IntroductionTop

Invasive species have been considered one of the main causes of habitat degradation (Gurevitch & Padilla, 2004). This invasion is closely linked to global change and globalization (Kaluza et al., 2010) and has had detrimental economic impacts on human enterprises caused by the spread of plant diseases and competition with agricultural species (Shacklenton et al., 2007; Arévalo & Fernández-Palacios, 2005; Arévalo et al., 2017). Specific elements of globalization that explain the spread of invasive species around the planet are: global changes (e.g. global warming, nitrogen deposition, habitat fragmentation, Dukes & Mooney, 1999); socioeconomic changes, such as gross domestic product (GDP) in relation to global trade and intercontinental trade (Lin et al., 2007; Sharma et al., 2010; Arévalo et al., 2017); transport (Westphal et al., 2008) and tourism (Sutherst, 2000) among others. Indeed, globalization has been one of the most substantial drivers behind the homogenization of insular biotas (Arévalo et al., 2010a) and is linked to a worldwide alien species expansion (Levine & D´Antonio, 2003; Westphal et al., 2008; Pysek et al., 2010). Ultimately, the consequences of invasive species affect different spheres such as the economy (Pimentel et al., 2005), health (Levine & D´Antonio, 2003) and ecological processes (Reaser et al., 2007). The cost of their eradication is also increasing as many more species are introduced into new habitats. The risks of new diseases and problematic weeds for agriculture and how these introduced species disrupt ecological processes in the native plant community have been widely reported (Pimentel et al., 2005).

On islands, infrastructure development is also directly related to the number of invasive species, whose impact can be either immediately evident or can lag for a period over 100 years (Weber & Li, 2008). Moreover, the extinction of native island biota and dismantling of island ecosystems worldwide have been some of the consequences of the high numbers of non-native species introduced (Donlan & Wilcox, 2008; Kueffer et al., 2010). Although we can consider non-native plants as important components of plant communities in many parts of the world (Mack et al., 2000), they have also become a global problem that requires analysis and evaluation of the ecosystems under threat. In some cases, the impact can even reach levels of ecosystem modification that completely alter the plant community, as is the case of Myrica faya in Hawaii (Lenz & Taylor, 2001) or Bromus tectorum in California (DiTomaso, 2000), while in others they only act by basically occupying disturbed areas and being more competitive than native species (Arévalo et al., 2010b).

The objective of this study is to test the recovery of native plant communities under the closed canopies of exotic species. Our hypothesis is that species composition of the understory below the canopy of the exotic Quercus suber may be favored by the environmental conditions and will be dominated by the native plant community, the laurel forest, following a “catalytic” effect of this exotic stand on succession. This effect can be revealed at different levels, such as a increases in soil humidity, organic matter or shade that favor shade intolerant native species (Arévalo et al., 2005), or other processes related to the provision of benign microhabitats that are more favorable for seed germination and/or seedling recruitment than their surrounding environment (Harris & Harris, 1997; Lugo, 1997; Parotta, 1995; Parotta et al., 1997; Whitmore, 1999).

We also aim to put forward proposals to manage the forest following a landscape, ecological, forestry approach involving the manipulation of stands in order to help them recover from natural or anthropogenic disturbances such as afforestation with an exotic species (Oliver & Larson, 1996).

MethodsTop

Study site

The study site is located in the northeastern part of the island of Gran Canaria (Fig. 1), in the Finca de Osorio, a public property of the Cabildo de Gran Canaria included in the Doramas Rural Park (geographical coordinates: 28° 4’ N, 15° 32’ W). A small part of the Quercus suber stand is on the border of the Park. The plots were established in an area of 5 ha where there is the highest density of Q. suber in the Park. Although disperse individuals can be found in an area of around 10 ha due to the interest in this species by the former owners of the property. The altitude of this stand is around 650-670 m. The climate is influenced by the trade winds, which produce a “sea of clouds” on the windward slopes and wetter conditions on the northern than on the southern slopes. The annual precipitation of the park reaches 550 mm, but can be twice this amount if fog drip is taken into account (Kämmer, 1974). The mean annual temperature is close to 15 ºC with an absence of frost events (and slight variation in minimal annual and daily fluctuations in temperature). These environmental conditions are similar to the conditions of the native areas of the species, Q. suber, such as in the south of Spain (Natural Park of Alconorcales (Pérez-Ramos et al., 2008), with a bioclime classified as thermomediterranea inferior subhumid, with temporal presence of fogs. The Doramas Park’s vegetation is dominated by laurel forest vegetation with dominant species such as Arbutus canariensis, Ilex canariensis, Persea indica, Morella faya, Visnea mocanera (Lauro-Perseetum indicae sigmetum Association) (Del Arco et al., 2002; Del Arco et al., 2006). In addition, the geology consists of a mainly volcanic area of 2.8-4 My (Balcells et al., 1992) with basaltic and pyroclastic rocks and high dominance of clays that classify these soils as endosols (Sánchez, 1975).

The species

Quercus suber L., commonly called the cork oak, is a medium-sized, evergreen oak tree in the section Quercus sect. cerris. It is a species of slow growth, which grows up to 20 m in height. The leaves are 4 to 7 cm long, weakly lobed or coarsely toothed, dark green above, paler beneath, with the leaf margins often downcurved. The acorns are 2 to 3 cm long, in a deep cup fringed with elongated scales. It is native to southwest Europe and northwest Africa (Spain, France, Morocco, Algeria and Tunisia). However, one species of the genus Quercus is considered to have been present in the Canary Islands 2,400 years ago, based on paleoecological studies (de Nascimento et al., 2016). In the Mediterranean basin, the tree is an ancient species with fossil remnants dating back to the Tertiary period. Although extensively intro-duced around the planet in appropriate environments, it is not considered an invasive species (Pausas et al., 2009). In natural areas, there are high acorn predation rates (both predispersal and post-dispersal) and high seedling mortality (during the summer drought). These are the main causes limiting the natural regeneration of Q. suber, which are related to microsite environmental variability and reduced probability of survivorship (Pérez-Ramos et al., 2008).

The species was introduced in Gran Canaria following the European conquest of the island, around the 16th century due to its many uses in agriculture, especially in cork production. The first reference is from the naturalist Viera y Clavijo (at the beginning of the 18th century) around the area of the study site, though due to the few references about this species, it is suspected that it had a restricted extension (Bory de Saint-Vincent, 1988).

Sampling design

In 2018, in the Quercus suber stand of the Finca de Osorio, we established a permanent network of 13 square plots (20 x 20 m) following a straight line from the centre of the stand and separated by at least 100 m (except for three located in a different Quercus suber stand, but located following the same method). These plots were located systematically in an area where no other anthropogenic disturbances apart from the introduction of the exotic species had occurred (areas of trails with human interventions were avoided in the sampling).

In each plot, we measured altitude and slope, aspect and estimated canopy cover of the stand using a convex spherical densitometer (Lemmon, 1957). We also visually estimated rock, bare soil, and litter cover within each plot on a scale of 1 to 100%.

We sampled the plant community of the plots at canopy level and the understory species composition. For the canopy, we sampled the trees, which were defined as stems with at least 2.5 cm diameter at breast height (dbh) and saplings as individuals over 50 cm height and less than 2.5 cm of diameter dbh, as an indication of regeneration of the stand. All the woody species were sampled. At the understory level, we noted species composition (all the plant community) and the cover of all species on the surface was estimated and noted on a scale of 1 to 10 (Cover classes: 1: traces, 2: >1% of cover in the plot, 3: 1-2%, 4: 2-5%, 5: 5-10%, 6: 10-25%, 7: 25-50%, 8: 50-75%, 9: >75%).

Soil samples to describe the plots' nutrient content were collected at 0-10 cm depth at 30 cm distance from each plot corner, mixing the four samples for each plot to form a composite sample. The pH, EC (exchangeable cations, dS/m), Polsen (Phosphorus Olsen extraction on ppm), organic matter (%OM), total nitrogen percentage (TN), nitrate percentage (NO3) and available cations in meq/100 (Na, K, Ca, Mg) were analyzed. The percentages of clay, sand, silt and limestone (Calcimeter Bernad method) were also analyzed. Standard methods of analysis were followed (AOAC, 1990; Anonymous, 1986). Soil samples were collected in May 2018.

Statistical analysis

To determine species diversity of the understory in the plant community dominated by oaks, we calculated species richness per plot as well as Smith and Wilson evenness index (Smith & Wilson, 1996).

Canopy and regeneration analysis

Basal area of trees and density of saplings were calculated per plot and used in the analyses. Ordination techniques help in explaining community variation (Gauch, 1982) and were used to evaluate trends in plant species composition (ter Braak & Šmilauer, 1998).

We based the analysis on the percentage (in order to standardized basal area of trees and density of samplings) of the species basal area of the plots together with the percentage density of saplings using a DCA (Detrended Correspondence Analysis). In the plane given by DCA axes I and II, we encircled the species of trees with an envelope, using the minimum possible area and in another polygon, the species of saplings.

Understory species composition analysis

A second ordination analysis was used to determine trends in understory species composition. We used partial Canonical Correspondence Analysis (CCA; Hill & Gauch, 1980) in CANOCO (ter Braak & Šmilauer, 1998) to examine how species composition changed over the different plots as a function of the environmental characteristics included in the analysis. In the environmental matrix, we included the following variables: altitude, slope, aspect (N or S), % of clay, silt, sand and limestone, pH, EC (exchangeable cations, dS/m), Polsen (Phosphorus Olsen extraction on ppm), organic matter (%OM), available cations in meq/100 (Na, K, Ca, Mg). As a biotic matrix, we used the species composition based on cover of the plots’ understory. We selected the five most informative environmental variables applying a forward selection procedure to remove the variables that did not explain a significant portion of the variability reported by the analysis when performing the axes (Monte Carlo permutation test 499 interactions for p<0.05). Axes I and II are graphically displayed with the selected environmental variables, plots and species.

ResultsTop

The plots did not present substantial environmental differences, with similar altitudes, slopes, canopy cover and litter cover. Some variability can be considered part of the microsite heterogeneity in the area (Table 1). In the case of the soil nutrient composition, there was greater variability in the case of some nutrients. For example %OM presented a range of variations from 4.29 to 13.8 and Polsen varied from 6 to 40 ppm (Table 2).

Table 1. General abiotic and plant community information of the plots.

Table 2. Nutrient content and soil structure characteristics of the soil of the plots.

The tree canopy composition was completely dominated by Quercus suber. In some plots, this was almost the only species (>99% total basal area of the plots QS2, 6, 7, 12 and 13). Even the plot with the lowest dominance in basal area of Q. suber reached almost 75% of the total (Average basal area of Q. suber of all the plots 93.1 ± 8.3 %; Table 3).

Table 3. Basal area of trees (>2.5 cm DBH) per species and total basal area per plot.

In the case of canopy regeneration, Q. suber also dominated in all plots with over 80% of density, with the exception of one plot (QS10). However, three more species were also present: Pinus canariensis, Laurus novocanariensis and Olea cerasiformis. In four plots, the woody regeneration of the plots was only Q. suber. The average regeneration of all plots for Q. suber was more than 92% including the non-dominant plot of the species (Fig. 2 represents the decimal logarithm of regeneration density in order to visualize species with low densities).

Figure 2. Regeneration densities per plot and species (individuals less than 2.5 cm dbh and over 50 cm tall).

The plant community analysis of the trees and regeneration shows the species scores in Fig. 3. These species are enclosed in the same polygon depending on whether their values are from basal area species or sapling species. Discrimination among polygons (as long as this discrimination is related to differences in relative abundance in the plots) revealed differences in the regeneration community and canopy tree com-munity. As the analysis shows, the polygon enclosing the regeneration species is included in the polygon of the tree species. There was much more variability in the canopy tree community than in the regeneration one (as we have seen previously, with only four species). This indicates that species in the canopy will be partially replaced and no major changes are expected based on regeneration. We should note the important impact on the regeneration plant community of Olea cerasiformis, a species favored by the present environmental conditions created by the canopy of Q. suber with respect to the others present in the canopy.

Figure 3. Detrended correspondence analysis axes I and II. Species coordinates of trees (based on percentage of basal area) are enclosed in an envelope with a solid line while species coordinates of saplings (based on percentage of density) are enclosed in an envelope with a dotted line. Eigenvalues for axis I: 0.15, eigenvalue for axis II: 0.08, cumulative percentage of total inertia for axes I and II: 57%. The codes of the species are the first three letters of the genus followed by the three first letters of the specific epithet (full name of the species in Table S1 [suppl.]).

A total of 93 species were found in the study (Table S1 [suppl.]), and there was some variation in the number of species among plots, from 16 (plot QS13) to 43 (plot QS8). However, Smith and Wilson evenness revealed high values on the plots and low differences among these (from 0.72 to 0.84).

The CCA for all the plant community revealed that the most significant environmental characteristics for the plant community (Fig. 4) were C/N, %Limestone, %Silt, NO3 and Na. There was also substantial vari-ability in the plant community, such as plots QS13, QS12 and QS11 had low levels C/N and limestone and were characterized by species such as Opuntia maxima, Erica arborea, Rubus ulmifolius of Hy-pericum canariense, whereas plots QS7, QS4 and QS5 had high values of NO3 and Na and were mainly characterized by species like Viburnum rigidum, Phagnalon purpurascens or Andryala integrifolia. Plots QS8 and QS6 had high C/N values and were characterized by species such as Carlina salicifolia, Teline microphylla or Castanea sativa (Fig. 4).

Figure 4. Canonical correspondence analysis using five selected environmental variables (Limestone, Silt, Na, C/N and NO3 as NO3), plot scores and species scores (based on species cover of the plot). Eigenvalues axis I: 0.27; axis II: 0.242; both axes cumulative percentages explained by species composition 26.6% and both axes cumulative percentages variance explained by species-environmental relationship: 49.9%. The codes of the species are the first three letters of the genus followed by the three first letters of the specific epithet (full name of the species in Table S1 [suppl.]). Invasive species are indicated in bold font.

Common species in all the plots were located as the origin of coordinates such as Quercus suber, Laurus novocanariensis, Sonchus acaulis, Oxalis pes-caprae or Olea cerasiformis. With respect to native or invasive species, there was no distribution specific pattern on the defined bi-dimensional space of CCA. Some of the species were common for all the plots like Q. suber, Oxalis pes-caprae, and others were only found near to some plots such as Scorpiurus muricatus, Brunsvigia rosea (QS8) or Opuntia maxima (QS11 and QS12).