Institute for Biodiversity and Ecosystem Dynamics (IBED), University of Amsterdam, 1018 XE Amsterdam, THE NETHERLANDS.
In the late eighties of the past century, two permanent plots were established in old-growth or mature forests near Araracuara, in the central part of Colombian Amazonia. The plots were installed in two contrasting land units: upland and floodplain. The overall aim of this dissertation is to provide basic knowledge about the structure, species composition, and forest dynamics of these two permanent forest plots. The plots were installed in the Nonuya community of Peña Roja, about 70 km downstream from Araracuara. In the area the mean annual rainfall is approximately 3000 mm per year, and a dry season is lacking. The upland plot was established on a land unit belonging to the Tertiary sedimentary plain. The soils were very poor Ultisols. The floodplain plot was set-up in a rarely inundated floodplain of the Caquetá River, characterized by shallow basins alternating with convex river bars, up to 2 m high. Soils here vary from poorly to moderately well-drained Inceptisols. Both plots had a final size of 1.8 ha.
Chapter 3 accounts of the description of Pseudomonotes tropenbosii Londoño et al. This new taxon is the second reported occurrence of a member of the family Dipterocarpaceae in the Neotropics. The new entity differs from the rest of this family in the absence of fasciculate trichomes and in having sepals conspicuously aliform (reaching 10-16 cm in length) and one ovule per locule with nearly basal (sub-basal) placentation. The species appeared as one of the dominant trees in the upland plot. Subsequent studies (wood anatomy and molecular phylogeny) confirmed its taxonomic position in the subfamily Monotoideae of the family Dipterocarpaceae.
Chapter 4 reports on an architectural analysis, which was carried out for three species of Myristicaceae (Iryanthera tricornis, Osteophloeum platyspermum and Virola pavonis). The three species showed a growth according to Massart's Model, just as other species of Myristicaceae (V. michelii and V. surinamensis), during which three orders of axes were reached. Reiterations are important in the growth and development of the three species. The architectural analysis suggested that the growth plan of these species would contribute to their ability to maintain healthy populations in the subcanopy forest. The case-study of the three subcanopy tree species of Myristicaceae shows the large potential of architectural analysis to build hypotheses why species might be adapted to particular conditions of forest development, especially regarding abundant subcanopy (or canopy) species.
A total of 1149 species, 347 genera and 98 families were found in the two plots (Chapter 5). Arborescent species comprised 65% of the total at both sites; climbers (woody and herbaceous) contributed 24%, shrubs 8% and terrestrial herbs 3%. The number of species was larger in the upland plot (698) than in the floodplain plot (511), but the number of genera was quite similar (236 in the upland and 235 in the floodplain plot), whereas the number of families was slightly higher in the floodplain plot (84, compared to 80 families in the upland plot). Between the plots 67% of the families were shared, 36% of the genera, and only 5% of the species. In total, the floristics in the plots showed a strong affinity to the flora recorded in central Amazonia, near Manaus.
The mortality in the upland plot was almost twice as low as in the floodplain plot (λ = 1.06 versus λ = 2.0, respectively, over 8.6 y)(Chapter 6). In terms of basal area and biomass, the discrepancies in mortality rates between the floodplain and upland plot were even more pronounced. Large trees contributed substantially to the mortality rates calculated for basal area and biomass. For the total census period (8.6 y) 29% of the trunks died standing in the upland plot, and only 17% died as such in the floodplain plot. Regarding death by uprooting (tree fall) the results were opposite: in the upland plot only 10% of the trunks had died by uprooting (over 8.6 y) whereas in the floodplain this proportion was two times higher. In terms of individuals the recruitment in the upland plot was almost two times lower than in the floodplain plot (κ = 0.87% against κ = 1.48% per year, respectively, over 8.6 y). Similar between-plot differences were found for recruitment in terms of basal area and biomass. The average annual growth in terms of diameter increment was 0.13 cm per year in the upland plot. In the floodplain plot is was almost twice as high (0.22 cm per year). The density-based annual mortality rates of the upland plot were near the lower limit reported for the northeast Amazonia. The density-based annual mortality rates of the floodplain plot were at the lower limit for the forests in northwestern Amazonia. For both plots the mortality rate based on number of individuals was higher than recruitment rate, indicating that mortality preceded recruitment. The average of the mortality and recruitment rates, which represent a measure of turnover, was lower in the upland plot than in the floodplain plot. Compared to other Neotropical forests, the turnover in both plots was relatively low.
The discovery of Pseudomonotes tropenbosii in the upland plot illustrates the strong progress in floristic inventories in Colombian Amazonia since the mid-eighties of the past century (Chapter 7). On the basis of wood anatomy and phylogenic analyses its taxonomic position has been confirmed. The architectural analysis of the three species of Myristicaceae suggested that the growth plan of these species would contribute to their ability to maintain healthy populations in the forest subcanopy. Architectural analyses by means of observations along transects complements demographic studies of forest dynamics in permanent plots. The higher species richness in the upland plot compared to floodplain plot and the low overlap in species composition between the two plots fitted well into the general scheme of upper Amazonian forest-landscape associations. In general these patterns are explained in terms of niche assembly or random walk processes. In the upland plot where soil nutrient availability and decomposition rates were far below those in the floodplain plot, tree turnover was lower. This corresponds to patterns found in Amazonia as a whole. However, the tree mortality rates in the upland plot were among the lowest recorded for Amazonian as a whole, and clearly far lower than would be expected purely because of the location of this plot in upper Amazonia. Despite its low levels of tree turnover, the species richness in the upland plot was higher than that of the floodplain plot. Therefore the results from the two permanent plots indicate that it is hazardous to apply insights based on regional studies to local situations without proper knowledge of the local physiography and terrain conditions. It is highly recommended to continue the monitoring of the forests in the permanent plots near Peña Roja, because the upland plot represents the type locality of Pseudomonotes tropenbosii because the two plots represent the oldest permanent plots in Colombian Amazonia.