Faculty of Biology, University of Bayreuth, 95440 Bayreuth, GERMANY.
Areas covered by mountain rainforests in East Africa have important functions as habitats for endemic species and as catchment areas for the streams in the savannah. Still, little is known about the biogeochemical processes in these often threatened ecosystems. The current study aims to present basic information on soils and different parameters of the water and nutrient cycle in intact as well as disturbed tropical mountain rainforest sites at Mt. Kilimanjaro in order to assess the long term effects of forest dissection on biogeochemical parameters.
For this purpose, the chemical and physical soil properties, soil water tension, rainfall, throughfall, litter percolate and soil solution were determined on twelve main study sites in mature forest, secondary forest and elder clearings between 2100 and 2300 m a.s.l. on the southwestern slopes of the mountain. The soil water tension was monitored from May 2000 to August 2002. The simultaneously collected water samples were combined to form fortnight samples and subsequently analysed for their content of K, Mg, Ca, Na, NO3-N, NH4-N, TOC, TON, TOS and TOP.
The soils, which comprised of several buried horizons, were classified as Andisols with high C and N stocks, some hydromorphic properties and comparatively low pH values. The latter led to very small ECEC values and low base saturation. Consequently, the retention of cationic nutrients in the soil was considered to be poor.
Rainfall amounts were slightly less at the higher sampling sites where the mature forest was located and differed between the two study years (2600 mm and 2480 mm at 2100 m, 2210 mm and 1960 mm at 2250 m). Rainfall interception was close to zero in the clearings, but reached up to 30% of incident rainfall in both studied forest types. The analyses of the soil water tension revealed the driest conditions to be under secondary forests. The reason for this was probably a combination of the topographic position of the secondary forest sites and the forest fragmentation. Since the secondary forest sites and the clearings had higher sand contents, the water holding capacity of these sites was lower than in the mature forest. In the clearings, the greatest throughfall amounts reached the ground surface so that soil water tensions were lower than in the adjacent secondary forest.
With the exception of NO3-N in the forests, all nutrients in rainfall were increased upon passage through the forest canopy, but nutrient fluxes in rainfall as well as throughfall were still low compared to other mountain rainforests, especially for Mg, Ca and K. Considering the lower biomass of the clearings compared to the forest, the amount of nutrients leached from the vegetation of the clearings was relatively high. Since NO3-N was not assimilated by the vegetation of the clearings, NO3-N throughfall fluxes were higher in the clearings as were the K fluxes, which was probably attributable to different leaf properties of the vegetation at the different sites. In the mature forest, the highest concentrations of nutrients were found in the litter percolate, followed by a pronounced decline in the soil solution. This was especially the case for K. NO3-N was an exception since it exhibited the highest concentrations only in the topsoil solution. Nutrient concentrations in litter percolate and the topsoil solution were usually higher in the clearings than in both forest types. This probably resulted in higher belowground nutrient fluxes since greater water amounts reached the ground surface in the clearings. The high belowground nutrient concentrations were likely the result of the higher mineralisation rates in the clearings induced by higher temperatures and the greater nutrient contents of the litter. With increasing soil depths, nutrient concentrations in seepage water below the clearings declined so that the differences among sites were not significant at deeper soil layers.
The contribution of organically bound nutrients to the total concentrations of respective nutrients was highest in throughfall water and lowest in the soil solution, in which OM concentrations were overall lower than in other mountain forests. This was most likely attributable to the adsorption of organically bound nutrients to the mineral phase of the Andisols studied. Again, the highest concentrations in seepage water were measured in the clearings.
The results show that mature forests at Mt. Kilimanjaro exhibit a more closed nutrient cycle, especially for basic cations, while the nutrient cycle in the clearings is more open. This probably reflects the different nutrient usage and conservation strategies of the pioneer and the late successional vegetation. Therefore, the forest disturbance on Mt. Kilimanjaro leads to long term changes in biogeochemical cycles. The opening of the forest at lower elevations, which resulted in the formation of large clearings with impeded regeneration and highly fragmented secondary forest patches, led to a higher spatial and seasonal variability of soil moisture and nutrient contents in seepage water. The large scale effects of these forest conversions on water yields and nutrient outputs need to be tested in the future.