Auswirkungen von ENSO-Trockenperioden und Landnutzungspraktiken auf die Dynamik von C, N und P in einem tropischen Regenwald und in Agroforst-Systemen in Zentral-Sulawesi, Indonesien

This thesis aimed to characterize the response of nutrient (C, N, P) and cation (Ca2+, K+, Mg2+, Na+) soil pools and dynamics to experimental drought and land-use practices in Central Sulawesi, Indonesia. To carry out a replicated throughfall exclusion experiment in an outlying district of the Lore-Lindu-National-Park, and in a for this region typical Cacao-Agroforestry area, a multi-parted roof construction was installed at two areas in the Kulawi-Valey. Both selected areas were divided into three roof-plots and three control-plots. This roof construction achieved a throughfall exclusion of approx. 67 % in the rain forest and of approx. 80 % in the Cacao-Agroforestry area. As a result of the artificial throughfall exclusion, the throughfall of the roof-plots in the rain forest showed decreases of nutrient-fluxes of about 46 - 58 % compared to the control-plots. In the Cacao-Agroforest decreases of about 75 - 79,2 % were observed. The litter of Gliricidia sep. showed significant higher C-, N-, P- und K+ input fluxes on the control-plots than at the roof plots. However, for the entire litter input the part of Gliricidia sep. was only prevailed for N and K+, otherwise the influences of Theobroma cacao were dominant. Neither the rain forest nor the Cacao-Agroforest area showed higher litter fall as a indicator for a dry-stressed ecosystem stand. During the rain exclusion period the litter layer of the Agroforest roof-plots showed significant higher C-, N- and P-stocks and reduced decompositions and mineralisation rates compared to the controls. In addition, C, N, P and K+ on the roof plots posses a 30 - 50 % longer turnover time which likely was the result of a decreased microbial decompostion and reduced nutrient release (C, N, P and K+) due to water stress. In opposite to the higher nutrient stocks in the litter layer, the nutrient stocks (C, P, K+) of the upper mineral soil layer (0 - 10 cm) were significant lower on the roof-plots. Because of the extremly dry litter layer, decomposition and mineralisation processes were hampered and only a local nutrient release was possible (via stem flow, roof holes). Neither in the rain forest nor in the Agroforest area the artificial roof construction showed an effect on the deeper soil layer. At both sites, the nutrient fluxes of all investigated parameters were highest in the litter layer and upper soil horizons reflecting the important source function of these two compartments within both ecosystems. Nutrient fluxes in the rain forest and Cacao stands were characterised by a large proportion of DON in DNb. Besides both sites showed higher DON-fluxes in deeper soil horizons than in the upper soil, because of the nitrogen release and delivery from the soil matrix. The high proportion of DON in DNb point to a N-limited system. In comparison to the DNb/DON ratio, DOP possesses a large part in Ptot. This fact and the litter N:P ratios marked a P-limitation for both locations (rain forest: 21,8, Cacao-Agroforest: 13,3). Especially in the Cacao-Agroforest a high release of Ca2+ , Mg2+ was observed, which is a result of cation leaching from the soil matrix of the Hypereutric Cambisol. Despite significant effects during the artificial throughfall exclusion, there was no indication of enhanced decomposition and mineralisation rates for DOC, DNb, Ptot, K+ and the other nutrients caused by the subsequent rewetting phase in the Cacao-Agroforest. Consequently, the nutrient mobilization during the rewetting period was too low to surpass the biological nutrient demand and the absorption capacity of soil matrix to set off an increased nutrient export with seepage water. This could be attributed to the intensified nutrient up-take of the microbial biomass and the plant roots in this highly productive ecosystems, or to a dilution effect caused by the high rainfall amounts entering the soil after the roof removal. In addition to the investigations on the nutrient dynamics in the replicated throughfall exclusion experiment, a second aim was to study the development of the soil nitrogen dynamic in view of diverse cultivation measures like variation in nitrogen fertilization (0 and 100 kg urea ha-1) as well as weeding (2-, 6-monthly). Before (2006) and after (2008) the experimental manipulations, soil samples (0 - 10 cm) were taken from the intensively and less- intensively treated subplot forming part of 44 plots each in the Kulawi- and Palolo-Valey. Analysis results neither showed differences between intensively and less-intensively treated subplots in 2006 and 2008. Although the soil compartment is capable to buffer smaller changes in water deficits and nutrient losses, the field results showed that with intensified throughfall reduction not only the water supply but also the release and availability of nutrients were significantly diminished, potentially leading to nutrient deficiency symptoms in forest ecosystems and Agroforestry systems in the long run.