Carbon and nutrients stocks in even-aged maritime pine stands from Portugal

To comply with the demands of the Kyoto Protocol, industrialized countries must reliably estimate the stored carbon (C) in different pools of forest ecosystems. The main objective of this study was to quantify the biomass, C and nutrients stocks in the forest floor, understory, downed dead wood (DDW) and mineral soil of even-aged maritime pine (Pinus pinaster Ait.) stands from three contrasting regions of Portugal. Assessing the specific contribution of DDW to the C and nutrients stocks and how C concentration in the plant material differs from 0.50, the reference value used in many practical applications, were also objectives of this research. Biomass content of forest floor was determined by a quadrat method. Sampling units of 1 m were used for the understory. The line intersection method was adopted for sampling DDW and the mineral soil was sampled at three depths. Concentrations of C and nutrients were obtained by chemical analysis of samples from soil and milled plant material. Biomass and C in the trees were obtained using published equations. Total C stocks ranged between 100.6 Mg ha and 308.6 Mg ha. Mineral soil shared up to 70-74% of global stored C, being the main cause of the global C stock differences among regions. Phosphorous and potassium were at low to very low levels in the mineral soil and plant material. The contribution of DDW to the C and nutrients pools was negligible. The percentage of C in the plant material ranged between 52% and 54%.


Introduction
Maintaining and enhancing productivity of forests and their contribution to global carbon cycle is an important component of sustainable forest management.This is true not only for timber production but also for other wood and non-wood forest products and services as it is the case of carbon sequestration, one of the most important services that are expected from the forests nowadays.Carbon dioxide (CO 2 ) is the most relevant anthropogenic greenhouse gas (GHG).According to the 2007 report from the Intergovernmental Panel for Climatic Changes (IPCC, 2007), CO 2 annual emissions increased from 21 to 38 × 10 9 Mg (Gt) between 1970 and 2004.Moreover it is believed that CO 2 is highly responsible for the changes in the world climate that are being observed.Kyoto Protocol demands for transparent and verifiable methods to estimate the amount of carbon (C) stored in different pools of forest ecosystems of industrialised countries and its changes in time (articles 3.3 and 3.4) (e.g., Bravo et al., 2007;LeMay and Kurz, 2008).Reliable estimates of C sequestration at the regional and national levels can help Governments to evaluate their success at meeting international agreements on climate and pollution.
About 46% of the total C in terrestrial ecosystems is in forests (Fujimori, 2001).In comparison with other terrestrial biomes, forests have the greatest capacity for storing and cycling carbon (Barford and Wofsy, 2001).The main C pools in a forest ecosystem are the biomass of living trees, understory vegetation, dead wood, forest floor and mineral soil (Waldendorp et al., 2002).It is important to quantify the C stock in each compartment, either by direct measurement or by estimation using models.The interest in the measurement and modelling of C storage in forests greatly increased in the last two decades.
In this study the C stocks in the trees, understory, forest floor, dead wood and mineral soil were quantif ied for even-aged stands of maritime pine (Pinus pinaster Ait.) in Portugal.Despite the main emphasis was on C stocks, nutrients stocks were also quantified, complementing and improving the information about the condition of these stands.The maritime pine is the most representative conifer species in Portugal, covering 926 × 10 3 ha (Tomé et al., 2007) which is around 27% of the total forested area in the country.The species is distributed along the Atlantic coast from the basin of the Tagus River, near Lisbon, to the Portuguese border in the north, spreading to inland in the north and centre regions, ascending to altitudes of 700 to 900 m where the influence of the Atlantic Ocean prevails (Oliveira et al., 2000).There are a few studies in Portugal that evaluate the biomass and C stocks in maritime pine stands.However, most of these works are mainly focused on one or few compartments (e.g., Fernandes et al., 2002;Lopes and Aranha, 2006;Martins et al., 2007) and were often complementary to other research objectives.Comprehensive studies of C stocks, accomplishing all the main pools in forests, including downed woody debris and mineral soil are scarce in the maritime pine specific case.Even more scarce are studies assembling information about C and nutrients stocks.
This study had the following specific objectives: (i) to quantify and compare the biomass, the C and nutrients stock in the forest floor, understory, dead wood, and mineral soil of even-aged maritime pine stands from different regions of Portugal; (ii) to analyse the specific contribution of downed dead wood to C and nutrients stock in managed stands and determine the basic density of different decomposition classes, and (iii) to assess how carbon concentration in the plant material differ from the reference value, 0.50, used as a default value in so many practical applications.

Study area
Three contrasting regions of maritime pine distribution area in Portugal were considered in this study (Minho, Tâmega and Sines).In each region, ten circular plots with an area of 500 m 2 were established in evenaged maritime pine stands (Fig. 1) from the end of July to the beginning of September, 2009.Therefore, a total of thirty stands were sampled for measuring C stocks in forest floor, understory vegetation, dead wood, live trees and mineral soil.
The sampling in each region was made with the purpose of including a large range of ages and stand densities.Thus, different combinations of age, density and site index were considered when setting the plots in the field.An effort was made to select similar stand structures between regions, including the thinning regime, which is mainly from below, in order to make comparison more reliable.Stands with uncommon structure of thinning regimes were excluded.
Using available digital information from the Portuguese Environment Agency (APA, 2007), mean values Carbon and nutrients stocks in maritime pine stands of climate variables for the period  were obtained for the sampled plots using ArcGis software (Esri Inc, 2009) to spatially link climate information in a grid with a resolution of 2 × 2 km 2 with plot coordinates (see Table 1).Following the soil nomenclature of the IUSS WRB (2006) classification, haplic umbrisols prevail in Minho as well as in a broad range of areas in Tâmega.Leptic regosols are also present in this last region.Haplic podzols prevail in the Sines region.
The understory vegetation in Minho was dominated by Pteridium aquilinum in many sampled plots sometimes along with Ulex sp.Ulex was present with Erica sp., Caluna vulgaris and Pterospartum tridentatum in other plots.Pterospartum tridentatum was more abundant in Tâmega than in Minho, being present either alone or together with Erica sp., Caluna vulgaris and Ulex sp.In Sines, the understory was dominated by Ulex sp.many times in consociation with Stauracanthus spectabilis (an endemism).Other observed species were Helichrysum picardii, Lavandula pedunculata, Cynara cardunculus, Corema album, Juniperus turbinata, Daphne gnidium and Asparagus acutifolius.

Measurements of trees inside the plots
In every plot, all the trees were measured for diameter at breast height (d), total height (h), crown width (cw), height to the base of live crown (hcb) and height to the level of the largest crown width (h lcw ).Azimuths and distances of each tree to the plot centre were taken,  dP: number of days with precipitation (days year -1 ).dF: number of days with frost (days year -1 ).RUN: runoff (mm year -1 ).P: annual precipitation (mm).RH: relative humidity (%).INS: insolation (hours year -1 ).ETP: evapotranspiration (mm year -1 ).RAD: radiation (kcal cm -2 ).T: temperature (°C).ELEV: elevation (m).SE: standard error.
beginning at the North direction and progressing clockwise.The plot centre coordinates were registered using a Trimble GeoXt GPS.The stand structural variables were computed based in the tree measurements (Table 2).

Measurements on the forest floor
The forest floor biomass was evaluated by a quadrat method, using metallic sampling units of 50 × 50 cm.Four samples were collected at a distance of 3 m from the plot centre in North, South, East and West directions.The forest floor thickness was measured in each sampling point.All the samples collected in a plot were placed inside identified plastic bags, one for the litter layer (L) and another for fermentative (F) and humus (H) horizons together (hereafter FH layer).The litter was further separated into leaves, twigs and cones.Total fresh weights of these components were registered and then sub-samples were oven-dried at 70°C until constant weight.The ratio of dry weight to fresh weight of sub-samples was used to calculate the oven-dried biomass of the global samples.Carbon and nutrients content in the several components of forest floor were obtained by chemical analysis of milled material (1 mm sieve).

Measurements of the understory vegetation
A sampling quadrat of 1 m 2 was set at a point with an understory cover representative of the understory conditions of the 500 m 2 plot, in terms of structure and composition.All the vegetation inside the 1 m 2 area was collected to a plastic bag and then separated by species and the respective fresh weight registered.Subsamples, for each species, were oven-dried at 70°C until constant weight.The material of these individual species was milled (1.5 mm sieve) for chemical analysis.

Measurements of dead wood
Logs (fallen trees, fallen branches, and pieces of stem or branches) were sampled inside the plot area using the line intersect method (Warren and Olsen, 1964;Van Wagner, 1968).Four 25 m length transects were used, going through the plot centre in North-South, East-West, Northeast-Southwest, Northwest-Southeast directions, were used.A length of 50 cm (e.g.Waldendorp et al., 2004) and a diameter of 2.5 cm at both ends (e.g.McCarthy and Bailey, 1994;Ganjegunte et al., 2004) were adopted as minimum dimension for logs.Diameters at the intersection points were measured at right angles with the logs longitudinal axis for volume calculation (m 3 ha -1 ).Each tallied log was classified into one of three decay classes similarly to Davis et al. (2003), equivalent to the five classes classification frequently found in forestry literature related to coarse woody debris (CWD) (e.g.Sollins, 1982 or Siitonen et al., 2000, andtheir adaptations) by coalition of classes II and III as well as IV and V. Samples for each decay class were extracted from three pieces of woody debris whenever present inside the plot area.After volume determination by water displacement (m 3 ), the samples were oven-dried at 103 ± 2°C to obtain mean values of wood basic density in order to convert volume into biomass.Samples from each decay class were also collected inside the plot area and milled (1.5 mm sieve) for chemical analysis.Snags (dead trees over 1.3 m tall and standing at an angle superior to 45°), if present inside the plot area, were measured for d and h and classif ied into one of the three decay classes using the same procedure already referred for the logs.

Procedures in the mineral soil
In the four points where forest floor was collected inside each pot, soil samples were taken, using a corer, at the depths 0-10, 10-30, and 30-60 cm and placed in separate plastic containers.These sampling depth levels were considered in order to match the requirements of the Kyoto Protocol level and to complement the information from two other studies for future model development.The use of corers in forest soils is difficult; therefore it was necessary to dig holes in order to facilitate the corer use to collect samples at deeper levels.Composite samples of about 2 kg from each depth level were used for chemical analysis.A total of 88 samples (instead of 90) were analysed.In two plots it was not possible to collect suitable samples from the 30-60 cm depth because of the low thickness of the prof ile.Undisturbed soil samples were taken from the wall in the opened holes at the three depths using metallic cylinders with 196.35 cm 3 .The bulk density (g cm -3 ) was determined after oven-dry the samples at 105°C during 24 hours, sieve with a 2 mm mesh and adjusted for the mass and volume of rocks (> 2mm size).The amount of organic carbon (C org ) in a layer of mineral soil was obtained by multiplying the proportion of C in this layer to the oven-dried mass of the fine earth fraction.

Biomass and C stock in the live trees
Above ground biomass of the live trees was obtained by using the system of simultaneous equations from Faias (2009), which estimates biomass for stem wood (ww), stem bark (wb), branches (wbr) and leaves (wl).Stem biomass (ws) is the sum of ww and wb and total aboveground biomass (wa) is the sum of the biomass of all the four components.The roots biomass was estimated by employing Montero et al. (2005) equations.The values of C concentration for the different tree components presented in the study of Lopes and Aranha (2006) were adopted to obtain C stock in the trees.

Chemical analysis
In relation to the samples from plant material (forest floor, understory, and dead wood), phosphorous (P), potassium (K), calcium (Ca) and magnesium(Mg) were extracted by dry mineralization at 500 ± 20°C and the analytical determinations in the extract were undertaken by using optical emission spectrophotometry (ICP-OES).Carbon, nitrogen (N) and sulphur (S) were extracted by catalytic pyrolysis and contents were determined by using a LECO CNS elemental analyser.
The samples of mineral soil were first dried at 40°C and then sieved.Chemical analysis were performed on the fine earth fraction (< 2mm).Contents of C org and total N were obtained by dry combustion using a LECO CNS analyser.Available K and P were obtained by ICP-OES after extraction with Égner-Rhiem method using a solution of ammonium lactate (0.1 N) and acetic acid (0.4 N) (Balbino, 1968).Available Mg was obtained by flame atomic absorption spectrophotometry (Flame-AAS) after using neutral 1N ammonium acetate (Warncke and Brown, 1998).Concerning exchangeable bases, available Ca and Mg were obtained by Flame-AAS and available K and Na by flame emission spectrophotometry (FES) after using neutral 1N ammonium acetate (modif ied from Chapman, 1965).Soil pH (H 2 O) was obtained by potenciometric determination in a 1:2.5 soil-to-water (v/v) suspension.

Data analysis
Field data and data from the chemical analysis were processed in order to characterize the carbon and nutrients stocks in the three sites and study the differences between them.Carbon and nutrients stocks for each pool were computed using concentration values in dry biomass obtained from chemical analysis and were condensed in tabular form.One-way analysis of variance (ANOVA) of the carbon and nutrient amounts were performed using region as a three level factor.Analyses of covariance were discarded as the tried covariates (Nt, t and SI) were not significant.If the null hypotheses, no significant effect of the regions, were rejected least significant differences tests were used to compare mean carbon and nutrient stocks values between the regions.Kruskal-Wallis test and multiple comparisons of mean ranks were used when normality assumption was not verified.This analysis was complemented by several graphics, namely absolute and percent values of C stock in the different C pools, as well as radial diagrams for nutrients concentration.Radial diagrams are graphical representations that show the characteristic pro-portions of the main mineral nutrients in plants or plant components (Larcher, 2007).

Carbon and nutrients stock in the forest floor
Carbon and nutrients stocks and concentrations in the forest floor are presented in Table 3
of C was higher in Minho (13.91 Mg ha -1 ) than in Tâmega (10.30Mg ha -1 ) and Sines (9.80 Mg ha -1 ) but the differences were not statistically significant.In Minho and Tâmega forest floor contains 5% of total C stock.In Sines the forest floor reaches 10% of the total C stock.
The layer FH contains higher C and nutrients stock than the other elements of forest floor.Leaves were the most relevant element in the litter layer (L).Average concentration of C was 54% for leaves and twigs and 53% for cones.The general pattern of nutrients concentration in leaves followed the decreasing order, N > Ca > Mg > K > S > P. In twigs and cones the general order (most to less abundant) was N > Ca > Mg > S > K > P.However, in Sines, Ca was higher than N in leaves and twigs.Nitrogen concentration was higher in the layer FH which presents a lower C:N ratio than the other components of the forest floor (Table 3).Nevertheless, except for Ca and Mg, nutrients concentration in the forest floor dry mater was quite low especially for P and also K.

Carbon and nutrients stock in the understory vegetation
Stocks and concentrations of C and nutrients in the understory vegetation are presented in Table 4.The average amount of C in the understory was 2.48 Mg ha -1 in Minho, 2.68 Mg ha -1 in Sines and 4.36 Mg ha -1 in Tâmega.These stocks are not statistically different.
Height of the understory vegetation in Minho ranged between 45 and 135 cm with a mean value of 70 cm.In Tâmega it ranged from 48 to 125 cm with a mean value of 74 cm.In Sines the height interval was 49 to 84 cm with a mean height of 64 cm.On average, a very regular pattern was observed for the nutrient concentration in all regions, decreasing in the order N > K > Ca > Mg > S > P (Table 4).Average percent of C was 52% for all the three regions; nevertheless different percent C values were observed along the different understory species.Phosphorous concentration in dry weight was at very low levels in all regions.Potassium and sulfur are also at low levels.Calcium and magnesium while not high, present acceptable concentration, especially in Sines.

Carbon and nutrients stock in the downed dead wood
Carbon and nutrient stocks in the downed dead wood (logs) are presented in Table 5. Observed average values of C stocks were 0.47 Mg ha -1 in Minho, 0.34 Mg ha -1 in Tâmega and 0.57 Mg ha -1 in Sines.These values were not statistically different.The dead wood is very poor in terms of nutrient concentrations.
Basic density decreased from 0.414 g cm -3 in decay class 1 to 0.321 g cm -3 in decay class 3 (Table 6).Carbon concentration in dead wood remains more or less constant throughout the decomposition process.The average dead wood C concentration was 52% for all decay classes.Nitrogen concentration increases with the decay level despite the losses in mass and thus, the C:N ratio decreases because variations in C are minimal.Patterns of Ca, Mg and K concentration are similar, declining with the level of decay.Phosphorous and sulphur show a different pattern, firstly increasing in early stages of decomposition and then declining.

Carbon and nutrients stock in the mineral soil
Carbon storage in the mineral soil was much higher than in the other studied pools.Table 7 shows that C concentration decreased with soil depth.On average, 229 Mg ha -1 , 144 Mg ha -1 and 40 Mg ha -1 of C were stored between 0 and 60 cm of depth of the mineral soil in Minho, Tâmega and Sines, respectively.These values are signif icantly different according to the statistical analysis and are equivalent to 70-74% of total C stock of the ecosystem in Minho and Tâmega and 40% of total C stock in Sines.The soils are acid to very acid in Tâmega and Minho and moderately acid in Sines.Phosphorous is at very low levels in all sampled layers.Potassium is at low to moderate levels in the 0-10 cm depth of mineral soil and at low to very low levels for deeper layers.Concentration of magnesium is low in the upper layer (0-10 cm) being very low in the other two layers.Soils in Minho are rich in N. Nitrogen is at moderate levels in Tâmega and at low levels in Sines.Concentrations of 8.7 ppm and 19 ppm where used to compute the amount of available P and K in Sines (kg ha -1 ) respectively as exact values were not available (Table 7).Exchangeable bases were at very low levels in Minho and Tâmega, and at low levels in Sines.Only results for Ca 2+ and Mg 2+ are present in Table 7. Reported values of Na + and K + were always < 0.10 cmol(+) kg -1 .

Carbon stock in the trees
Aboveground and root biomass for live trees and corresponding C stocks are presented in Table 8.
A great variability was observed in aboveground C stocks which is inherent to the diversity of ages, productivity levels and densities in the studied stands.As mentioned before, similar stand structures were selected at sampling design, trying to minimize variability between regions.Live trees, on average, store 62.6 Mg ha -1 of C in Minho, 49.7 Mg ha -1 in Tâmega and 47.9 Mg ha -1 in Sines (aboveground + roots).C stock in Minho was higher than in the other two regions but the Between parentheses are standard errors.Different letters means significant differences (P < 0.05) between regions, using ANOVA and least significant differences test.differences were not significant according to the statistical analysis.Snags were present in only five plots all over the three regions.Their contribution for C stock was only 0.79 ± 0.43 Mg ha -1 in aboveground and 0.19 ± 0.04 Mg ha -1 in roots.

Discussion
In this study, biomass, C and nutrients stocks in the main carbon pools of even-aged maritime pine stands in Portugal were quantif ied for three contrasting regions.The sampling in each region was made so that a large range of ages and stand densities could be selected inside each region trying to sample stands of similar size and structure in the three regions.The average values for C storage in Minho, Tâmega and Sines were 308.6 Mg ha -1 , 208.7 Mg ha -1 and 100.6 Mg ha -1 , respectively (Fig. 2).
While statistical analysis did not found significant differences among regions for the C stocks in forest floor, understory, downed dead wood and live trees, these differences were detected for the mineral soil till 60 cm of depth (Table 7).Therefore mineral soil is the main cause of differences among regions concerning total stored C. The amount of C in soil in terrestrial ecosystems, including forests, is far larger than that in biomass (Ravindranath and Ostwald, 2008).According to these authors, the ratio of C in soil and living biomass  *: cumulative values for C stock.†: as reported in the results of chemical analysis.SBT is the sum of exchangeable bases.Between parentheses are standard errors.Different letters means significant differences (P < 0.05) between the same component by region, using ANOVA and least significant differences test or Kruskal-Wallis and multiple comparison of mean ranks for non normality situations.Wa: aboveground biomass.Wr: root biomass.Between parentheses are standard errors.Different letters means significant differences (P < 0.05) between regions, using ANOVA and least significant differences test.
in Europe is, on average, 64:25%.In this study, mineral soil till 60 cm of depth stored 70-74% of the total C stock in Minho and Tâmega, decreasing to 40% in Sines.Organic carbon stocks in the mineral soil presented in Table 7 are in agreement with results of Correia's (2004) work and are also inside the ranges reported for the Portuguese forest soils by Martins et al. (2009) which rely on a large sampling effort all over the country.These authors point out that there is a strong linear relationship between mean annual precipitation in Portugal and the C stock in mineral soil (0-100 cm), with higher stocks in soils from north-western Portugal (Minho region), decreasing to the South and to the inland regions as the climate becomes more arid.In Minho, the precipitation (P) surplus over evapotranspiration (ETP) is high (Table 1) with a climate Csb (dry summer cool) according to the Koppen classification.
Haplic umbrisols are the prevailing soil type, with high levels of organic matter.In Tâmega, the surplus of P over ETP is also high but annual precipitation values are lower than in Minho.In this inland region, Koppen Csb and Csa (dry summer warm) types are present but maritime pine is mainly present where the Atlantic Ocean influence prevails.Despite the fact that high values of organic matter were observed in the Tâmega soils, on average, soil profiles are less developed than in Minho and C stocks decreased fast with depth (Table 7).From the three studied regions, the lowest C stock in the mineral soil till 60 cm of depth was observed in Sines.Haplic podzols are the main soil type in Sines.Cardoso (1965) states that in Portugal about 600 mm of precipitation are required to podzolization to occur, which is the case in Sines (Table 1).Dune sands from Sines are particularly prone to leaching because of their high permeability and low water storage capacity.Sevink (1991) states that pedological processes as the rates of decalcification, acidification and podzolization are strongly dependent on precipitation surplus over evapotranspiration and on the soil temperature regime.High precipitation surplus and cold temperatures (as in boreal zones) are very favourable for the podzolization process.However this is not the case in the Sines region where the low surplus of precipitation and a long dry summer (Csa) may require a very long period of time for podzolization to occur (Sevink, 1991).On average, soils in Sines are only moderately acid which is not common on «climatic» podzols.The spodic horizon appearing at most at 200 cm from the mineral soil surface according to the WRB classification was not observed.However, in this study, soil was only sampled to a depth of 60 cm because the objective was to complement available data on C stocks for the depths 0-30 and 30-60 cm from studies in two other regions.The whole data set will be used for modelling purposes of the C stock at the default level of Kyoto Protocol for reporting changes of C stock in mineral soil which is 0-30 cm.Moreover, it is important to notice that maritime pine stands in Portuguese coastal areas were mainly planted during the first half of the 20 th century (1930's and 1950's) for dune fixation.These anthropic disturbances have certainly affected the initial condition of stored C and most of these soils are many times referred as «podzolized sands» or «podzolized soils» having low content of organic matter.
Carbon and nutrients stocks in maritime pine stands 443 Comparing the nutrient content of the mineral soil, with the standards for Portugal reported by Santos (1996) and LQARS (2000), N, significantly different among regions, was at the highest levels in Minho and lowest levels in Sines.Deficiency of phosphorous was observed in all the soils, more pronounced in Tâmega and Sines.Only one plot in Minho showed a reasonable concentration of this nutrient.At the plot level, in Minho and Tâmega K was, on average, present in moderate concentrations in the 0-10 cm layer with one plot in Tâmega presenting a high concentration.These concentrations decreased to low levels in the 10-30 cm layer and to very low levels in the layer 30-60 cm in most of the plots from both regions.In Sines, K levels were low in the superficial layer and very low in the other sampled layers.For the Mg, moderate values of concentration were observed in some plots of the Sines region, decreasing to low and very low levels in the 10-30, and 30-60 cm layers, respectively.In Minho and Tâmega, concentrations on Mg were low at the 0-10 cm layer, decreasing to very low from then to deeper levels in all plots.
Forest floor represented 5-10% of total stored C and about 20% of total aboveground C stock.The values reported here (Table 3) are similar to the mean value (10 Mg ha -1 ) indicated by Martins et al. (2009) for the organic horizons of forest soils in Portugal.Canopy litterfall represents the major above-ground pathway by which carbon and nutrients are returned to the forest floor, being an important path to the biogeochemistry cycles of the forest ecosystems (Saarsalmi et al., 2007).The mineral composition of plants ash reflects the chemical nature of the soil on which they grow (Larcher, 2007).Comparative evaluation of the concentration of the nutrients in the plant material was based on reference percent values in dry matter presented by Prével et al. (1984), Bowen and Nambiar (1984) and Marschener (1986).Leaves are the main component of the L layer being rich in N > Ca > Mg > K > S > P by this decreasing order.Similar patterns for nutrients concentration in the leaves were observed by Martins et al. (2007) for 52 years old even-aged maritime pine stands in the Tâmega region with 988 trees per hectare and dominant height of 16.2 m as well as by Kavvadias et al. (2001) for old growth maritime pine forests in Greece.The P concentration in the leaves and in all the other studied components of the forest floor was at lower levels than that reported in the above referred studies on maritime pine.There is a deficiency of P in the soils of all the regions and this is reflected in the P concentration in the plant material, both in the forest floor (mainly in the leaves) (Table 3) and in the understory vegetation (Table 4).Potassium concentration in leaves was significantly higher in Sines than in the other regions (Table 3 and Fig. 3).The same trend is observed for the understory (Table 4 and Fig. 4).However the K levels in the forest floor for all regions are quite low which was expected from the soil results.Calcium concentration in leaves and understory was significantly higher in Sines than in Tâmega and Minho.This was also the pattern for Mg in leaves but no significant differences were detected for the understory (Tables 3 and 4) (Figs. 3 and 4).The moderate levels of Ca and Mg in the plant material in Sines reflect the concentration of these nutrients in the soil.In fact soils in Sines are only moderately acid in comparison to the other regions.Twigs had the higher concentration of Ca in the forest floor.
The ratio C:N is an indicator of available N to the plants.Recent litter usually presents C:N values greater than 100 but in the lowest layers of forest floor (humus), the C:N ratio can decrease beneath 10 ( Oliver and Larson, 1990).The FH layer presented a C:N ratio around 50 in Minho and Tâmega and 72 in Sines (Table 3).
Table 3 shows that N concentration in leaves is significantly higher in Minho and Tâmega than in Sines, following the same sequential decreasing order (Minho > Tâmega > Sines) as was observed for the mineral soil (Table 7).Nitrogen concentration in the understory did not significantly differ between regions (Table 4 and Fig. 4).Porté et al. (2009) found an average value of 3.5 Mg ha -1 for the biomass stock in the understory of French maritime pine forests.In this study, biomass stocks in the understory were above the mean value reported by these authors (Table 4).Carbon stock in the understory ranged between 2.5 Mg ha -1 in Minho to 4.4 Mg ha -1 in Tâmega.On average, this carbon pool represented 2% of total C in the studied maritime pine stands and about 5% of total aboveground C stock.
Concerning dead wood, large woody debris can represent substantial amounts of biomass in old-growth natural forests (Carmona et al., 2002).However, in managed forests large pieces of dead wood are scarce (Fridman and Walheim, 2000); thus the nutrient content of dead wood in such managed ecosystems (as it is the case of the sampled stands) is of little evidence (Laiho and Prescott, 2004).This was confirmed in this study where N, P, K, Ca, Mg, and S stored in the logs represented less than 1% of the forest floor, understory and dead wood nutrient stocks altogether.The values of C stock in woody debris (logs) roughly corresponded to 0.3% of the total stored C in the studied stands, including mineral soil, and to 0.8% of the total aboveground biomass (Table 5).
Wood density decreased with decay classes (Table 6), as observed in many studies (e.g.Harmon et al., 1987;Yatskov et al., 2003).Birdsey (1992) indicated average conversion factors of 0.521 and 0.491 for softwoods and hardwoods, respectively, to estimate carbon from wood biomass values.In this study, an average conver-sion factor of 0.520 was estimated.Carbon concentration remained practically constant along the decay classes.Nitrogen concentration increased throughout the decomposition process despite losses in mass, leading to a decrease of the C:N ratio.Thus, an immobilization process seems to occur.Phosphorous and S showed a similar increasing pattern along the decomposition classes.Quite opposite, K, Ca and Mg declined with the level of decomposition.Similar patterns were reported by other authors (e.g., Idol et al., 2001;Wilcke et al., 2005).
Finally, C stock in the live trees (including roots) was higher in Minho than in the other regions.However these differences were not statistically signif icant (Table 8).Despite the rusticity of maritime pine with respect to the soil and climate conditions, in Minho there are more favourable conditions for tree growth.These climate and soil conditions may probably explain the slightly higher tree biomass in the Minho region.However it is important to note that although the sample selection tried to minimize differences between regions, trees in Minho presented slightly higher dominant height (Table 2).

Conclusions
Global C stocks of 308.6 Mg ha -1 , 208.7 Mg ha -1 and 100.6 Mg ha -1 were obtained in even-aged maritime pine stands of Minho, Tâmega and Sines regions, respectively.The mineral soil accounted for the larger proportion of the total amount of stored C and also for the significant differences observed between regions.Phosphorous was at low concentrations in the soils of all the regions and K was also at low to very low levels.This is reflected in the nutrient concentration of the plant dry matter.Calcium and magnesium were found at moderate levels in the understory vegetation and in the forest floor (leaves) from Sines.These levels statistically differed from Tâmega and Minho where soils are more acid and lower levels of these nutrients were found.Nitrogen concentration in soil signif icantly decreased in the order Minho > Tâmega > Sines.The importance of downed dead wood as a C pool in these managed stands was very small as well as was its contribution to the nutrient status of the ecosystem.The basic density of dead wood decreased along the decomposition stage.
The percent of carbon in the plant material (forest floor, understory and dead wood) ranged between 52% and 54%, slightly above the value 50% commonly used in many practical applications.

Figure 2 .
Figure 2. Carbon stocks in the main forest pools for the three studied regions.Tr: trees.Ff: forest floor.U: understory.Dw: dead wood.S: mineral soil.

Table 1 .
Summary statistics for climate variables of the studied regions

Table 2 .
Summary statistics of the stand structure variables

Table 3 .
. Total amountCarbon and nutrients stocks in maritime pine stands 439 Carbon and nutrients: stock and concentration in the forest floor Twigs 0.84 a 0.45 a 3.60 a 0.17 a 0.38 b 1.44 a 0.62 a 0.37 a 4.30 a 0.20 a 0.45 a 1.72 a 0.74 a 0.45 Cones 1.31 a 0.70 a 4.29 a 0.21 a 0.20 a 0.76 a 0.46 a 0.43 a 3.27 a 0.16 a 0.15 a 0.57 a 0.35 a 0.

Table 4 .
Carbon and nutrients: stock and concentration in the understory vegetation

Table 5 .
Carbon and nutrients stocks in maritime pine stands 441 Carbon and nutrients stock in the dead wood (logs)

Table 6 .
Average values for basic density, carbon and nutrients content of dead wood decay classes

Table 7 .
Carbon and nutrients: stock and concentration in the mineral soil by layers down to 60 cm depth

Table 8 .
Biomass and C stock in the live trees