Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction

Human-induced ecosystem and habitat degradation is a widespread phenomenon throughout the world; and “Restoration Ecology”, which studies all the issues of repairing, renewing and restoring ecosystems to a more natural or more sustainable situation, is an actively unfolding field in ecology . The Society for Ecological Restoration (SER) defines ecological restoration as an “intentional activity that initiates or accelerates the recovery of an ecosystem with respect to its health, integrity and sustainability ” (SER 2004). The practice of ecological restoration includes several activities such as: erosion control , reforestation , use of native species collected locally, removal of alien species and weeds, revegetation of disturbed areas, etc. The inexorable growth of human population worldwide, empowered by technological changes and the rise in commodity demands, causes ecosystem destruction at a pace that outraces natural recovery and current restoration activities. Our work is within the frame of the global necessity of gathering efforts to restore ecosystems for our health and well being. Ecosystems have a remarkable potential for self-repair . Given enough time, under stable conditions without human-induced severe disturbance , most degraded ecosystem s can restore themselves. This self-healing capacity is an important and valuable attribute (Bradshaw 1987), but most frequently the time needed to reach a mature and complex stage is so long that active restoration can be a valuable option. Restoration, in this sense, would be a means to accelerate the natural recovery process of a degraded ecosystem.

Exact reproduction of the original community is generally impossible because some site conditions will have been altered irreversibly and some species cannot survive in them any longer or have become locally extinct (Balaguer 2002). Thus, when reflecting on the aims and intention of starting such a project, we assumed that restoration of a forest ecosystem , particularly when starting from a severely degraded stage, is not a perfect reconstruction but only an imitation of the natural forest . We can create a system that is not identical but which can be similar in essential aspects and can function in the same way (Jordan et al. 1987).

Among the different aspects of restoration, the reconstruction of natural forest s is of growing importance in many areas of the world (Jordan et al. 1987) and has been considered an increasingly important activity for environmental sustainability (Miyawaki and Box 2006). In our case, we have performed a forest reconstruction experiment , i.e., an attempt to restore several types of natural forests using only woody species of the native flora . The importance and utility of acquiring knowledge for restoring natural forests is strategic in the context of the conservation policy of any country, and particularly in the European Union, where the Habitats Directive (92/43/EEC) focuses much attention in forests ecosystems (Loidi 1999).

The principles of ecological restoration are based on succession theory. This means that restoration should follow the steps accepted by the geobotanical survey for succession. Precise descriptions of vegetation succession in the past may enable us to make accurate predictions about the nature of future succession (Harper 1987). This has lead to the practice of making chronosequences (space-for-time substitution s), which have certain flaws in their predictive power (Chiarucci et al. 2010). In the absence of a better approach supported sufficiently by scientific documentation, Clementsian climax theory adapted to the concept of potential natural vegetation (PNV, Tüxen 1956) has provided the fundamental conceptual basis describing succession and the basic guidelines for restoration programs (Loidi and Fernández-González 2012). In this sense, mapping PNV types is essential for guiding the spatial aspects of restoration, and some countries have worked on this extensively (Rivas-Martínez 1987, 2007; Miyawaki 1980–1989; Blasi 2010, etc.). PNV analysis and cartography provide the models to be tested in the practice of ecological restoration and can also be useful tools in the field of land management (Biondi 2011). Intensive research on PNV units in the Basque area has been carried out in previous years (Loidi et al. 2005, 2011; Loidi and Báscones 1995) and has described the forest types existing in the area at the moment of project design.

Forest planting in the Basque Country throughout most of the twentieth century was done almost entirely using exotic tree species: Pinus radiata , Eucalyptus globulus , Pseudotsuga menziesii , Chamaecyparis lawsoniana , Larix kaempferi , Quercus rubra , Robinia pseudoacacia, etc. Those plantations were extensive and nowadays cover nearly 30 % of the total area (Inventario Forestal CAE 2005). This intense planting activity has been taking place from the 1930s onwards, because before that the territory had been deforested severely by intensive land use for animal husbandry and agriculture. Such exotic tree plantation s cannot be considered ecological restorations in any case, since they ignored completely the essential requirement that native species be used.

When planning the restoration project, we decided not to follow all the successive steps described by theory based on space-for-time observations. In order to ensure relative success in forest restoration we assumed that: (1) Natural forest restoration requires a minimum area and number of components; monospecific plantation s are not suitable (Kirby and Patterson 1992); and (2) Some typical stages of forest development are not essential and may be bypassed (Ashby 1987). We decided to use several native species which are components of natural forest s in the area. We also wanted to test whether planting mature-stage trees directly was more or less effective than planting secondary forest trees and shrubs, as representatives of earlier stages in succession. We assumed that younger-stage species may modify site conditions favoring the establishment of mature-stage species: this is the facilitation paradigm, which can be applied to forest succession and was meant to be tested in our survey.

Ecosystem restoration has become a very important issue in the Iberian Peninsula in the last years, but most of the projects have been done in the Mediterranean area, due to the urgency created by the desertification threat characterizing that environment (Benayas et al. 2003). In the Atlantic zone of the Iberian Peninsula, such experiences are scarce, particularly those focusing on restoration of natural forest s. For this reason, in 1993, we started a Project entitled Criterios de restauración de la vegetación natural en el territorio de Bizkaia mediante la utilización de especies leñosas de la flora silvestre autóctona (Criteria for restoration of the natural vegetation in the territory of the Basque Country using woody species of the native wild flora). We established a network of plots to plant forest tree species and monitor their long-term growth. The main aim was to gain knowledge of the criteria and techniques for re-vegetation of degraded areas with natural forests. In this paper we present the results of three of those plots after 15 years of monitoring.

The hypothesis we established for this survey were that:

  1. 1.

    Species from seral stages have higher survival rates than long-living trees of mature stage s;

  2. 2.

    Growth during the first years is faster in seral-stage species than in mature-stage trees.

Material and Methods

Study Area

The survey area is located in the Basque Country (northern Spain ), within the area of the Urkiola Natural Park, in the southern sector of the territory of Bizkaia (Basque Country) (Fig. 1). It is a mountainous area crossed by the Duranguesado range, with its highest elevation on the Anboto peak (1330 m); most of the Natural Park area is above 600 m. The climate is temperate-oceanic with cool winters, mild summers, abundant precipitation ranging from 1400 to 2000 mm per year, and no summer drought . In the biogeographic classification of Rivas-Martínez et al. (2007), the area is included in the Basque-Cantabrian sector of the European Atlantic province of the Euro-Siberian region .

Fig. 1
figure 1

Location of the three experimental plots: 1 Subcantabric Mesophytic Oak Forest; 2 Eutrophic Montane Beech Forest; 3 Oligotrophic Montane Beech Forest

Full size image

The Urkiola Natural Park is an area of high ecological diversity , and almost all the natural forest types of the Atlantic Basque Country can be found there. For this survey we have selected three of these forest types:

  • Subcantabrian Mesophytic Oak Forest (Crataego laevigatae-Querctum roboris)

  • Eutrophic Montane Beech Forest (Carici sylvaticae-Fagetum sylvaticae)

  • Oligotrophic Montane Beech Forest (Saxifrago hirsutae-Fagetum sylvaticae).

Plot Design

Three plots were selected, one in each of the three forest types. The characteristics of the three plots are described in Table 1. Each plot involves two replicates of 40×40 m. Both replicates were located close to each other so that environmental conditions would be similar. Each replicate was divided into 6 subplots of 200 m2, and different treatments (plantation density and species combination ) were carried out in each subplot. There was also a control subplot for each replicate (Fig. 2).

Table 1 Description of the three plots
Full size table
Fig. 2
figure 2

Experimental design of the plot replicates with the spatial distribution of 6 subplots plus the control. Species combinations and densities: C, Pre-forest trees mixed with mature forest trees; B, Pre-forest trees mixed with some shrubs and mature forest trees; A, Shrubs mixed with pre-forest trees; Control, Non-planted plot, in order to monitor natural succession of vegetation. 1, Density of 5000 trees/ha; 2, Density of 2500 trees/ha

Full size image

Species Selection

Species were selected using phytosociological information about the area (Loidi et al. 1997), following the successional hypotheses of the vegetation series of the Basque Country (Loidi et al. 2011). The species are shown in Table 2.

Table 2 Species planted in each of the three plots
Full size table

Data Sampling

Plant survival and growth were monitored annually during the first 5 years (1994–1999) and every 5 years thereafter (1999–2008). Survival was quantified as the proportion of living saplings relative to the total number of seedlings planted in each plot. Plant growth was estimated as the absolute growth rate in height and diameter (AGRH and AGRD) and the relative growth rate in height and diameter (RGRH and RGRD) (Hunt 1982):

  • $$ \mathrm{Absolute}\ \mathrm{growth}\ \mathrm{rate}\ \mathrm{in}\ \mathrm{height}:\kern1.25em \mathrm{AGRH}\ \left({\mathrm{t}}_{\mathrm{n}}\hbox{-}\ {\mathrm{t}}_0\right) = \left(\mathrm{ALT}\ {\mathrm{t}}_{\mathrm{n}}\hbox{-}\ \mathrm{ALT}\ {\mathrm{t}}_0\right)\ /\ {\mathrm{t}}_{\mathrm{n}}\hbox{-}\ {\mathrm{t}}_0 $$
  • $$ \mathrm{Absolute}\ \mathrm{growth}\ \mathrm{rate}\ \mathrm{in}\ \mathrm{diameter}:\kern1.25em \mathrm{AGRD}\ \left({\mathrm{t}}_{\mathrm{n}}\hbox{-}\ {\mathrm{t}}_0\right) = \left(\mathrm{GROS}\ {\mathrm{t}}_{\mathrm{n}}\hbox{-}\ \mathrm{GROS}\ {\mathrm{t}}_0\right)\ /\ {\mathrm{t}}_{\mathrm{n}}\hbox{-}\ {\mathrm{t}}_0 $$
  • $$ \mathrm{Relative}\ \mathrm{growth}\ \mathrm{rate}\ \mathrm{in}\ \mathrm{height}:\kern1em \mathrm{RGRH}\ \left({\mathrm{t}}_{\mathrm{n}}\hbox{-}\ {\mathrm{t}}_0\right) = \left( \ln\ \mathrm{ALT}\ {\mathrm{t}}_{\mathrm{n}}\hbox{-}\ \ln\ \mathrm{ALT}\ {\mathrm{t}}_0\right)/\ {\mathrm{t}}_{\mathrm{n}}\hbox{-}\ {\mathrm{t}}_0 $$
  • $$ \mathrm{Relative}\ \mathrm{growth}\ \mathrm{rate}\ \mathrm{in}\ \mathrm{diameter}:\kern1em \mathrm{RGRD}\ \left({\mathrm{t}}_{\mathrm{n}}\hbox{-}\ {\mathrm{t}}_0\right) = \left( \ln\ \mathrm{GROS}\ {\mathrm{t}}_{\mathrm{n}},\ \hbox{-} \ln\ \mathrm{GROS}\ {\mathrm{t}}_0\right)\ /\ {\mathrm{t}}_{\mathrm{n}}\hbox{-}\ {\mathrm{t}}_0 $$

Analysis

Negative growth has been counted as null for the calculations of the growth indices. Normality analysis was performed by using the Kolmogorov-Smirnov test (Sokal and Rohlf 2012). To compare growth between species, an analysis of variance (ANOVA) was performed using the minimal-significant-differences test (DMS) when the distribution of residuals was distributed normally (Norušis 2012). In order to understand temporal changes in growth rate, relative growth in height and diameter were compared over two periods: 1994–2004 and 1994–2008. Pearson correlation analysis showed high correlation between absolute and relative growth rate, indicating similar tendencies. So we present only the results for relative growth in the three plots.

Survival accumulation rate was calculated as the percentage of surviving plants in each sampling, taking the first sampling as the baseline. Cumulative survival rates were calculated by the Kaplan-Meier method , and survival curves were compared using the log-rank test (Kleinbaum 1996).

Results

Sapling Survival

In the Subcantabrian Mesophytic Oak Forest, the survival rate was low (25.2 %) if we consider the total number of individuals planted (Table 3). Nevertheless, some species were more successful, approaching 50 % survival, such as Quercus robur (QURO, 48.3 %), Corylus avellana (COAV, 48.3 %) and Crataegus monogyna (CRMO, 40 %). On the other hand, other species presented a low or very low survival rate, and Ligustrum vulgare and Euonymus europaeus did not survive at all. The survival diagram (Fig. 3) shows three separate groups. The first group involves the three species (QURO, COAV and CRMO) with over 80 % survival over the first 10 years. These species suffered high mortality during the last 5 years due mainly to routine maintenance activities, such as clipping and clearing, which eliminated some of the saplings. This group shows significant differences relative to the other species planted in the plot, except Corylus avellana and Ilex aquifolium . The second group is formed by Acer campestre , Ilex aquifolium, Fraxinus excelsior and Prunus spinosa . These had high mortality between 5 and 10 years after planting, with around 50 % survival, and then suffered high mortality again during the last 5 years. Finally, there is a third group formed by two species, Ligustrum vulgare and Euonymus europaeus, which had very low survival (less than 10 %) after the first 5 years and did not survive at all by the end of the 15 years. Average seedling life extended from 54.95 months for Euonymus europaeus to 167.6 months for Crataegus monogyna (Table 4).

Table 3 Total number of seedlings of each species planted and percentage of survival in the Subcantabric Mesophytic Oak Forest
Full size table
Fig. 3
figure 3

Accumulated survival curves for each of the species planted in the Subcantabric Mesophytic Oak Forest, the dotted line indicates 50 % survival

Full size image
Table 4 Mean survival times (months) and log rank test results for the planted species in the Subcantabric Mesophytic Oak Forest
Full size table

A good survival rate (59.3 %) was observed in the Eutrophic Montane Beech Forest. Some species showed a survival rate of ca. 90 %, i.e., Crataegus monogyna (91 %), Salix atrocinerea (87.93 %) and Corylus avellana (86.6 %). The rest had a survival rate over 50 % except for Sorbus aria , with only 24.6 % (Table 5).

Table 5 Total number of seedlings of each species planted and percentage of survival in the Eutrophic Montane Beech Forest type
Full size table

The survival-curve analysis (Fig. 4) showed three groups of species with significant differences among them. The first group is that with the highest survival during the monitoring time, including Crataegus monogyna , Salix atrocinerea , Corylus avellana and Rosa squarrosa (ROSQ). These species had a slight decline during the period 2004–2008, especially Rosa. The second group showed gradual mortality through the entire period, with over 50 % survival except for Prunus spinosa (36 %); it includes Prunus spinosa, Populus tremula , Fraxinus excelsior and Fagus sylvatica . The third group involves only Sorbus aria , which suffered high mortality during the second year after planting. Average seedling life extended from 83.36 months for Sorbus aria to 172.4 months for Rosa squarrosa (Table 6).

Fig. 4
figure 4

Accumulated survival curves for the species planted in the Eutrophic Montane Beech Forest type, the dotted line indicates 50 % survival

Full size image
Table 6 Mean survival times (months) and log rank test results for the planted species in the Eutrophic Montane Beech Forest type
Full size table

The Oligotrophic Montane Beech Forest showed very low survival in general (17.5 %), and only one species, Sorbus aucuparia (77 %), had over 50 % survival. The next highest survival rate was that of Sorbus aria at 35.7 %. Other species, such as Betula celtiberica (17.9 %), Ilex aquifolium (15 %), Fagus sylvatica (8.60 %) and Salix atrocinerea (1.90 %) presented very low survival rates. Crataegus monogyna and Salix caprea did not survive at all (Table 7).

Table 7 Total number of seedlings of each species planted and percentage of survival in the Oligotrophic Montane Beech Forest type
Full size table

Survival curves (Fig. 5) showed that all the planted species suffered high, gradual mortality throughout the monitoring period and that most seedlings, except Sorbus aucuparia , died during the first 5 years. Three groups with significant differences between them can be distinguished. The first group was formed by only Sorbus aucuparia, with the highest survival scores and significant differences from all the other taxa. The second group was formed by Sorbus aria (SOAR), Betula celtiberica , llex aquifolium, Fagus sylvatica and Salix atrocinerea ; in this group, llex aquifolium deviates slightly because it is the only species presenting significant differences from all the other taxa. The third group was formed by Crataegus monogyna and Salix caprea , with no survival. Average seedling life extended from 44.07 months for Salix caprea to 170 months for Sorbus aucuparia (Table 8).

Fig. 5
figure 5

Accumulated survival curves for the species planted in the Oligotrophic Montane Beech Forest type, the dotted line indicates 50 % survival

Full size image
Table 8 Mean survival times (months) and log rank test results for the planted species in the Oligotrophic Montane Beech Forest type
Full size table

Sapling Growth

Plant growth was analyzed by combining all the data for height and diameter increases by the different species for the two periods monitored (1994–2008 and 1994–2004). The standard deviation was very high, due to the high variation in growth rate among individuals (Figs. 6, 7 and 8).

Fig. 6
figure 6

Mean and standard deviation for (a) relative growth rate in height and (b) relative growth rate in diameter in Subcantabric Mesophytic Oak Forest type. Red bars: 1994–2004 and black bars: 1994–2008

Full size image
Fig. 7
figure 7

Mean and standard deviation for (a) relative growth rate in height and (b) relative growth rate in diameter in Eutrophic Montane Beech Forest type. Red bars: 1994–2004 and black bars: 1994–2008

Full size image
Fig. 8
figure 8

Mean and standard deviation for (a) relative growth rate in height and (b) relative growth rate in diameter in Oligotrophic Montane Beech Forest type. Red bars: 1994–2004 and black bars: 1994–2008

Full size image

Subcantabrian Mesophytic Oak Forest Type (Crataego laevigatae-Querctum roboris)

In all, more than 75 % of the surviving plants had positive growth. Quercus robur had the best result, with almost all surviving seedlings showing positive growth (97.4 %); this was followed by Crataegus monogyna (83.3 %). The relative rate of height growth was greater for the 15 years than for the first period in all species, but the opposite was true for diameter growth , except for Corylus avellana .

Over the 15 years of monitoring, the species with the highest growth rate in height was Quercus robur (RGRH1994–2004 = 0.24 and RGRH1994–2008 = 0.28) and Acer campestre had the second fastest growth per year. On the other hand, the lowest rates of height growth were by Corylus avellana (RGRH1994–2004 = 0.11 and RGRH1994–2008 = 0.11), followed by Crataegus monogyna and Prunus spinosa (Fig. 5a). Quercus robur (RGRD1994–2004 = 0.24 and RGRD1994–2008 = 0.22) and Acer campestre were the two species with the fastest relative rates of diameter growth , while Prunus spinosa (RGRD1994–2004 = 0.15 and RGRD1994–2008 = 0.10) was the slowest, followed by Fraxinus excelsior (Table 9, Fig. 6b).

Table 9 Results of Fisher LSD post-hoc tests for ANOVA between relative growth rate for planted species in Subcantabric Mesophytic Oak Forest type
Full size table

Eutrophic Montane Beech Forest (Carici sylvaticae-Fagetum sylvaticae)

In spite of having the highest survival rates (59.3 %), only 54 % of the surviving seedlings in this forest type had positive growth. The highest growth percentages were by Populus tremula (87.21 %) and Salix atrocinerea (80.39 %).

Relative rates of height and diameter growth were higher in all species during 1994–2004 than during 1994–2008. This suggests a tendency for growth to slow down during later years.

The highest rate of height growth over the whole 15 years of monitoring was by Fagus sylvatica (RGRH1994–2004 = 0.32 and RGRH1994–2008 = 0.26), followed by Populus tremula and Salix atrocinerea ; the lowest rate corresponded was by Corylus avellana (RGRH1994–2004 = 0.18 and RGRH1994–2008 = 0.15) (Fig. 7a).

For relative rate of diameter growth , Populus tremula (RGRD1994–2004 = 0.29 and RGRD1994–2008 = 0.22) had the highest values, followed by Fagus sylvatica and Salix atrocinerea (Fig. 6b); Corylus avellana (RGRD1994–2004 = 0.16 and RGRD1994–2008 = 0.13) had the slowest growth. Differences in growth between most of the species pairs declined over time, not only for relative height growth but also for relative diameter growth (Table 10).

Table 10 Results of Fisher LSD post-hoc tests for ANOVA between relative growth rate for planted species in Eutrophic Montane Beech Forest type
Full size table

Oligotrophic Montane Beech Forest Type (Saxifrago hirsutae-Fagetum sylvaticae)

Although survival was very low (17.5 %), the proportion of seedlings with positive growth reached 67.8 % and Betula celtiberica stood out with 100 % survival, followed by Sorbus aucuparia with 72.3 % (Table 7).

Relative rates of height and diameter growth were higher in the period 1994–2004 in all the species, indicating a slowing-down in their growth rate.

Over the 15 years of monitoring, the species with the highest rates of height and diameter growth was Betula celtiberica (RGRH1994–2004 = 0.32; RGRH1994–2008 = 0.26; RGRD1994–2004 = 0.30 and RGRD1994–2008 = 0.23), followed by Fagus sylvatica ; lower rates were showed by Sorbus aucuparia (RGRH1994–2004 = 0.18; RGRH1994–2008 = 0.14; RGRD1994–2004 = 0.15 and RGRD1994–2008 = 0.13) (Fig. 8a, b). Differences in growth between all the species pairs declined with time, not only for relative height growth but also for relative diameter growth (Table 11).

Table 11 Results of Fisher LSD post-hoc tests for ANOVA between relative growth rate for planted species in Oligotrophic Montane Beech Forest type
Full size table

Discussion

Seeding survival can be attributed to the combined effects of human disturbance and microclimatic conditions (Ramírez-Marcial 2003) and also to the quality of the seedlings, in terms of their survival capacity (Grossnickle 2012). The high survival (almost 60 %) among seedlings in the Eutrophic Montane Beech Forest (EMBF), as compared to the Oligotrophic Montane Beech Forest (OMBF, less than 20 %), is related to soil and climatic conditions where the plots are located. The Oligotrophic beech forest , besides having poorer nutrient availability, is located on a steeper slope that experiences high runoff , thus captures less water, and consequently is drier. In addition, the sandstone substrate facilitates water loss and the southeast orientation confers greater summer drought . The better soil conditions in the Eutrophic beech plot and its northwest orientation, which promotes higher humidity especially in summer, has led to better plant survival. On the other hand, the poor quality of the seedlings of several species, such as Sorbus aria , Ligustrum vulgare , Euonymus europaeus and Prunus spinosa , caused high mortality during the first years (Grossnickle 2012). Another factor was the action of herbivores that affected some species, such as Ilex aquifolium .

Already in 2004 it was observed that a high proportion of the trees showed very little growth; by 2008 high mortality was recorded. Root-system size is proportional to the diameters of the above-ground main stem (Mexal and Landis 1990), and the ability to grow roots has a significant effect on survival (Grossnickle 2012). Thus, the low growth rate could indicate a small root system that leads to plant death. Moreover, the routine maintenance work conducted in the plots increased the mortality, especially in the Subcantabrian Mesophytic Oak Forest, where about 30 % of the seedlings were cut.

There was some variation in survival between different functional groups . In general, shrubby species showed better survival than species of mature stage s in later succession. Thus, the hypothesis that species of seral stages (secondary forests and shrubland ) have higher survival rate than mature-stage species seems to be confirmed. This is to be expected for beech, since it is well known that it has lower survival rates under high light intensity (Petritan et al. 2007; Jarcuska 2009). This was not the case with oak ( Quercus robur ), however, which had a survival rate similar to that of Corylus avellana or Crataegus monogyna .

Relative rate of height growth has been greater in the last 5 years than in the previous period in mesic oak forest (SMOF), but in the beech forest plots growth in the first 10 years was greater than in the last period. Even the species planted in both the mesic oak forest and eutrophic beech forest , such as Corylus avellana , Crataegus monogyna and Fraxinus excelsior , showed this tendency. This suggests that greater water availability in deeper soil layers of the flat ground, where the plots are located, could induce more growth in the last period, once the tree roots had reached these more moist levels. The different relative growth rates of Fagus sylvatica in the eutrophic versus oligotrophic beech forest can be explained by the worse soil and microclimatic conditions at the Oligotrophic Montane Beech Forest plot.

There are also some other differences between the functional groups of plants. Trees of mature and secondary stages ( Quercus robur , Fagus sylvatica , Populus tremula and Acer campestre ) tend to show greater rates of height and diameter growth than do shrubs of successional stages ( Crataegus monogyna , Prunus spinosa and Corylus avellana ). The higher growth rate of the mature-forest species is a phenomenon that has also been observed in tropical forests by Shono et al. (2006), where a number of primary forest species had better growth rates than did secondary-forest species.

In summary, the results have proved that the initial topographic and soil conditions of the plots determine the afforestation success to a great extent, which was already evident and reported by the second year of monitoring (Sesma et al. 1997). The Eutrophic Montane Beech Forest plot, with more structured soil and humid conditions, especially in summer, had better survival and growth rates, probably due to the orientation of the plot. On the other hand, the steeper slope and sandy soil of the Oligotrophic Montane Beech Forest plot, together with its SW orientation, resulted in worse conditions for seedling establishment , increasing mortality and reducing growth rates. In fact, water stress related to high temperatures and lower moisture availability in pastures was reported as one of the main factors causing mortality in several reforestation attempts (Hooper et al. 2002, Doñobeytia et al. 2007). Finally, in the Mesic Oak Forest plot, there were some areas with very compacted soil due to the presence of paths; this resulted in high plant mortality in the first years. Another factor that caused a remarkable decline in the survival of all species was the weeding and clearing work carried out in the plots, especially in the oak forest . Although weeding has been documented to have a strong positive effect on seedling and sapling growth (Löf and Welander 2004), it must be done by hand or very carefully, in order to avoid cutting the plants.

Despite low overall survival, the general growth rate of the surviving seedlings allows us to conclude that the study results were positive, accelerating the natural forest successional processes and controlling invasion by shrub species like Rubus and Ulex , which limit natural succession in degraded lands (cf Shono et al. 2006). Tree-planting strategies should include improving the soil conditions before afforestation and selecting seedlings with good root systems that contribute to better survival. With regards to species selection, even though growth rates were higher in trees than in shrubs, we recommend using shrubs also in order to improve the biodiversity of the restored forests.