Changes in eco-hydrological functioning after tropical rainforest transformation to rubber and oil palm plantations

This work presents findings from eco-hydrological research carried out in Jambi, Sumatra, Indonesia in the framework of the CRC990 (Sub-Project A02). Our results point to substantial changes in central ecosystem water fluxes after tropical rainforest transformation to monoculture plantations, which affects the availability of water at the landscape scale and thus impacts on the well-being of rural communities. Rainforest transformation to agricultural systems is generally expected to alter ecosystem water cycles at local and regional scales. Transpiration, i.e. water use by plants, is central to the hydrological cycle and biosphere-atmosphere feedback mechanisms. In Amazonia, e.g., substantial reductions in land-atmosphere water fluxes after large-scale rainforest transformation to pasture altered precipitation patterns. The hot spot of current rainforest transformation is the ‘Maritime Continent’ of Indonesia. In contrast to Amazonian rainforest transformation to pasture or soy-bean, rainforests in Indonesia are largely being transformed to rubber (Hevea brasiliensis Müll.) and oil palm (Elaeis guineensis Jacq.) monocultures. Scattered in between, locally some traditional agroforestry systems (e.g. ‘jungle rubber’) remain in the landscape. They are considered a glimmer of hope regarding the balancing of economics and ecosystem services. For the ‘maritime’ tropics, eco-hydrological consequences of rainforest transformation to the prevalent productive systems (i.e. oil palm and rubber) have not yet been convincingly addressed. To assess changes in the magnitude of the central water flux of stand transpiration as well as in its spatial and temporal variability after rainforest transformation, we simultaneously studied four tropical lowland land use types in Jambi, Indonesia with a sap flux technique. Our 39 study sites were located in reference forests, ‘jungle rubber’ agroforests and rubber and oil palm monocultures. The main objectives were (1) to verify and if necessary adjust a broadly used sap flux technique for measurements on different species in a lowland landscape in Jambi, Indonesia, (2) to shed first light on the thus far relatively unknown water use characteristics of oil palm as a globally rapidly expanding crop species, and (3) to assess the consequences of the continuing ‘maritime’ rainforest transformation to monoculture plantations for landscape-scale eco-hydrological functioning. First, a commonly applied sap flux technique, the thermal dissipation probe method (TDP, Granier 1985), was tested and adjusted for measurements on several tropical monocot and dicot species. Experiments confirmed that the method should be calibrated when working on previously unstudied (monocot) species. Using the original Granier calibration equation, good agreement was found between TDP derived water use rates and reference gravimetric measurements for four tropical tree species including rubber, but substantial deviations became apparent for oil palm leaves. We thus derived new, oil-palm specific parameters for the TDP calibration equation. Based on sap flux measurements on 56 leaves on ten oil palms, we derived a sampling scheme for soundly estimating stand-level transpiration rates of oil palms including error margins. Statistically-derived ‘optimal’ sample sizes suggest measurements on a per-site minimum of 13 leaves on four different palms in an un-stratified scheme, which results in sample-size related estimation errors of stand transpiration of 13%. In tree based land use types, we measured sap flux in the North and South of the trunks of six (rubber) and eight trees (jungle rubber, forest) per site, which was associated with potential estimation errors of less than 10% and up to 35%, respectively. We subsequently focused on investigating the thus far little explored eco-hydrological characteristics of oil palms with the newly-established measurement scheme. We studied effects of stand characteristics on transpiration along an age gradient in 15 oil palm plantations in Jambi, Indonesia. Stand transpiration rates increased almost eight-fold from an age of two to an age of five years and then remained constant with further increasing age, but were highly variable among medium-aged oil palm plantations. Other water fluxes than transpiration (e.g. evaporation) contributed substantially and variably to evapotranspiration (eddy covariance derived), reducing a 12-fold difference between the stands transpiring at the lowest and highest rates, respectively, to a less than two-fold difference in evapotranspiration. Our results suggest that both transpirational and total evapotranspirational water fluxes from oil palm plantations can be substantial (i.e. as high as from rainforests) under certain site or management conditions. Relatively high water use rates of oil palm plantations were confirmed by an assessment of sap flux derived transpiration rates of 32 sites in four land use types in Jambi (forest, jungle rubber, rubber, oil palm). Stand transpiration rates of rubber plantations were much lower than those of rainforests and jungle rubber agroforests, partly due to (partial) leaf shedding. Oil palm transpiration, on the other hand, was almost as high as in the more ‘natural’ land use types despite e.g. a much lower biomass per hectare than in forests. Additionally, the transpirational day-to-day response of oil palm was ‘buffered’ compared to tree based land use types, i.e. even pronounced fluctuations in micrometeorological conditions resulted in relatively low temporal heterogeneity of oil palm transpiration rates. The pronounced differences in eco-hydrological characteristics that we observed between oil palm and rubber plantations were found to play a key role in explaining periodically occurring local water scarcity in oil palm dominated landscapes, as was reported by villagers in a qualitative social study in Jambi. Soil erosion and thus reductions in soil water infiltration capacity were similar in oil palm and rubber plantations; landscape-scale water losses by run-off after pronounced precipitation were thus high in both plantation types. In contrast to rubber, however, oil palms had relatively high water use rates and transpired relatively constantly despite fluctuating environmental conditions. Paired with the reduced water storage capacity of the eroded soils under oil palms, their relatively high water use can lead to local water scarcity during pronounced dry periods, i.e. low streamflow and groundwater levels in oil palm dominated landscapes. Our work thus points to potentially severe and thus far neglected social- and eco-hydrological consequences of ‘maritime’ rainforest transformation to monoculture plantations, particularly in the case of transformation to oil palm