RAPID ASSESSMENT OF TROPICAL RAIN-FOREST SUCCESSIONAL STATUS USING AERIAL PHOTOGRAPHS

Four forest structural types were recognized and mapped from 1:20,000 black-and-white aerial photographs of the 2000 ha Central Catchment Na ture Reserve in Singapore. These types were an essentially treeless un it of grassy areas and fern thickets (Type 1), a low forest with unifo rm canopy of many small trees (Type 2), a taller forest with larger tr ees (Type 3), and a yet taller forest with a structurally more heterog eneous canopy with some very large-crowned trees present (Type 4). The three vegetation units containing trees were sampled using clusters o f circular plots totalling 0.2 ha each, in which all trees greater tha n or equal to 30 cm gbh were measured and identified Increasing canopy height and tree size among the structural types were reflected in inc reasing total basal area of the sample clusters. Mean tree species ric hness also increased with structural complexity. Ordination of the clu sters based on their floristic composition showed that Type 4 was very varied but generally distinct from Types 2 and 3, which were much les s diverse and not distinct from one another floristically. TWINSPAN wa s used to analyze the important floristic differences between the samp le clusters. It was found that Types 2 and 3 were dominated by early s uccessional species including Adinandra dumosa, Macaranga conifera and species of Calophyllum and Garcinia. Type 4 was typified more by the presence of species of Burseraceae and Dipterocarpaceae - primary fore st trees. This mosaic of different successional stages is supported by the known history of the area, with massive deforestation in the mid and late 19th century. We conclude that the interpretation of aerial p hotographs is a useful tool for the fine-scale vapid assessment of the successional status of lowland tropical forest and can be used to inf er relative levels of forest diversity. As such it can be of value in the rational management of tropical forest conservation areas. (C) 199 6 Elsevier Science Ltd