In 2000, we established a 24-ha plot in Peninsular Thailand to investigate how forest composition, structure and dynamics vary with spatial heterogeneity in resource availability. Detailed soil and topographic surveys were used to describe four edaphic habitats in the plot. Disturbance history was inferred from historical records and floristic analysis. The plot included >119 000 trees ≥1 cm dbh in 578 species, and was recensused in 2010. Species distributions, floristic turnover, stand structure, demographic rates and biomass dynamics were strongly influenced by heterogeneity in soils, topography and disturbance history. Over 75% of species were aggregated on specific edaphic habitats leading to strong compositional turnover across the plot. Soil chemistry more strongly affected species turnover than topography. Forest with high biomass and slow dynamics occurred on well-drained, low fertility ridges. The distribution and size structure of pioneer species reflected habitat-specific differences in disturbance history. Overall, above-ground biomass (AGB) increased by 0.64 Mg ha−1 y−1, from 385 to 392 Mg ha−1, an increase that was entirely attributable to recovery after natural disturbance. Forest composition and stand structure, by reflecting local disturbance history, provide insights into the likely drivers of AGB change in forests. Predicting future changes in tropical forests requires improved understanding of how soils and disturbance regulate forest dynamics.

Structure and demographics in many tropical forests is changing, but the causes of these changes remain unclear. We studied 5 y (2005–2010) of species turnover, recruitment, mortality and population change data from a 20-ha subtropical forest plot in Dinghushan, China, to identify trends in forest change, and to test whether tree mortality is associated with intraspecific or interspecific competition. We found the Dinghushan forest to be more dynamic than one temperate and two tropical forests in a comparison of large, long-term forest dynamics plots. Within Dinghushan, size-class distributions were bell-shaped only for the three most dominant species and reverse J-shaped for other species. Bell-shaped population distributions can indicate a population in decline, but our data suggest that these large and long-lived species are not in decline because the pattern is driven by increasing probabilities of transition to larger size class with increasing size and fast growth in saplings. Spatially aggregated tree species distributions were common for surviving and dead individuals. Competitive associations were more frequently intraspecific than interspecific. The competition that induced tree mortality was more associated with intraspecific than interspecific interactions. Intraspecific competitive exclusion and density-dependence appear to play important roles in tree mortality in this subtropical forest.

Changes in the lowland dipterocarp forest structure and composition are described for a 34-year period between 1947 and 1981 in Sungei Menyala Forest Reserve, Peninsular Malaysia. Although tree density declined by about 10%, basal area changed little, averaging 32.4 m2 ha−1. Size class distributions in 1947 and 1981 were not significantly different. Mortality rates were independent of size class. Mortality exceeded recruitment during the first 16 years to 1963, but was thereafter almost exactly balanced by recruitment. Recruitment rate increased significantly for the latter part of the study.

Mortality was greater than average for understorey species and lower for emergents. Pioneer and late seral species together showed significantly higher mortality rates. Amongst the commoner species, emergent, late-seral and pioneer species showed the highest annual diameter growth rates with species averages over all size classes exceeding 3 mm yr−1; rates for main-canopy species were between 1.5 and 2.5 mm yr−1, and understorey species generally less than 2 mm yr−1.

Species richness was almost identical in 1947 and 1981 (243, 244 species), but a clear decline (to 229 species) between 1953 and 1971 was recovered by 1981. Most species showed little net change in density over the 34-year period, but eight of the 32 commoner species showed significant changes, all unidirectional. These changes in species composition are not negligible, but further analysis is needed to determine if, for example, they are related to disturbance of part of the plot in 1917.

In January 1983, two 50 m × 50 m rain forest plots in the University of Ife campus in southwestern Nigeria were demarcated for a baseline study of species composition, litterfall and tree girth increments. By accident, a severe ground fire burnt one of the plots on 31 January, barely two weeks after litter traps had been set and species listing and first girth measurements completed.

The effect of this fire was assessed in the burnt plot after 14 months, in April 1984. The fire affected small trees especially, and there were indications that some species were more sensitive than others.

Manihot glaziovii, which was present in the plot and abundant in its vicinity before the fire showed a big increase in density after the fire because it was able to germinate in the openings created by the fire. These observations were related to succession, and it is concluded that the observation by some workers that the initial composition of a regrowth forest may be partly predicted from a knowledge of the seed bank in the soil may be applicable in the present case.

Elephants (Loxodonta africana (Blumenbach 1797)) can have a major transforming effect on savanna structure through felling, debarking or uprooting trees (Dublin et al. 1990, Laws 1970, Mapaure & Campbell 2002). However, it is difficult to separate their influence from that of other causes of tree mortality, including wind storms (Spinage & Guinness 1971), drought (Lewis 1991, van de Vijver et al. 1999), fire (Higgins et al. 2000), and in some situations frost (Childes & Walker 1987, Holdo 2006), especially when interactions among them may occur (de Beer et al. 2006, Laws et al. 1975, Pienaar et al. 1966). Furthermore, the consequences for woodland dynamics depend on the size classes of the trees affected, as well as on how the disturbance is concentrated in time and space. Mortality of canopy trees has a much greater and longer-lasting impact than losses among the regenerating stages of these trees. However, the consequences may be less adverse for ecosystem function and biodiversity if the disturbing effects are locally concentrated, generating a patch mosaic of stands at different stages of regeneration (Remmert 1991).