Subantarctic forest ecology : case study of a conifer-broadleaved stand in Patagonia, Argentina

In the temperate rainforests of southern South America, the tree genus Nothofagus (Nothofagaceae) is the dominant in extension and abundance on zonal soils at different latitudes and altitudes, as well as on intrazonal (e.g., wetlands) and azonal soils (e.g., morrenic and fluvioglacial deposits). Although concern on the global role of this biome is currently important, the existing level of ecological knowledge on its functioning is still inadequate to design a sound management to maintain or enhance forest values, services, and commodities. Vegetation misuse triggers severe biota and physical deterioration, particularly in the intrinsically fragile Andean Patagonic region characterised by seasonality of high rainfall levels, low temperature, strong intensity and frequency of winds, and extensive and deep mountain slopes. In Argentina, at low and mid elevations between 38° and 43°S and the annual isohyets of 1,500 to 2,000mm, the forest is composed of the xeric cypress Austrocedrus chilensis ("ciprés", Cupressaceae) and the mesic southern beech Nothofagus dombeyi ("coihue"). The effect of the strong east - west environmental gradient, caused by the Andes' rain shadow, on community composition and dynamics is striking. This is reflected in a clear vegetational zonation, whose edges are the sparse A.chilensis woodland surrounded by the Patagonian steppe towards the east and the Valdivian rainforest co-dominated by N.dombeyi towards the west. Both communities are separated only by tens of kilometres. The conifer-angiosperm association, characterised by two groups of plants with contrasting evolutionary histories and ecological adaptations, has received little scientific attention.At 41°11'S and 71°25'W, a mixed, virgin, post-fire stand, located on a dry north-facing slope was examined regarding regeneration, size, age, and spatial structures, and stand and tree growth. Inferences on community dynamics were made. The minimum area of the community was also estimated. Density, basal area, and volume of adult trees is calculated in 658 ind ha -1(66.6% of A.chilensis ), 72.1 m 2ha -1(65.6%), and 608.7m 3ha -1(51.2%), respectively. Total density of saplings and seedlings is 2,991 (27.1% of A.chilensis ) and 7,143 ind ha -1(34.3%), respectively. Stand growth is estimated in 7.3m 3ha -1yr -1(42.5% of A.chilensis ) and 3.7 t ha -1yr -1(32.4%). Between species, individual increments significantly differ within development stages. Adults A.chilensis and N.dombeyi grow in diameter at an individual rate of 0.36 and 0.57cm yr -1, respectively, while in height at a rate of 15.8 and 29.3cm yr -1, respectively. Saplings of A.chilensis and N.dombeyi grow in diameter at a rate of 0.11 and 0.21cm yr -1, respectively, while in height at a rate of 7.8 and 17.2cm yr -1, respectively. Within species, adults grow in diameter at rates between 2.7 (in N.dombeyi ) and 3.2 (in A.chilensis ) times significantly faster than their respective saplings, and also old trees show significantly larger diameter increments than young trees. Within species, adults grow in height at rates between 0 (in N.dombeyi ) - 2.0 (in A.chilensis ) times faster than their respective saplings. In A.chilensis , old trees grow in height at significantly higher rates than young trees, whereas this is not the case for N.dombeyi . For each A.chilensis tree, a negative relationship is found between individual diameter (range= 0.22 to 0.42cm yr -1) and height growth (12 to 21cm yr -1), and the number of neighbour trees (0 to > = 7) taller than itself within a 5m distance. Contrarily, for N.dombeyi no relationship is observed between these variables. In A.chilensis , individual growth rates do not differ significantly between sex classes: for males, mean and current diameter, and height growth rate is 0.360, 0.132, and 17.6cm yr -1, respectively, while for females is 0.348, 0.124, and 17.2cm yr -1, respectively.The inherently inferior growth capacity of A.chilensis over N.dombeyi is explained by its lower leaf photosynthetic rate. Conifers from the northern hemisphere show a primary productivity similar to their mostly associated deciduous angiosperms. However, southern hemisphere conifers show a lower productivity than broadleaved. This is probably because they cannot benefit differentially from an extended period of net photosynthesis as they co-exist mainly with broadleaved evergreen species. The greater growth disadvantage of A.chilensis in comparison to N.dombeyi during the juvenile stage is consistent with the same trend observed in gymnosperms as a group: seedling represents the ontogenetic phase of slowest growing, caused by multiple factors related to leaf productivity and carbohydrate allocation. This constraint has been used to explain the decline and retreat of gymnosperms along evolutionary scales. Within species, the dependence of growth to age and crowding is indicative of one-sided, asymmetric competition for light: trees that arrived earlier to the site have pre-empt resources, developed a larger size, and therefore at present they interfere asymmetrically with those arriving later. However, contrasting responses are found according to species and growth variables: i) In N.dombeyi , the independence of individual growth to crowding, and of height growth to age/development stage of trees suggests that this species is rather insensitive to this competition type. On the contrary, A.chilensis is highly susceptible judging from the dependence between growth to age and taller neighbours. ii) In comparison to diameter growth, height growth differs slightly between age classes/stages, which reflect the common priority that trees give to height during development. Lack of gender-related growth differences in the dioecious A.chilensis would be masked by the low reproductive development exhibited in the sampled individuals. If trees in structurally well-developed stand are unable to express the reproductive potential as observed in isolated trees, then its incidence on the individual energetic budget would be marginal and marked intersexual growth contrasts would not be expected. This hypothesis would explain the incongruence with previously results suggesting a trade-off between vegetative and reproductive investments, and needs to be tested during future studies given the ecological and silvicultural relevance of this issue.In 1860, both tree species began to colonize a heterogeneous site, following a fire that eliminated the original vegetation. This first regenerative pulse lasted 60 to 70 years after which recruitment ceased probably in response to canopy closure, resulting in the present even-aged clustered adult tree populations. Because of its lower density and higher growth rates, N.dombeyi constitutes widely spaced, big emergent trees of the stand. The build-up of the A.chilensis -dominated mixed stand probably improved autogenously local moisture conditions, encouraging establishment of N.dombeyi seedlings and defining a second continuous regenerative pulse starting ca.1930. In the understorey, the sapling population dominated by N.dombeyi represents a third regenerative pulse with ages between 1 to 10 years. The current-day stand structure is represented by a young population dominated primarily by N.dombeyi , and a mature overstorey dominated primarily by A.chilensis . In absence of large-scale impacts, changes in forest structure over time would be accounted for interspecific differences in recruitment, growth rate, and sensitivity to competition, probably resulting in a local decline of the conifer component. However, the combination of a great longevity and a light-demanding temperament of A.chilensis implies that frequent to very infrequent large-scale impacts would be sufficient to maintain its abundance in the landscape. These results, together with others described in literature, suggest that divergent development patterns occur in the A.chilensis - N.dombeyi stands, probably because these forests grow under a spatially varied environment and their responses differ consequently.Based on the concept of "quantitative minimum area", the statistically optimum plot size for the ecological study of a temperate forest was examined. Changes in the estimated tree density (N), basal area (G), stemwood volume (V), and volume growth rate (I v ) per area unit, in relation to the increase of sample plot size (X), were analysed. Means of N, G, V, and I v fluctuated considerably within a range of small plot sizes, showing stability at X >= 1,000m 2. The accuracy of the parameters estimates measured as relative error (RE) increases as long as the plot size is enlarged. For X= 1,000m 2, ER varies between ± 30% and ± 49%, and for X= 2,500m 2, RE varies between ± 20% and ± 32%. Around X= 5,000m 2should be sampled to obtain RE < = ± 20% in all variables, according to the developed regressions RE= f (X) (P<0.05). This estimated plot size is larger than those often recommended in literature for ecological studies of temperate forests.