Natural Forests of Oak in NW Spain: Soil Fertility and Main Edaphic Properties

The species of Quercus genus dominates a large part of the temperate forests of the Northern hemisphere and in subtropical transition areas. At present, about 400 species are recognized, most of which occur in Mexico and extend all over North America, Europe and most of Asia (Díaz-Maroto et al., 2005). In the Northwest Spain, the climax vegetation that currently covers the largest area is the broad-leaved forest characterized by various oak species (Buide et al., 1998). According to different studies (palynological and dendrochronological studies, principally) these forests were established in the NW Iberian Peninsula between five and seven thousand years ago, after the last glaciation in Quaternary (Guitián, 1995). Historical factors, site conditions, and requirements of the species give rise to different forest types with various floristic compositions and structures (Peterken & Game, 1984). All the forests belong to Querco-Fagetea class (Atlantic Province) with the maximum possible degree of evolution, and they would remain in the current state if environmental conditions did not change (Rivas-Martínez, 1987; Rivas-Martínez et al., 2001).


Introduction
The species of Quercus genus dominates a large part of the temperate forests of the Northern hemisphere and in subtropical transition areas. At present, about 400 species are recognized, most of which occur in Mexico and extend all over North America, Europe and most of Asia . In the Northwest Spain, the climax vegetation that currently covers the largest area is the broad-leaved forest characterized by various oak species (Buide et al., 1998). According to different studies (palynological and dendrochronological studies, principally) these forests were established in the NW Iberian Peninsula between five and seven thousand years ago, after the last glaciation in Quaternary (Guitián, 1995). Historical factors, site conditions, and requirements of the species give rise to different forest types with various floristic compositions and structures (Peterken & Game, 1984). All the forests belong to Querco-Fagetea class (Atlantic Province) with the maximum possible degree of evolution, and they would remain in the current state if environmental conditions did not change (Rivas-Martínez, 1987;Rivas-Martínez et al., 2001).
The forests correspond to the following phytosociological associations (Rivas-Martínez, 1987;Díaz & Fernández, 1994;Rivas-Martínez et al., 2001)  This area has a complicated topography, an average altitude higher than 550 meters, and slopes steeper than 20% in half of the surface area. It is a territory with a rich and different lithological composition .
The climate is quite varied, but follows a general oceanic pattern with annual rainfall from 600 to more than 3000 mm (Carballeira et al., 1983). The study area was defined and the sampling zones were chosen based on the Forest Map of Spain (Ruiz de la Torre, 1991).

Experimental design and analysis techniques used
The number of sampling points was different for each species, but in all cases, we studied the physiographic features (altitude, slope, and orientation), general soil characteristics (parent material, depth, and description of horizons), soil profile, and physicochemical properties (Food and Agriculture Organization [FAO], 1977). To enable description of the soil profiles, a pit approximately 2 m long and 80 cm wide was dug at each site. The horizons were differentiated according to color and/or texture. One sample of each horizon (about 2 kg) was collected, and the following variables were analyzed (FAO, 1977;Klute, 1986): Munsell color, moisture (by oven drying to constant weight at 105 ºC), pH (potentiometrically in water at 1:2.5 soil-solution ratio), total nitrogen (N) (semi-micro-

6
Kjeldahl method), carbon (Walkley and Black method), organic matter (OM) (multiplying the percent of C by a factor of 1.724), C/N ratio, exchangeable cations [by extraction with 1 N ammonium acetate at pH 7 and quantified by atomic absorption (Ca and Mg) and by emission (K) using a Perkin-Elmer 603 spectrophotometer, Waltham, Mass.] and available P by the Bray and Kurtz method (Bray-1 P).
At each sampling point, soil different parameters were estimated. The parameter total value in the whole soil profile was determined by using a standard weighted mean for Sand (S), Silt (Si), and Clay (Cl). For the rest of studied parameters, the weighted mean was considered by applying the method of Russell & Moore (1968). In the case of surface values, the information from the upper 20 cm of the soil was considered, except when more than one horizon was present at this depth. In that case, it was calculated as a weighted mean .

Statistical analysis
Univariate statistics were calculated (Walpole et al., 1999) and the following values were presented for the measured soil parameters: lower limit (LL), parameter minimum value; lower threshold (LT) or 10th percentile; mean (M); upper threshold (UT) or 90th percentile; upper limit (UL), parameter maximum value. Based on these values, the optimal edaphic habitat of the oak forests, in relation to a parameter, can be defined as the interval between UT and LT. This interval is composed by 80% of the sampling points, excluding the areas where the parameter showed the highest and the lowest values, marginal habitats . The optimal habitat defines the most suitable soil conditions for a particular species, whereas, in the marginal is the condition where some parameter is not adequate so the area has questionable suitability for oak growth (Hardtle et al., 2005).
Normality of the soil parameters was evaluated (Walpole et al., 1999), the result allowed a bivariable correlation analysis by Pearson (normal) or by Spearman's rank (Rozados et al., 2000;Sas Institute Inc. [SAS], 2004) to be applied. When soil parameters were significantly (p<0.05) correlated, possible linear relationships between the parameters was studied.

Physiographic characteristics of the stands
Most of the oak stands are located in zones with a steep slope or at the thalwegs, due to historical anthropogenic activities in these forests, at low and medium altitude with variable orientations, but shady sites are dominated in Q. robur and Q. petraea, while sunny sites in Q. pyrenaica stands . Some stands of Q. robur are located in continental influence zones that are relatively distant from their suitable habitat. In these cases, this species is very easily hybridized, with Q. petraea and Q. pyrenaica , thus it is complicated to distinguish between these three species in such zones (Rivas-Martínez et al., 2001).

Parent rock and soil type
The geological material is heterogeneous with rocks from Precambrian and Paleozoic ages, tertiary and quaternary sediments (Ruiz de la Torre, 1991). Siliceous substrates were present in most of the stands, with dominance of granite and schist . Soil depth is elevated, and the presence of loose, fresh, deep soils is especially suitable for Quercus spp., in general, and above all for Q. robur, because these species has an extensive root system that penetrates very deep into soil (Göransson et al., 2006).

Quercus robur soils
The main properties of the soils under Q. robur forests in NW Spain were characterized. Eleven edaphic parameters (pH, OM, N, C/N ratio, P, K, Ca, Mg, S, Si and Cl) were estimated in 39 soil samples (Table 1).
The parameters that reflect the chemical and biological soil properties (pH, OM, N, and C/N ratio) show average values similar to the values reported as adequate by Castroviejo (1988) and Rozados et al. (2000) for these formations, but, the values obtained for OM are an exception because they are slightly higher than optimal, its average value is equal to 8.64 ± 5.19. According to the acidity scale suggested by Wilde (1946), most soils are to strongly acidic (4.7 ≤ pH < 5.5). Some of them present a value lower than 4 (extremely acidic) or higher than 6 (moderately acidic).
However, the studied soils show satisfactory humification, which is revealed by high values of the C/N ratio (Mansson & Falkengren-Grerup 2003;Hagen-Thorn et al. 2004). The pH of the upper horizons is the most relevant pH in the biological processes (Nornberg et al., 1993). This is because it depends on the residues of the dominant species in the vegetal cover and because the first roots of young plants are developed in the upper layer. The leaching due to rainfall is very important, because it can move dissolved salts down through the soil and cause an overall increase in pH at depth. Therefore, the analysis of the data obtained in the topsoil, the upper 20 cm of the soil, would mean a low mineralization rate due to the relatively high OM content, the low pH, and the C/N ratio near 15 (Nornberg et al., 1993;Neirynck et al., 2000;Landgraf et al., 2006) (Table 1).
Nutrient concentration is considered in the low or medium category (for instance, the exchangeable K concentration is only 73 mg kg -1 ) as compared with other forests in the study area, except the Bray-P concentration, which is considerably higher (21.8 mg kg -1 versus 1-6 mg kg -1 ) (Castroviejo, 1988;Rozados et al, 2000).

www.intechopen.com
Soil Fertility Improvement and Integrated Nutrient Management -A Global Perspective 8 The physical characteristics values are close to those reported by Castroviejo (1988) ( Table 1). These data can be influenced by the steep slope on which the stands are located. Moreover, slight erosion occurs despite the vegetation cover and the soil protection by the extensive root system of Q. robur. This erosion leads to a gradual loss of the fine fractions, which are dragged downhill. The particle-size distribution of the soil determines the physical properties and influences the soil acidity ( Van de Moortel et al., 1998).

Parameter
Average

Quercus petraea soils
Soils under natural forests of Q. petraea in northwest Spain were characterized in an inventory of 52 stands in which the chemical properties and nutrient contents were determined by analyzing ten edaphic parameters (Table 2). Siliceous substrates (mainly slates) of high pH predominated in all stands. Umbric Regosols occurred in almost 80% of the stands, indicating that the soils were less well developed than those under Q. robur. The C/N ratio was particularly high and the OM content was less than 10 mg kg -1 .
In general, the soils are very acidic, with pH values that present little variation. The high rainfall in a large part of study area accentuates the acid character of the soils.
Although the mean value of the C/N ratio was close to 18, the low pH values do not allow adequate humification conditions (Mansson & Falkengren-Grerup, 2003). The rate of mineralization may be considered as slow, as analysis of the data corresponding to the topsoil shows that the mean content of OM is 10.09, the pH is 4.59 and the C/N ratio 17.92.
The concentrations of all macronutrients, except phosphorus, were higher than in the soils under pedunculate oak, and those of phosphorus and calcium were the most variable. The contents of base cations, fine earth (FE) and total gravel (GRA) were high in the optimal habitat (West of Galicia, East of Asturias and NW León).

Quercus pyrenaica soils
The soils where natural stands of "rebollo" oak (Quercus pyrenaica Willd.) occur in NW Iberian Peninsula were characterized in an inventory of 40 different stands; the main properties of the soils were also determined by analyzing ten different edaphic parameters (Table 3).
The substrates are principally siliceous, with schist and quartzite predominating; the dominant textural composition is sandy-loam with small pockets of clay-loam and sandy textures. Soils are mainly the Umbric Regosol-type (about 60% of soils), although other soil types like Cambisols are found in areas with a soil depth greater than 80 cm and, normally located in thalwegs.
The values of the parameters that reflect the chemical properties of the soil, i.e., the pH, organic matter, nitrogen, and C/N ratio, are similar to those reported as suitable for Quercus spp. forests, in general, to the study area. However, the concentrations of macronutrients, except for phosphorous, were higher than in soils under Q. robur (Table 3) .  Table 3. Average, standard deviation (±SD), coefficient of variation (CV), and the range of values (Max-Min) of soil parameters (n=40) in natural forests of Q. pyrenaica in NW Spain

Characterization of edaphic habitat of Quercus robur forests
The optimal habitat of Q. robur forests, which was considered statistically to fall between the 10th and 90th percentiles (80% of plots ≈ sampling points), allows determination of the most suitable ecological range, and may be considered as the minimum potential area of the species in the study area. Marginal habitats are defined as those between the limits of the optimal habitat and the absolute extremes; they include 20% of plots and provide an estimation of the species adjustment to site conditions according to parameters values that remain outside the optimal habitat, since not all of them have the same level of significance as biotope descriptors (Fig. 2)  .
In Q. robur stands, 80% of soils are Cambisols, so we calculated the optimal and marginal edaphic habitats for this soil-type only (Fig.2).   The wideness of the upper marginal intervals (upper threshold and limit) for Ca and SCa in the Cambisols edaphic habitat (Fig. 2), suggests that these parameters are not good indicators of the upper limit of optimal habitat : - The soils located within the optimal edaphic habitat ( Fig. 2) are mainly strongly acidic, with an average pH value equal to 4.91 for the whole profile, and equal to 4.75 for the topsoil. Within the marginal habitat, there are moderately acidic soils, with pH higher than 5.5 (Wilde, 1946). OM content is variable, with a minimum that does not reach 2 mg kg -1 for the whole profile and corresponds to the lower threshold of the edaphic habitat (Fig. 2), although the average is 7.58 mg kg -1 for the whole profile and 11.53 mg kg -1 for the topsoil, which characterizes a mull-moder humus (Camps et al., 2004). -Optimal edaphic habitat shows low content of exchangeable cations (Fig. 2). However, the P has a considerably high value, with an average of 24 mg kg -1 for the whole profile. - The soils have an elevated percentage of sand with an optimal habitat that ranges between 49-82 mg kg -1 and an average value of 67 mg kg -1 . Conversely, the clay proportion reaches a maximum of nearly 23 mg kg -1 , with an average of 11 mg kg -1 , because the substrates overlie parent material that releases a greater proportion of sand than clay during weathering (Castroviejo, 1988).

Characterization of edaphic habitat of Quercus pyrenaica forests
The same methodology was followed to distinguish the optimal and marginal edaphic habitats of Quercus pyrenaica forests in NW Spain. In this case, as about 60% of soils are Umbric Regosols. In the Fig. 3, we only calculated the optimal and marginal edaphic habitats to this soil-type. Also, it can be seen that certain edaphic parameters are not suitable for fixing the upper value of the optimal habitat, given the wide range of upper marginal values: OM, total and surface N, total and surface Mn and surface Ni as, to a lesser extent, total and surface P. The most notable characteristics were as follows (Fig. 3): - The pH values of the optimal habitat ranged between 4.89 and 5.38, corresponding to strongly to moderately acidic soils (Wilde, 1946). In marginal habitats soil pH was less than 4.30 or greater than 6.13 (Fig. 3). - The OM content was low and for optimal habitats ranged between 1.39-6.06 mg kg -1 for the total profile and between 1.89-7.72 mg kg -1 for the upper 20 cm (Fig. 3). The optimal values of the C/N varied widely and indicate adequate humification (Mansson & Falkengren-Grerup, 2003) and a normal rate of mineralization. - The optimal habitat showed higher contents of exchangeable base cations than in the soils under Quercus robur within the study area. However, the concentration of P was lower, probably because the stands correspond to mature forests that had not undergone changes in land use . -Wide suitable ranges of fine earth and total gravel contents were also apparent.

Characterization of edaphic habitat of Quercus petraea forests
Finally, the same methodology was followed to distinguish the optimal and marginal edaphic habitats of Q. petraea forests in NW Spain. In this species, more than 75% of the stands also occur on Umbric Regosols. Then, we analized the optimal and marginal edaphic habitats to this soil-type (Fig. 4). Parameters related to Ca concentration are not suitable for determining the edaphic habitat of sessile oak in northwest Spain, given the wide range of upper marginal values (Fig. 4)  The total pH in H 2 O (WPH) values of the optimal habitat ranged between 4.28 and 5.14, corresponding to moderately or even extremely acidic soils (Wilde, 1946). In marginal habitats soil pH was less than 4.23 or greater than 5.65 (Fig. 4). These values varied little and, principally, the surface pH in H 2 O (SWPH), partly due to a high rainfall that does not allow adequate humification. - The OM matter content was intermediate (Hagen-Thorn et al., 2004) and for the optimal habitat ranged between 3.04 and 12.60 mg kg -1 for the total profile, and between 4.40 and 17.02 mg kg -1 for the topsoil (Fig. 4), giving rise to moder type humus with a slow rate of mineralization, in spite of suitable C/N ratio of more than 18 (Mansson & Falkengren-Grerup, 2003). - The optimal habitat showed higher contents of base cations, except P, than in Q. robur soils ). - The concentration of the micronutrient Mn varied widely (4.92-79.80 mg kg -1 ). Copper was the micronutrient present at lowest concentrations, with an average value of less than 0.5 mg kg -1 (Fig. 4). - The total fine earth (FE) and total gravel (GRA) contents also varied widely, 29-59 mg kg -1 and 41-72 mg kg -1 , respectively.

Conclusions
Most of the studied soils correspond to stands located in areas with a complicated topography, a medium altitude and variable orientation. The substrates are mainly siliceous, with sandy-loam or loamy textures dominating. Soil depth is greater than 100 cm in more than half of the stands, although some stands can also develop on rocky substrates, mainly granite and schist.
Within the study area, Quercus robur is located on more developed soils than the other two species, principally, Dystric cambisol, with an A/Bw/C profile. Q. pyrenaica occurs predominately on Regosols and Q. petraea is present on poorly developed soils, with more than 75% of stands occurring on Umbric Regosols with an "A" horizon that is usually no deeper than 50 cm.
The chemical parameters of Q. robur soils (pH, OM, N and C/N ratio), show average values similar to those considered optimal for this species, except for the OM, which had higher values. They are strongly acidic (4.7 ≤ pH < 5.5), however, they have a good humification. The macronutrient content can be considered low or medium, except for P, since soils develop on substrates that are poor in base cations.
Organic matter, pH and clay tended to be significantly (p<0.05) correlated with other soil parameters, and with each other. A positive linear relationship (p<0.001) was noted between pH and Bray-P concentration, suggesting that many oak forests are located on abandoned agricultural land. In these stands, the high P content increases with the value of pH. Clay content shows significant differences for surface and total OM and nitrogen.
The analysis of edaphic habitat by soil type does not differ greatly from the results of the stands considered as a whole. Q. robur forests show a tendency to occur on the best soils, cambisol type, and not on poorer soils, where the occurrence of this species is often due to anthropogenic pressure.
In Quercus pyrenaica soils, the pH values were quite homogeneous, giving rise to very acidic soils, although these were not conducive conditions for humification, as verified by the C/N ratios and the presence of moder or mull type humus. Mineralization occurs at a normal rate and the chemical-nutritive characteristics are suitable for "rebollo" oak forests. Although these forests are developed on parent rock, which are poor in base cations, all studied macronutrients, except P, were present at higher concentrations than in soils under Q. robur, probably because the stands correspond to mature forests that had not undergone changes in land use.
The statistical analyses reveal that Ca and Mg contribute to the greater number of significant relationships. The influence of the physical properties (fine earth and gravel) has a smaller importance.
Finally, in Q. petraea soils, the pH values were also quite homogeneous, giving rise to very acidic soils, although, in this case, this did not allow adequate conditions for humification too, giving rise to moder type humus. Mineralization occurs slowly and the biologicalchemical characteristics are suitable for sessile oak forests. Although these soils are developed on substrates that are poor in base cations, all macronutrients except P were also present at higher concentrations than in soils under Q. robur, probably because no changes in soil use have taken place, and also because of the loamy textures of the substrates.
Analysis of the edaphic habitat reveals that most highly developed forest soils, within the study area, correspond mainly with Quercus robur, and even Q. pyrenaica. The optimal habitat in Q. petraea stands shows a wide range of values for base cations, pH and physical parameters (FE and GRA).