Introduction

Silvopastoral systems (SPS) are defined as a land-use system in which tree species grow in association with forage species, deliberately designed to optimize the use of spatial, temporal, and physical resources. Both thinning and pruning are common practices in forest plantations, but in SPS are carried out more frequently than in “traditional” forestry systems designed to obtain only wood. In SPS, these silvicultural practices must be done in order to manage resource use by the different components of the system, trying to increase facilitative interactions between trees and grasses or crops, and decrease negative ones (Jose et al. 2000). Pruning and thinning of trees decrease the total leaf area of the stand, reducing radiation interception by trees and thus decreasing their negative effects on understorey species. In addition, pruning may be a way to control water use by individual trees reducing competition for this resource (Jackson et al. 2000; Gyenge et al. 2008). However, these practices produce different and potentially opposite effects over tree growth depending on the species, plantation density, inter- and intraspecific competition (Oliver and Ryker 1990; Balandier and Dupraz 1999) and environmental characteristics of the site (Balandier and Dupraz 1999).

Considering the effects of pruning on tree growth, it was reported that it usually produces a decrease in annual diameter growth, with no influence on height growth (Kozlowski and Pallardy 1997). Results of Mowat (1947) in ponderosa pine agree with this observation and showed that diameter growth was more sensitive to pruning than height growth. Pruning also negatively affect branch growth of ponderosa pine trees (Gyenge et al. 2008). Based on the mentioned results, our expectation was that stem growth, in particular diameter of ponderosa pine growing in SPS in NW Patagonia will be affected by pruning.

As was mentioned, annual growth depends on intra- and inter-specific competition for resources. Volumetric growth of 102 plantations with different tree densities of ponderosa pine (Pinus ponderosa Doug. ex Laws) in NW Patagonia Argentina covering different site quality, age and management, was 18.5 ± 11 m3 ha−1 year−1, with 9 ± 5 mm year−1 of increment in diameter at breast height (dbh; Andenmatten et al. 2002). These results were similar than those found in ponderosa pines growing in Oregon, USA, where an annual increment of 9 mm at dbh was measured in trees growing without understorey against 1.2 mm in trees growing with understorey (Oliver and Ryker 1990). In a similar way, higher growth rates were found in USA ponderosa pines growing at low tree density and without understorey (309–153 trees ha−1; Oren et al. 1987) compared to systems with high competition.

From the ecophysiological standpoint, the foliar area:sapwood area ratio (F a:SA) of a tree depends on water resource availability. For example, Maherali and DeLucia (2001) found that ponderosa pines acclimate to different site conditions changing this relationship, being lower in dry than in wetter sites. In spite that pruning changes this ratio, theoretically increasing the efficiency of water transport to the remnant foliage, not all species can take advantage of the released sapwood area. In the particular case of ponderosa pine, previous results indicate that pruning creates an imbalance in hydraulic system at the plant level, leading to water stress conditions (Gyenge et al. 2008). Therefore, it is possible that the plant try to return to the original F a:SA ratio, which is important to determine from the management point of view.

At the present time, scarce information is available about ponderosa pine growth performance under SPS conditions in NW Patagonia. Previous studies in this region have considered the forage component of the systems (e.g. Fernández et al. 2002, 2004, 2006), the effect of pruning on branch production (for bole quality) and water relations in ponderosa pines (Gyenge et al. 2002, 2003, 2008), and competition for belowground resources between trees and grasses (Fernández et al. 2008). However, no studies are available considering pruning and low density plantation effects on growth rate of ponderosa pines. Therefore, we performed three experiments aiming at: (a) quantifying the effects of different stand densities (SPS vs. conventional density) on annual wood productivity; (b) quantifying the effects of pruning on stem growth and (c) on the foliar to sapwood area ratio of ponderosa pines growing in NW Patagonia. This information will be useful for practical purposes considering the performance of the tree component in SPS in NW Patagonia.

Materials and methods

Study site and experimental treatments

The study site was located in Lemu Cuyen Ranch (40.3°S, 71.1°W, 900 m of altitude) in Neuquén Province, Patagonia, Argentina. The climate of the region is characterized by cold and wet winters, and hot and dry summers. During the period 1978–1999 average annual rainfall was 684 ± 283 mm (with approximately 579 mm in fall-winter and 105 mm in spring–summer). Maximum and minimum annual average temperatures are 17.1°C ± 0.5 and 4°C ± 2.1, respectively. The soil is a silty clay loam, with 27–40% loam and 20% clay (A. Marcolín, Personal Communication, INTA Bariloche 1998).

All measurements were carried out in experimental silvopastoral plots (SPS) and a high density pine stand (HPP, 1,300 pines ha−1, approx. 5 ha) established in 1985 (approx. 15 years old at the beginning of the experiment). Considering the SPS, in winter of 1999, ten study plots (40 m × 40 m each) were randomly distributed within a 2 ha stand with an original density of 500 trees ha−1 (500 SPS). After that, five plots were randomly selected and thinned to decrease the original density to 350 trees ha−1 (350 SPS). In order to avoid edge effects, only trees of the central 30 × 30 m subplots were measured. Average tree height and dbh at the beginning of this experiment were 7.77 ± 0.31 m (±SD, n = 163) and 21 ± 0.61 cm, respectively. At winter 1999 (beginning of the experiment), all trees were pruned with handsaw up to a height of 3 m. At winter 2001 and 2003, a second pruning at 4.5 m height was imposed to trees in half of the plots of both densities (see Experiments 2 and 3). The HPP was located 50 m apart from the SPS plots. In this stand, only few trees formerly selected by the owner of the ranch (may be due to their good stem taper) were pruned at 4 m height during winter 1998, before the beginning of the experiment. Both sites presented similar slope, soil depth, and exposure conditions. Understorey of the SSP plots was mainly composed by the native grass tussocks Festuca pallescens ((St.-Yves) Parodi) and Stipa speciosa (Trin. And Rupr.), whereas no understorey was present within the HPP stand. A detailed description of both stands could be found in Fernández et al. (2002, 2004, 2006, 2008) and Gyenge et al. (2002, 2008).

Initial leaf area index (LAI, m2 m2) in all treatments was measured in the summer 1999–2000 with a sunscanner (Sunshine Sensor type BF2, Delta-T Device Limited, Cambridge, England, UK). Mean LAI was 0.97, 1.83 and 6.4 for plots with 350, 500 and 1,300 pines ha−1, respectively.

Experiment 1. Annual stem growth at different plantation densities

To estimate annual stem growth of the trees, we measured dbh (using a diameter tape) of all the trees of the SPS plots every winter (from 1999 to 2004). In the HPP treatment, four plots 100 m2 each were randomly selected for growth measurements. Diameter growth was estimated measuring the rings from increment cores with a digital caliper in the laboratory. In each plot, we selected nine trees in each HPP plot, three per social status: dominant (dbh > 20 cm), co-dominant (15.1 < dbh < 19.9 cm) and suppressed trees (dbh < 15 cm). Annual volume of each tree was estimated yearly following the formula:

$$ {\text{dry mass of stem (g)}} = \, 93.648{\text{ dbh (cm)}}^{2.189} \, \left( {{\text{Laclau }}2003} \right) $$

were dbh represent the diameter in each year estimated from cores. We used a mean wood density of 0.434 kg dm−3 (Laclau 2003). This formula was developed from pines growing in the same site (Lemu Cuyen Ranch) but in stands other than those measured in this study. In addition, we compared the estimated volume of SPS trees with that estimated with Andenmatten’s double-entry formula developed for ponderosa pines of NW Patagonia (Andenmatten et al. 1995):

$$ {\text{Vol}}\left( {{\text{m}}^{3} } \right) \, = \, 0.02985 \, + \, 0.00003272{\text{ height (m)[dbh (cm)]}}. $$

No differences were found between stem volumes estimated with both models (P > 0.05, F test, Neter and Wasserman 1974).

Stand annual growth (m3 ha−1) of SPS plots was estimated adding individual growth of all the trees in the sampled area, and scaling that value to a hectare. In the HPP stand, stand annual growth was estimated from the diameter distribution of trees measured in the four plots.

We applied ANOVA and Tukey tests (α = 0.05) to determine statistical differences in individual growth (mm dbh tree−1 year−1) and stand productivity (m3 ha−1 year−1) between both SPS treatments (350 and 500 trees ha−1) and between trees growing in HPP stand of different size range, within each year. We qualitatively compared these values to those of HPP stand because we could not applied ANOVA due to the lack of true replicates of plots in this last system.

Experiment 2. Effect of pruning on stem growth

To study the effect of pruning on dbh and height growth in SPS, we compared results of measurements obtained in 4.5 m-pruned trees against 3 m-pruned trees. Height (measured with a clinometer) and dbh of ten randomly selected trees in each SPS 500 and SPS 350, were measured every winter (from 2000 to 2003).

We determined relative growth rate following the formula:

$$ {\text{Relative growth rate }} = \, \left( {{\text{value}}_{t} \, - {\text{ value}}_{t - 1} } \right)\left( {{\text{value}}_{ \, t - 1} } \right)^{ - 1} $$

where t was time and “value” represented the dbh or height values of each tree in each measurement date.

To determine statistical differences in tree growth between pruning treatments and densities we applied ANOVA with two factors for each studied year, with α = 0.05. We apply post hoc Holm-Sidak method for comparison between factors.

Experiment 3. Effect of pruning on the foliar area:sapwood area ratio (F a:SA)

Based on the results of relative growth rates (see Results Experiment 2), we only measured SA and F a annual growth of pines growing in SPS of 500 ponderosa pines ha−1. As was determined in ponderosa pines growing in the same place by means of sapflow density measurements at different radial depth, a total reduction of sapflow density could be observed approximately at 15 cm of the radii. For this reason, hydroactive sapwood area was estimated as the basal area at the breast height minus the bark in those trees with a dbh lower than 30 cm (Gyenge et al. 2003). Leaf area of each tree was estimated following the methodology described in Gyenge et al. (2008). We firstly developed a model to estimate F a of each branch based on its diameter at the base of the stem (DIAM) (Gyenge et al. 2008):

$$ F_{\text{a}} = 0.0142 + 0.001\,{\text{DIAM}}^{2} $$

We measured the dbh and DIAM of all branches in 3–4 trees of each pruning treatment of the SPS 500 plots, in winter 2003, 2004 and 2005.

Results

Experiment 1. Annual stem growth at different plantation densities

Both absolute and relative (standardized by their dbh) annual diameter growth were similar in pines growing in both SPS (P > 0.05), except in season 2000–2001 when pines growing at 350 trees ha−1 showed higher growth rates (Fig. 1). On average, annual dbh growth of SPS trees was 17.6 ± 3.6 mm. On the other hand, dominant trees of HPP plots showed a higher absolute growth rate (9.3 ± 1.7) than medium (5.4 ± 2.2) and suppressed trees (3.3 ± 1.2; P > 0.05; Fig. 1) of the same stand. The mean growth values of HPP trees was 6 (±3) mm per year. Averaged relative growth rate of SPS trees almost doubled the growth of HPP trees (7.7–3.1%, respectively; Fig. 1).

Fig. 1
figure 1

Annual absolute (mm, a) and relative diameter growth (b) of Pinus ponderosa growing at three plantation densities (1,300, 500 and 350 trees ha−1) in NW Patagonia

Full size image

In the HPP plots, individual annual volume growth was no statistically different in suppressed and co-dominant trees: 0.003 and 0.007 m3 tree−1 year−1, respectively (P > 0.05). These values were lower than individual growth of dominant trees of the same plantation (0.024 m3 tree−1 year−1, P < 0.05). A higher volume annual growth was observed in the trees growing in SPS than in the HPP stand, and no statistical differences were observed between both low plantation densities (0.042 m3 tree−1 year−1, P > 0.05).

At the stand level, different growth rates were observed in years 2000 and 2002 (corresponding to growing season 1999–2000 and 2001–2002, respectively) between both SPS treatments (Table 1; P > 0.05). The higher growth rate of SPS 500 during these seasons incidentally corresponded to periods with lowest absolute values of diameter growth, in which trees of the lowest plantation density (SPS 350) decreased in a higher magnitude their growth compared to “good” years (Fig. 1). This higher sensibility to interannual environmental conditions of SPS 350 trees increases whole stand differences with SPS 500 in “bad” years but decreases the differences under good growing conditions. On the other hand, growth rates in the HPP plots showed a lower variability between years, being similar than the stand growth rates of SPS of 350 trees ha−1 in spite of the very different amount of trees per hectare.

Table 1 Annual volume growth rates (average ± SD) at individual (m3 tree−1 year−1) and stand (m3 ha−1 year−1) level of ponderosa pine (Pinus ponderosa Doug. ex Laws) plots at three plantation densities: 1,300 (HPP), 500 (SPS 500) and 350 (SPS 350) trees ha−1
Full size table

Experiment 2. Effect of pruning on stem growth

After the first pruning (3 m height) treatment was applied to all trees, diameter growth differed between trees of SPS 350 and 500, being higher in the first treatment (Table 2). Then, some trees were subjected to another 1.5 m of green pruning (up to 4.5 m height). After that, those trees presented lower diameter growth rates in comparison with trees with 3 m pruning, both in SPS 350 and 500. Significant differences were observed in the first (2002) and in the second year (2003) after pruning (Table 2).

Table 2 Mean relative annual growth in height and diameter at the breast height (dbh) during 3 years (2001–2003, average ± SD) of ponderosa pine (Pinus ponderosa Doug. ex Laws) stands of two plantation densities 500 (SPS 500) and 350 (SPS 350) trees ha−1, and two green pruning height levels [3 m pruning (winter 1999) and 4.5 m pruning (winter 2001)]
Full size table

Considering height growth, individuals differed in this variable even before the second pruning was applied (Table 2), and because of these initial differences, pruning had a general effect of homogenizing the different treatments (Table 2). As a whole, and as was expected, no clear effect on height growth can be attributed to pruning treatments.

Experiment 3. Effect of pruning on the foliar area:sapwood area ratio (F a:SA)

Neither absolute nor relative annual growth of F a was statistically different between both pruning levels, but this was probably due to the high variability observed in data of 3 m-pruned trees. Values of annual F a growth was 12.7 ± 3.3 (SD) m2 for 4.5 m-pruned trees and 21.2 ± 7.8 m2 for 3 m-pruned ones. Relative growth (weighed by the previous F a) was 31 and 25% for 4.5 m-pruned and 3 m-pruned pines, respectively.

The relationship between F a and SA, as was expected, decreased immediately after the pruning treatment in a higher magnitude in the more-pruned trees (0.17 and 0.11 m2 cm−2 for 3 and 4.5 m-pruned trees, respectively; Table 3). The F a:SA ratio increased annually in both treatments, indicating a higher relative foliar growth in comparison to stem (sapwood area) growth (Table 3). It is interesting that this variable continued increasing in both pruning treatments during the studied period (3 years after pruning), but with a higher magnitude in the 4.5 m-pruned trees compared to the 3 m-pruned trees (considering changes from 2002 to 2003 and 2003 to 2004, a mean increment from 1 year to the following year of about 16 and 7%, respectively). This led to a similar F a:SA ratio in both treatments 1 year after the pruning (Table 3).

Table 3 Ratio between leaf area (F a, m2 ± SD) and hydroactive sapwood area (SA, cm2) in ponderosa pine (Pinus ponderosa Doug. ex Laws) trees (mean and standard deviation)
Full size table

Discussion

Stem growth of 350 SPS trees tended to be higher than stem growth of the 500 SPS trees (at least in some years), and both were higher than HPP trees demonstrating the negative effect of intra-specific competition even at low density plantations such as SPS 500. At the stand level, maximum annual productivities were observed in 500 SPS compensating their lower growth rates compared to SPS 350 trees with the higher amount of trees per hectare. This demonstrated that the wood productivity in SPS could be higher (in SPS 500) or similar (SPS 350) than the traditional ponderosa pine plantations in NW Patagonia (with approx. 1,100 trees per ha). However, grass productivity begins to decrease at the 500 trees ha−1 density (Fernández et al. 2004), emphasizing the trade-off between a high productivity of both wood and forage in the same land unit. Average diameter growth rates observed in the HPP stand (6 mm year−1) were within the range reported previously for other commercial plantations in NW Patagonia (9 ± 5 mm year−1; Andenmatten et al. 2002). Similarly, the advantage in growth as a consequence of the decrease in resource competition, as was observed in both SPS densities (18 mm year−1), has also being measured in ponderosa pines growing without understorey and at low plantation density in California (USA) (Oren et al. 1987; Oliver and Ryker 1990). In spite that this species has several mechanisms leading to its high drought resistance by means of drought avoidance (Piñol and Sala 2000), it seams that the shortage in soil water resources due to a combination of low precipitation and a high amount of trees, has marked consequences on individual tree growth. The intra-specific competition by soil water resources in NW ponderosa pine forests has being studied from the ecophysiological standpoint by Licata et al. (2008), who showed that SPS stands had relatively high and steady rates of transpiration even during the summer drought period, while HPP transpiration rates consistently decreased during that part of the season. The associated carbon exchange was surely being reduced in parallel, as was observed in other ponderosa pine forests (e.g. Anthoni et al. 2002), explaining the differences in annual productivity between different density stands.

The negative effect of pruning over diameter growth was observed in both 350 and 500 SPS pines. Pruning produced symptoms of water stress such as lower leaf water potential and stomatal conductance (Gyenge et al. 2008) with a consequent decrease in water use (Gyenge et al. 2008; Jackson et al. 2000). These changes in water exchange due to a decrease in canopy conductance surely led to a reduction in carbon assimilation. The magnitude of the negative effect of pruning may depend on the extracted branch position in the canopy (shaded or sunny branches) and/or the magnitude of extracted F a (Montagu et al. 2003). In this regard, significant differences were observed between 3 and 4.5 m pruning treatments in both density stands, with a higher reduction in growth in the higher pruning level.

The F a:SA ratio of a tree or a stand depends on species-specific characteristics and on the water balance of the site (e.g. Piñol and Sala 2000; Maherali and DeLucia 2001). The values of annual F a growth observed in the present study (25 and 31%) were higher than those reported for a commercial ponderosa pine plantation in USA (15%; Oren et al. 1987). The F a:SA ratio observed (ranging from 0.167 to 0.192 m2 cm−2) was similar than that measured in the most humid places of the ponderosa pine natural distribution in USA (around 0.21 m2 cm−2; Callaway et al. 1994; Maherali and DeLucia 2001). In addition, after the pruning treatments, trees continuously maintained increasing values of F a:SA ratio, with higher increments in the more pruned trees. All these evidences suggest that water is not an important limiting resource for productivity in ponderosa pines growing at low plantations densities (SPS) in semiarid Patagonia. In this regard, in addition to a potential decrease in carbon fixation after pruning (due to the direct reduction in leaf area and/or the stress symptoms previously described), it is possible that the lower stem growth rates after pruning observed in the studied pines could be the result of changes in carbon allocation within the plant. As was described in the results section, the increase in F a:SA ratio indicates a higher relative growth of the foliage than the stem (sapwood area), reflecting a compensatory or a balanced response (after foliage extraction) for the one hand, but at the same time, due to the high F a:SA values, that trees are sensing good hydric conditions as a whole. On the other hand, we did not study the response to pruning of HPP trees. Their lower growth rates even without pruning could be indicating their worst hydric conditions, suggesting that the water stress imposed by the pruning itself may probably have a more marked negative effect on stem growth than that observed in SPS trees. Much more has to be learned about the interaction between stand density-site quality (water stress level)-pruning level in order to develop a management program based on these issues.

Management implications

Stem productivity values measured in the treatment with 500 pines ha−1 (LAI 1.83) were similar or higher to previously reported annual productivity of ponderosa pine plantations in NW Patagonia [average: 18.5 m3 ha−1 year−1, (Andenmatten et al. 2002); average: 10 m3 ha−1 year−1 (SAGPyA 2001)]. However, it is important to note that the previously reported values correspond to stands with much more trees per hectare than those measured in this study indicating the differences in individual stem growth. In Patagonian region, as in other parts of the world, there are important concerns about environmental impacts of forest plantations. Studies in this region have shown that low density or open canopy plantations can have a significant lower impact on important ecosystems processes (biodiversity, water use, resistance to plagues) than high density stands (Gyenge et al. 2009). Our results indicate that it is possible to maintain high wood productivity values by planting or leaving relatively few trees per land unit, which in addition to their lower environmental impacts, have the potential economic advantage related to concentrating growth in few large trees. This could contribute to solve the challenge of a sustainable forest management in Patagonian region.

In conclusion, high growth rates measured in pines of SPS suggest that it is possible to shorten the rotation period in semiarid Patagonia, producing high quality wood (knot-free boles) and decreasing environmental impacts of fast growing species-based forestry. However, our results indicate negative effects of pruning on relative diameter growth, and that this effect may be mediated by resource availability. For this reason, future research is needed in order to quantify pruning effects over stem form in interaction with variable climate and site conditions. Another important topic to be investigated is the relationship between the high growth rates of pines in SPS and their wood density, the most important variable determining wood properties.