Einfluss der Waldkonversion auf den Wasserhaushalt eines tropischen Regenwaldeinzugsgebietes in Zentral Sulawesi (Indonesien)

In the first phase of the collaborative research project Stability of rainforest margins financed by the DFG (German Research Council) a experimental catchment and its sub-catchments have been instrumented with extensive measurement instruments to investigate the influences of land use changes on the water and nutrient fluxes at the rainforest margins area along the border of the Lore Lindu National Park in Central Sulawesi (Indonesia). The research design allows the investigation of the long-term influence of gradual land use change due to forest conversion by smallholders on the water and nutrient balances. In the first phase of the project three main goals and objectives were pursued in the research area: (1) The establishment of a spatial and temporal high resolution water balance for the research catchment and its subcatchments. (2) The application of the physically based water balance model WASIM-ETH to the research area Calculation of water balances for three land use change scenarios (forested catchment, agricultural land use, cocoa plantation). (3) In the first part of the study the pattern of precipitation, runoff and storage change has been investigated from September 2001 to February 2003.The experimental results can be summarized in the following way: (1) In 2002 an area precipitation of 2360 mm has been observed. The annual movement of the intertropical convergence zone creates two wet and dry seasons. Precipitation demonstrates an enormous spatial and temporal variation and increases with elevation. (2) Runoff was app. 966 mm in the year 2002. Due to the rainy periods a seasonal runoff dynamic was observed. The catchment exhibits a very fast precipitation-runoff reaction. The statistical parameters of the runoff behaviour show a wide range between low flow and high flow. A typical behaviour for tropical mountainous catchments. (3) The dynamics of soil moisture and soil suction is influenced by the seasonal precipitation pattern. During dry seasons soil water and groundwater storage contribute to evapotranspiration and base flow. Wet seasons are important for storage recharge. Forested areas show a faster depletion of soil moisture during dry periods in comparison to agricultural and agroforestry areas. (4) An evapotranspiration from 1350 to 1518 mm is calculated from the water balance equation. In the second part of the study the water fluxes of the research area was modelled with the physically based water balance model WASIM-ETH.The application of the model requires extensive spatial and temporal data, which are available due to the experimental work in the catchment. The modelling results can be summarized: (1) A good agreement between the simulated and measured water balance equation was reached. Using a multiple response validation approach runoff, soil moisture, evapotranspiration (from the water balance equation) and separated runoff components could be used for model validation. (2) The model allows the separation of the different evapotranspiration processes. The real evapotranspiration in the catchment was app. 1450 mm in 2002. Due to the high annual precipitation amounts real evapotranspiration was merely reduced in comparison to the potential evapotranspiration. Interception was app. 28%. (3) The ability of WASIM-ETH to calculate the influence of land use changes on the water balance was examined by testing the sensitivity of changes in the land use parameterisation of the model. (4) Modelling in daily time steps is suitable for water balance considerations. Because of the fast precipitation-runoff reaction, further process-oriented modelling of the runoff generation process should be performed in hourly time steps.Finally a first assessment of the influence of land use changes on the water balance in the catchment was modelled. Three scenarios were investigated (a natural forest scenario, agroforestry scenario with cocoa plantation and a agricultural scenario with mixed land use). All scenarios clearly show influences on the water balance. (1) The forest scenario decreases annual runoff. Agroforestry and agriculture increase annual runoff. (2) Forest vegetation decreases runoff peaks. Agricultural land use increases runoff. (3) The increasing ground water table - as a result of reduced evapotranspiration and unchanged infiltration capacities - in the agroforestry scenario - leads to higher flow runoff in the dry season. The modelling results are consistent with other comparable experimental studies in the tropics. The advantage of a modelling approach is a fast assessment of possible water balance equation changes due to land use changes without long-term measurements. However a reliable parameterisation of the model is a prerequisite for reliable results.