Spatial Variation in the Structure and Litterfall of a Sumatran Rain Forest

Abstract

In this paper an explanation for the spatial variation in the structure of a Sumatran rain forest is suggested. The forest area consisted of a series of topographical units (terraces, mountain slopes) of increasing age and decreasing pH and fertility. With increasing soil age the production of leaf litter and fruit decreased whereas the proportion of mature-phase forest increased and gap cover decreased. There is variation in the growth strategies of trees depending on soil fertility; this results in higher tree longevity on poor soils, and hence in forest with fewer gaps and with more large trees, than on rich soils. Thus, Janzen's (1974) hypothesis of the effect of soil fertility on the prevalence of anti-herbivore defenses and fruiting periodicity is extended to include aspects of forest structure. THERE IS LONG-STANDING EVIDENCE of spatial variation in the floristic composition of the Malesian rain forests (van Steenis 1935; Richards 1952; Ashton 1964, 1982; Whitten 1982). More recently, much attention has been given to the temporal variation in floristic composition in both the Malesian (Poore 1968) and the neotropical (e.g., Hartshorn 1980) regions. The interest in forest dynamics has led to a number of studies in which structure and productivity are linked with secondary successional processes (e.g., Odum et al. 1970). For forests in the wet and moist tropical life zone it is established that leaf area index and leaf biomass are restored within a few years of the start of secondary succession (Golley 1978, Snedaker 1980) and that the rate of increase in total biomass and litter production slows down after some 20 years (Brown and Lugo 1982). In the mature phase, however, net primary production and perhaps litterfall tend to increase again (Whitmore 1975). There are no such dear patterns of spatial variation in structure and productivity of the forests for the wet and moist tropical life zone. The general concensus is that soil factors affect neither litter production (Tanner 1980, Leigh and Windsor 1982, Proctor et al. 1983b) nor structure and biomass (Proctor et al. 1983a). The aim of the present study is to compare the structure, dynamics, and litter production of a number of habitat types within one small area of tropical rain forest containing soils of various ages and fertilities. The results suggest that variation in soil fertility may be responsible for the observed spatial variation in structure and productivity. STUDY AREA The upper half of the course of the Alas river lies in the Semangko Rift zone, a graben running the full length of Sumatra and flanked by the Barisan mountain range (Verstappen 1973). The Ketambe field station (3041'N, 97?39'E) lies just below the confluence of the fast-flowing and rocky Alas and Ketambe rivers (Fig. 1) at an altitude of about 3 50 m. The graben is only a few km broad here and contains a number of accumulation terraces (van Beek 1982). The study area (about 200 ha) includes these terraces and the lower slopes of the mountain range (see fig. 16 of Rijksen (1978) for altitudes) which consists mainly of crystalline metamorphic schists, alternating in places with hard limestones of late Paleozoic origin (van Beek 1982). When the Toba volcano erupted some 75,000 years ago (Francis 1983) the exit from the rift valley was blocked by inflowing volcanic material. As a result the valley filled with ignimibrites and became a lake. The erosion base has since been lowered discontinuously by tectonic activity and, due to variation in the weathering resistance of the layered sediments, resulting in a series of terraces. It is estimated that the highest of these terraces became dry about 10,000 yr ago (van Beek, pers. comm.). The area is considered sacred by the local inhabitants of the Alas valley and, with the possible exception of one small part of the lowest terrace near the field station, has not been disturbed by human activities other than the collection of forest produce. Some people say that this part of the lower terrace was prepared for cultivation of rice around 1940 but was abandoned after an outbreak I Received 5 December 1983, revision accepted 14 June 1984. 196 BIOTROPICA 17(3): 196-205 1985 This content downloaded from 157.55.39.255 on Sat, 16 Apr 2016 04:28:25 UTC All use subject to http://about.jstor.org/terms