On 15 October, 2017, major wildfires contributed toward Portugal being the European country with the highest number of forest fires and burnt areas in 2017, corresponding to 60% of the total area burnt in the European Union [1
]. A 9400 ha fire in the largest Portuguese public forest, Leiria National Forest (Mata Nacional de Leiria—MNL), also known as “King’s Pine Forest”, contributed to these dramatic figures. The main losses included 85% of the maritime pine stands (Pinus pinaster
Aiton), emblematic ancient pines in coastal dunes, and riparian forests, 84% of which were affected by the fire [2
]. This forest is more than 800 years old. First plantations of maritime pine were ordered by the Portuguese King D. Afonso III, and date back to the 13th century, with the initial goal of protecting the croplands from the coastal winds loaded with salt and sand [3
]. The greatest expansion took place during the D. Dinis kingdom (1279–1325), which was crucial for the provision of timber for shipbuilding for the Portuguese Maritime Discoveries. Timber production is the main function, followed by recreational activities and protection from erosion [4
MNL has a long history of wildfires, c. 5000 ha, burnt in 1824, and smaller but damaging wildfires were recurrent with short fire return intervals in the final decade of the 20th century [3
]. Nevertheless, the most devastating wildfires occurred in the last two decades, during the heat waves of summer 2003 and autumn 2017. Fire-prone climate, climate change (heatwaves and droughts), societal changes, with deficient forest management, and human activity—either deliberate or negligent—were regarded as the main drivers [6
]. Portugal is considered a hot-spot of land-use and land cover change and, accordingly, it has recently been observed that, at a local level, Portugal (and Southern Europe) showed a higher concern for civil protection actions than the rest of Europe, which is probably related to the increasing number of wildfires, in addition to problems caused by extreme climate events, such as heatwaves and floods [7
Mediterranean regions are naturally fire-prone environments, supported by direct plant regeneration mechanisms, coined in literature as “autosuccession” or “direct succession” [9
]. Different strategies for post-fire survival and regeneration contribute to the autosuccession phenomenon. Resprouter species regenerate from lignotubers belowground or aboveground organs, whereas obligate seeders rely on a fire-cued seed bank for re-establishment after the fire [10
]. This last group includes the rosemary, Rosmarinus officinalis
L., species of the genera Pinus
sp. and Cistus
sp., and many others. Some species show passive resistance via morphological and anatomical features, such as leaves with low water content or fire-resistant bark (e.g., cork oak, pines), and some species display different combinations of pyrophytic life-history traits. For instance, the Tasmanian blue gum (Eucalyptus globulus
Labill.) shows vigorous epicormic resprouting after crown fires in Southern Europe, and the capsules are thermo-dehiscent favoring post-fire re-establishment of populations [11
]. Maritime pine is fire-killed, though the relatively thick bark conveys good resistance to fire. A great capacity of post-fire regeneration from soil seed banks has also been recognized, dependent on the age of the stand before the fire, and also from the fire-mediated opening of the thermo-dehiscent cones. The production of viable seeds by maritime pines occurs when pines are 15–20 years old [13
]. The pine fruiting structures have a certain degree of serotiny, i.e., the cones have a resinous bond among scales that allow remaining closed at maturity, and free seeds when affected by heat [14
]. So, the success of natural regeneration of pine forests is strongly influenced by fire severity (i.e., measurable loss or change in aboveground and belowground biomass) [16
], fire frequency, and on many environmental factors, including topography, soils, and climate [12
]. Moreover, biotic factors, such as the age of the stand, and also the biotic legacies, such as propagule pressure and seed bank, are key to the post-fire recovery [19
Fire promotes short-term successional processes of the understory vegetation that mostly occur in the first two years after fire [22
]. Plant-community responses are complex and multiple abiotic and biotic factors interact for the recovery of natural vegetation [18
]. In this sense, the riparian forests are similar, holding a variety of woody and herbaceous plant communities under the canopies, and provide for important ecosystem services [24
]. In the MNL, riparian forests cross through the maritime pine stands along small-sized streams, which support a relevant share of fauna and flora biodiversity [25
]. Specifically, the São Pedro River catchment, the most important watercourse in the area, was directly affected by the 2017 fire in around 84% of the area [25
]. Many studies are devoted to understanding the effects of fire on the physical, chemical, and ecological characteristics of rivers and riparian zones [26
]. However, few studies are addressing, simultaneously, the effects of a fire event on rivers and managed forests (e.g., [29
This work aims to quantify the effect of a large fire on pine regeneration and plant community composition in pine stands, riparian areas and in river channels. In addition, we aim to answer the following questions:
Are there differences in plant composition between understory vegetation of pine stands and riparian ecosystems after the fire?
Does fire severity affect the regeneration potential of pine stands and riparian forests?
Does the age of pines significantly influence the regeneration of pine stands?
For this, we made field surveys on pine stands crossed by the rivers and related riparian forests along of São Pedro River. The study area constitutes a good experimental area, as the stream was partially affected by the fire and there were historical data on vegetation (pre-fire data), allowing comparison between pre- and post-fire communities and reference/control unburnt areas.
2. Materials and Methods
2.1. Study Area
Sampling sites are in MNL, a state forest (mostly of maritime pine, Pinus pinaster
Aiton) located in the central west coast of Portugal (Figure 1
). It has a Cool-summer Mediterranean climate (Csb) according to the Köppen classification, with a strong oceanic influence, which contributes to high average air humidity levels (81–83%). It is characterized by dry summers and mild winters, with a mean annual temperature of 14–16 °C, and mean annual precipitation of 710–909 mm [31
]. On the day of the fire (15 October, 2017), the temperature ranged from 25 °C to 35.8 °C and there was no precipitation. On the first period of the survey (May–July 2018), the average monthly temperature and precipitations were 18.0 °C and 28.9 mm, and on the second survey (May 2019), they were 17.4 °C and 17.7 mm, respectively (data from the Portuguese Institute for Sea and Atmosphere, I.P. (IPMA, IP) for the Leiria Meteorological Station; www.ipma.pt
; accessed on 25 March 2021).
The area is included in the Lusitanian Basin and is composed mostly of quaternary and neogenic sediments (silts and sands) above Jurassic and Cretaceous formations. The relief is flat to wavy with three main strands parallel to the coast. The Arenosols are dominant, followed by relatively incipient Podzols [32
MNL is crossed by diverse surface water bodies, of which the most important are the lagoons Lagoa da Ervideira (CORINE Biotope Pinhal de Leiria, C12300073), Lagoa da Saibreira, and the watercourses Tábuas River and São Pedro River. Apart from the maritime pine, MNL is home to other forestry species that were planted, including Eucalyptus globulus, Pinus pinea L., Quercus rubra L., Quercus robur L., Taxodium distichum (L.) Rich., Laurus nobilis L., and Acacia melanoxylon R.Br., amongst others. Riparian forests along the MNL small-sized streams are mainly composed of alders (Alnus glutinosa (L.) Gaertn.), poplars (Populus nigra L.), ash (Fraxinus angustifolia Vahl), willows (Salix sp.), and several shrubby species (e.g., Frangula alnus Mill., Crataegus monogyna Jacq.). In past years, several patches of exotic invasive species were observed, such as diverse species of Acacia, Hakea sericea Schrad. & J.C.Wendl., Robinia pseudoacacia L., especially near roads and streams.
2.2. Sampling Design and Surveys
MNL is composed of 342 rectangular numbered management units for timber production (T) of 35 ha each (430 × 800 m2
). Pine ages within units were relatively homogeneous and ranged in the MNL from young saplings to more than 90 years old, due to management planning and the effect of recurrent wildfires (Supplementary Materials, Figure S1
). Our study area includes five management units, namely T259, T260, T261, T262, and T277, placed along São Pedro River, and for which there was no timber extraction in the first year after the 2017 fire. Concerning the sampling locations on the channel and riverbanks of São Pedro River, we set out eight locations from stream headwaters (two tributaries) to the river mouth. Sampling locations were randomly defined at the office and slightly altered during a prospecting field campaign conducted in April 2018 to evaluate the accessibility restrictions. SP4 was an exception, as it was a location with pre-fire data of plant species composition (Figure 1
For each management unit, five to six square plots (3.5 × 3.5 m2
) on maritime pine stands were defined in transects parallel to the largest dimension of the management unit, drawn approximately in the middle, and distanced from each other on sequences of 50, 100, and 200 m. Some adjustments to this sequence had to be made during the field campaign given the differences in slope, soil disturbances, and composition by diverse species of trees other than pine (Figure 1
and Figure 2
a). All understory plant species observed in the 12.25 m2
plots were identified and abundances were estimated by percent aerial cover. Surveys took place in May–July 2018 (n
= 28 plots). Pine seedlings were counted in each plot (Figure 2
b). A seedling is defined as a young plant grown from seed up to 25 cm high [33
]. It was not possible to repeat these surveys in 2019 due to widespread damages in vegetation and pine seedlings caused by timber extraction.
Surveys were done on six sampling locations along the 6350 km of São Pedro River, SP#, and two surveys on upstream tributaries, namely RRT and RBA (Figure 1
a). Plots SP6 and SP5 were used as control sites (unburnt), and SP4 (burnt site) has historical data, from a survey done in May 2004 by one of the team members (Figure 2
c). The survey of 2004 plotted together both riverbanks. Sampling surveys were done in May 2018 and May 2019 (Figure 2
d). Sampling locations SP3, SP2, and SP1 were disturbed from forestry and logging machinery in 2019, and it was not possible to survey SP5 due to safety concerns (risk of trees falling). Three sampling plots (5 × 20 m2
) were outlined in right and left riverbanks and inside the channel, parallel to the thalweg, totaling 24 plots at each campaign (2018 and 2019). All vascular plant species (woody and herbaceous) were recorded, and their abundance cover estimated by visual assessment of the percent aerial cover of each species (100 m2
). Surveys were made by zigzagging across the sampling plot starting from downstream to upstream. Then a downstream observation was done to ensure that all species were recorded and to confirm the species abundance attributed in the first assessment. The percent cover of each species was estimated by two surveyors and then compared to minimize estimation errors. We used a scheme of the rectangular sampling plots divided into grids (100 cells) to support the visualization of percent species cover in the field, which is included in the Spanish field protocols for fluvial plants (macrophytes) [34
Specimens that could not be identified in field were collected for later identification in the João Carvalho e Vasconcellos Herbarium (LISI), Lisbon. We used the “New Flora of Portugal: Continent and Azores” (in Portuguese language) for the identification, nomenclature, and biogeographic origin (native; exotic) of the species [35
]. Exotic plant species are plants whose presence beyond their natural range is due to intentional or accidental introduction as a result of human activity. Exotic plants can become invasive—i.e., naturalized plants that produce reproductive offspring, often in very large numbers—at considerable distances from the parent plants and, thus, have the potential to spread over a large area [40
]. Invasive species in Portugal are listed in Annex II of Decree-Law no. 92/2019 of July 10, a legislation that promotes the early detection and regulates the possession, cultivation, growing, and trade of the listed speciesSampling plot vertices were georreferenced. Further details can be found in [41
2.3. Fire Severity and Age of Pine Stands
We used the age classes of pine stands defined by the forest inventory carried out in MNL by the National Forest Authority (AFN) (Supplementary Materials, Figure S1
) validated by the forest engineers field observations. On the sampling area, pine stands were from two ages classes: (i) <25 years and (ii) >60 years old.
Fire severity maps were elaborated in 2018 by a group of researchers from the Forest Research Centre, School of Agriculture, University of Lisbon, based on remote sensing observation of Sentinel-2 imagery [41
]. Fire severity levels are related to measurable biomass loss and were derived from a spectral index related to the effects of fire on biomass (Normalized Burnt Ratio—NBR) calibrated with pre-fire data to obtain the spectral variation (DNBR). The fire severity levels are thresholds of DNBR values considered into seven categories (more details in [42
]). In our study area, the management units are included into three classes (low, moderate, high fire severity) for pine stands, and São Pedro River, we used the classification into unburnt, moderate-low, and high severity levels for the sampling plots.
2.4. Statistical Analysis
We assessed the Importance Value (IV) of the species using the sum of the relative percent aerial cover assessed in each sampling plot and the relative frequency, for pine stands and riparian and aquatic vegetation. The relative percent aerial cover was calculated by dividing the total percent aerial cover of species by the total percent aerial cover of all species in each plot, multiplied by 100. The relative frequency was calculated using the number of plots where a certain species was observed divided by the total number of plots of all species, multiplied by 100.
We tested whether the pine stands age significantly contributed to the number of pine seedlings recorded. A t-test was performed to test the null hypothesis that the mean values of natural regeneration for the set of plots, aged less than 25 years and more than 60 years, were equal. Previously, and to select the most appropriate t-test, an F-test was performed to test whether the variance of the two data sets was equal or significantly different [43
We analyzed the dissimilarity between the two vegetation types (pine stands and river) using all species that were observed using an Analysis of Similarity (ANOSIM). Then, we used a non-metric Multidimensional Scaling procedure (nMDS) to address the effects of fire severity on understory vegetation (all species from the surveys of 2018 were included), separately for pine stands and the river. The ordination was performed on sampling plots, which included the relative abundance (percent aerial cover) of all species. Analysis of Similarity tests were used to address the differences of vegetation affected by diverse fire severity levels.
To assess the spatial (longitudinal river gradient) and temporal (2004, 2018, 2019) variation of species composition and abundance, we used a hierarchical classification of sites derived from the Bray–Curtis dissimilarity matrix, and the unweighted pair-group average method, applied to 2018 and 2019 floristic data (percent aerial cover of plants) of the São Pedro River. We validated the groups obtained by observing the significance (p < 0.001) and the degree of segregation in ANOSIM. Analyses were performed with PRIMER (PRIMER-e software ver. 6).