bunce B (2005) Agricultural abandonment in the North Eastern Iberian Peninsula: the use of basic landscape metrics to support planning. J Environ Plan Manag 48:85–102

    Article  Google Scholar 

  6. Blondel J, Aronson J (1999) Biology and wildlife of the Mediterranean region. Oxford University Press, Oxford

    Google Scholar 

  7. Broncano MJ, Retana J, Rodrigo A (2005) Predicting the recovery of Pinus halepensis and Quercus ilex forests after a large wildfire in northeastern Spain. Plant Ecol 180:47–56

    Article  Google Scholar 

  8. Brotons L, Aquilué N, de Cáceres M et al (2013) How fire history, fire suppression practices and climate change affect wildfire regimes in Mediterranean landscapes. PLoS ONE 8:e62392

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  9. Carnicer J, Coll M, Pons X et al (2014) Large-scale recruitment limitation in Mediterranean pines: the role of Quercus ilex and forest successional advance as key regional drivers. Glob Ecol Biogeogr 23:371–384

    Article  Google Scholar 

  10. Costa P, Castellnou M, Larrañaga A et al (2011) La prevenció dels grans incendis forestals adaptada a l’Incendi Tipus. Unitat Tècnica del GRAF (ed), Divisió de Grups Operatius Especials, Direcció General de Prevenció, Extinció d’Incendis i Salvaments, Departament d’Interior, Generalitat de Catalunya

  11. De Cáceres M, Brotons L, Aquilué N, Fortin M-J (2013) The combined effects of land-use legacies and novel fire regimes on bird distributions in the Mediterranean. J Biogeogr 40:1535–1547

    Article  Google Scholar 

  12. Debussche M, Lepart J, Dervieux A (1999) Mediterranean landscape changes: evidence from old postcards. Glob Ecol Biogeogr 8:3–15

    Article  Google Scholar 

  13. Duane A, Piqué M, Castellnou M, Brotons L (2015) Predictive modeling of fire occurrences from different fire spread patterns in Mediterranean landscapes. Int J Wildl Fire 24:407–418

    Article  Google Scholar 

  14. Fernandes P (2009) Combining forest structure data and fuel modelling to classify fire hazard in Portugal. Ann For Sci 66:1–9

    Article  Google Scholar 

  15. Fischer EM, Schär C (2010) Consistent geographical patterns of changes in high-impact European heatwaves. Nat Geosci 3:398–403

    CAS  Article  Google Scholar 

  16. Fried JS, Gilless JK, Riley WJ et al (2007) Predicting the effect of climate change on wildfire behavior and initial attack success. Clim Change 87:S251–S264

    Article  Google Scholar 

  17. IPCC (2000) Summary for Policymakers: Emissions Scenarios. A Special Report of Working Group III of the Intergovernmental Panel on Climate Change

  18. Keenan T, Maria Serra J, Lloret F et al (2011) Predicting the future of forests in the Mediterranean under climate change, with niche- and process-based models: CO2 matters! Glob Chang Biol 17:565–579

    Article  Google Scholar 

  19. Lloret F, Calvo E, Pons X, Díaz-Delgado R (2002) Wildfires and landscape patterns in the Eastern Iberian Peninsula. Landsc Ecol 17:745–759

    Article  Google Scholar 

  20. Lloret F, Pausas JG, Vila M (2003) Responses of Mediterranean plant species to different fire frequencies in Garraf Natural Park (Catalonia, Spain): field observations and modelling predictions. Plant Ecol 167:223–235

    Article  Google Scholar 

  21. Loepfe L, Martinez-Vilalta J, Piñol J (2012) Management alternatives to offset climate change effects on Mediterranean fire regimes in NE Spain. Clim Change 115:693–707

    Article  Google Scholar 

  22. Montserrat Aguadé D (1998) Situaciones sinópticas relacionadas con el inicio de grandes forestales en Cataluña. Nimbus 1–2:93–112

    Google Scholar 

  23. Moreno MV, Conedera M, Chuvieco E, Pezzatti GB (2014) Fire regime changes and major driving forces in Spain from 1968 to 2010. Environ Sci Policy 37:11–22

    Article  Google Scholar 

  24. Moriondo M, Good P, Durao R et al (2006) Potential impact of climate change on fire risk in the Mediterranean area. Clim Res 31:85–95

    Article  Google Scholar 

  25. Moritz MA, Parisien M-A, Batllori E et al (2012) Climate change and disruptions to global fire activity. Ecosphere 3:art49

    Article  Google Scholar 

  26. Moya D, Alfaro-Sánchez R, López-Serrano F et al (2014) Post-fire management of Mediterranean forests: Carbon storage in regenerated areas in eastern Iberian peninsula. Cuad Investig Geográfica 40:371–386

    Article  Google Scholar 

  27. Pausas JG (2006) Simulating Mediterranean landscape pattern and vegetation dynamics under different fire regimes. Plant Ecol 187:249–259

    Article  Google Scholar 

  28. Pausas JG, Fernández-Muñoz S (2011) Fire regime changes in the Western Mediterranean Basin: from fuel-limited to drought-driven fire regime. Clim Change 110:215–226

    Article  Google Scholar 

  29. Pausas JG, Paula S (2012) Fuel shapes the fire-climate relationship: evidence from Mediterranean ecosystems. Glob Ecol Biogeogr 21:1074–1082

    Article  Google Scholar 

  30. Piñol J, Terradas J, Lloret F (1998) Climate warming, wildfire hazard, and wildfire occurrence in coastal eastern Spain. Clim Change 38:345–357

    Article  Google Scholar 

  31. Piqué M, Valor T, Castellnou M, et al (2011a) Integració del risc de grans incendis forestals (GIF) en la gestió forestal: Incendis tipus i vulnerabilitat de les estructures forestals al foc de capçades. Sèrie: Orientacions de gestió forestal sostenible per a Catalunya (ORGEST). Centre de la Propietat Forestal. Departament d’Agricultura, Ramaderia, Pesca, Alimentació i Medi Natural, Generalitat de Catalunya

  32. Piqué M, Vericat P, Cervera T, et al (2011b) Tipologies forestals arbrades. Sèrie: Orientacions de gestió forestal sostenible per a Catalunya (ORGEST). Centre de la Propietat Forestal. Departament d’Agricultura, Ramaderia, Pesca, Alimentació i Medi Natural, Generalitat de Catalunya

  33. Podur J, Wotton M (2010) Will climate change overwhelm fire management capacity? Ecol Modell 221:1301–1309

    Article  Google Scholar 

  34. Puerta-Piñero C, Espelta JM, Sánchez-Humanes B et al (2012) History matters: previous land use changes determine post-fire vegetation recovery in forested Mediterranean landscapes. For Ecol Manage 279:121–127

    Article  Google Scholar 

  35. Regos A, Aquilué N, Retana J et al (2014) Using unplanned fires to help suppressing future large fires in mediterranean forests. PLoS ONE 9:e94906

    Article  PubMed  PubMed Central  Google Scholar 

  36. Retana J, Espelta JM, Habrouk A et al (2002) Regeneration patterns of three Mediterranean pines and forest changes after a large wildfire in northeastern Spain. Ecoscience 9:89–97

    Google Scholar 

  37. Rocca ME, Brown PM, MacDonald LH, Carrico CM (2014) Climate change impacts on fire regimes and key ecosystem services in Rocky Mountain forests. For Ecol Manage 327:290–305

    Article  Google Scholar 

  38. Rodrigo A, Retana J, Picó FX (2004) Direct regeneration is not the only response of Mediterranean forests to large fires. Ecology 85:716–729

    Article  Google Scholar 

  39. Ruiz-Labourdette D, Nogués-Bravo D, Ollero HS et al (2012) Forest composition in Mediterranean mountains is projected to shift along the entire elevational gradient under climate change. J Biogeogr 39:162–176

    Article  Google Scholar 

  40. Sheffer E (2012) A review of the development of Mediterranean pine–oak ecosystems after land abandonment and afforestation: are they novel ecosystems? Ann For Sci 69:429–443

    Article  Google Scholar 

  41. Stephens SL, Agee JK, Fulé PZ et al (2013) Managing forests and fire in changing climates. Science 342:41–42

    CAS  Article  PubMed  Google Scholar 

  42. Vega-García C, Chuvieco E (2006) Applying local measures of spatial heterogeneity to Landsat-TM images for predicting wildfire occurrence in Mediterranean landscapes. Landsc Ecol 21:595–605

    Article  Google Scholar 

  43. Verburg PH, van Berkel DB, van Doorn AM et al (2010) Trajectories of land use change in Europe: a model-based exploration of rural futures. Landsc Ecol 25:217–232

    Article  Google Scholar 

  44. Zavala MA, Espelta JM, Retana J (2000) Constraints and trade-offs in Mediterranean plant communities: the case of holm oak-Aleppo pine forests. Bot Rev 66:119–149

    Article  Google Scholar 

Download references


MEDFIRE development was supported by the Spanish Government through the FORESTCAST (CGL2014-59742-C2-2-R), BIONOVEL (CGL2011-29539/BOS) and MONTES-Consolider (CSD28008-00040) projects, NEWFORESTS (EU’s 7th programme, PIRSES-GA-2013-612645) and the ERA-NET FORESTERRA project INFORMED (29183). We thank the Servei de Prevenció d’Incendis de la Generalitat de Catalunya for providing data on fire perimeters and ignitions. Miquel Ninyerola and Meritxell Batalla (UAB) generate spatially explicit climatic predictions from data provided by the Spanish Meteorological Agency and the Spanish Ministry of Marine and Rural Environment within the MONTES-Consolider project. A. Gil-Tena (Juan de la Cierva fellow, JCI-2012-12089) and M. De Cáceres (Ramón y Cajal fellow, RYC-2012-11109) are funded by Ministerio de Economía y Competitividad (Spain), Andrea Duane (PhD student grant FPU13/00108) by the Ministerio de Educación, Cultura y Deporte (Spain) and N. Aquilué (Forest Complexity Modelling fellow) by Natural Sciences and Engineering Research Council of Canada. We thank Mario Beltrán for his valuable help in age initialization.

Author information



Corresponding author

Correspondence to Assu Gil-Tena.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary material 1 (DOCX 463 kb)

Supplementary material 2 (DOCX 450 kb)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Gil-Tena, A., Aquilué, N., Duane, A. et al. Mediterranean fire regime effects on pine-oak forest landscape mosaics under global change in NE Spain. Eur J Forest Res 135, 403–416 (2016).

Download citation


  • Afforestation
  • Firefighting
  • Ignition probability
  • IPCC-SRES scenarios
  • Mediterranean forests
  • Post-fire regeneration