| dc.description.abstracteng | Oil palm cultivation has vastly expanded in the last decades and led to high economic returns, but has also induced environmental problems including reduced biological diversity and impaired ecosystem services. One potential way of reconciling economic and ecological needs would be to foster oil-palm agroforests by planting native tree species with the objective to enhance biodiversity and ecosystem services. In Sumatra, Indonesia, such an oil palm agroforestry experiment was established and a series of studies evaluated tree performance, oil palm yields and biodiversity changes. In this context, spatial information was relevant as the survival of planted trees may be influenced by neighborhood or the occurrence of specific taxa in experimental treatments may be influenced by the distance of source habitats. Drone-based assessments offer opportunities to support such spatial-structure related ecological studies. In this study, I used low and high flying drones (1) to analyze crown metrics of trees and palms to predict plant water use, (2) to study canopy cover in oil palm agroforest and the effect of oil palm canopy cover on tree mortality, and (3) to assess land use types surrounding the experiment. Objectives 1 and 2 were addressed with an octocopter drone, flying approximately 40 m above ground; while objective 3 was addressed by a fixed-wing drone, flying approximately 300 m above ground.
For study objective 1, I collaborated with a colleague who measured the water-use rates of individual trees and palms within the agroforestry experiment. Transpiration is often estimated from direct water-use measurements in a limited number of plants and then scaled up to the stand-level by using plant size related variables for the remaining plants. Presently, drone-based methods offer new opportunities for plant size assessments. We tested crown variables, derived from drone-based photogrammetry, for predicting and scaling plant water use. Aerial images were taken from an octocopter equipped with an RGB camera and the structure-from-motion approach was used to compute several crown variables including crown length, width and volume. Crown volumes for both palms (69%) and trees (81%) explained much of the observed spatial variability in water use; however, the specific crown volume model differed between palms and trees and there was no single linear model that fit both. With respect to trees, crown volume explained more of the observed variability than stem diameter, and in consequence, uncertainties in stand level estimates resulting from scaling were largely reduced. For oil palms, an appropriate whole-plant size related predictor variable was currently not available. In conclusion, we consider drone-derived crown metrics very useful for scaling up to stand-level transpiration from single plant water use.
For study objective 2, using a drone-based analysis, we compared oil palm canopy conditions in thinned and non-thinned plots and also examined how oil palm canopy cover affected the mortality rates of planted tree species. Three years after planting, canopy cover was assessed by drone-based photogrammetry using the structure-from-motion technique. Additionally, these surveys were augmented with tree positions and mortality rates recorded by colleagues. Drone-derived canopy cover was highly correlated with ground-based hemispheric photography along the equality line, indicating the usefulness and comparability of the approach. Canopy cover was further partitioned between oil palm and tree canopies. Oil palm canopy cover was then extracted at the level of individual trees and combined with ground-based mortality assessment for all 3819 planted trees. For three tree species (Archidendron pauciflorum, Durio zibethinus, Shorea leprosula), probability of mortality during the year of the study were dependent on the amount of oil palm canopy cover. Thinning of oil palms before tree planting created a more open and heterogeneous canopy cover. We regard the drone-based method for deriving and partitioning spatially explicit information as promising for many questions addressing canopy cover and the management of agroforestry systems.
For study objective 3, I assessed land use types in the landscape surrounding the oil palm agroforestry experiment. In two consecutive years, 2015 and 2016, 1121 ha were analyzed where the experiment is in the center. Beginning in 2015, oil palm covered 81% of the area and the remaining 19% comprised other land use types including bare soil, rubber plantation, secondary forest, orchard, fallow, water and urban. During this time frame, the area under oil palm continued to expand. In just one year, 50% of bare soil (47 ha), 27% of fallow (10 ha), 18% of secondary forest (10 ha) and 15% of rubber plantation (3.2 ha) were converted to oil palm plantation. We found oil palm cultivation in large and continuous tracts with very little fragmentation. Secondary forests were found in relatively small patches, some of which occurred in close proximity to the northern plots of the agroforestry experiment, and possibly influenced the occurrence of some mobile taxa. We conclude that it was feasible to derive detailed land cover maps from drone-based assessments that enable the detection of even small-scale land use change in the oil palm landscape.
Overall, drone-based assessments can play vital roles for ecological studies in oil palm landscapes and agroforests. The data from the photogrammetric approach, such as SfM point clouds and aerial imagery, can derive quality assessments of canopy structures. Specifically, such assessments can support the prediction of tree and oil palm water use, transpiration at the stand level, and the evaluation of oil palm canopy conditions between thinned and non-thinned plots. According to these drone-based assessments, oil palm canopy cover appeared to influence the mortality of some tree species in the oil palm agroforestry plots. In addition, drone-based monitoring could be used to detect the expansion of oil palm cultivation around the oil palm agroforest. Other applied workflows in forestry and agricultural studies can integrate our procedures to attain spatial information. I conclude that drone-based assessment is a well-suited tool for monitoring oil-palm agroforests, in both small scale and large areas. | de |