Approaches to characterise forest structures for wind resource assessment using airborne laser scan data

The transformation of the energy system pushes the expansion of wind power in Germany towards the installation of new wind farms in complex and forest covered terrain. The primary objective of this thesis is to derive forest structure parameters in high resolution as input for flow simulations used in wind energy site assessment. For this purpose, airborne LiDAR data is processed, and a digital terrain model (DTM), as well as forest heights and forest densities, are determined. Airborne laser scanning (ALS) is a new technique for land surface applications. A fast development could be observed in recent years, and it is about to supersede the traditional ground-based measurement techniques. This bears new challenges for the forest parametrisation, especially the estimation of the vegetation density. In literature, different approaches to determine the plant area index (PAI), which is a measure of the canopy density, can be found. Three methods are analysed within this work: One method is based on the number of first echoes that can be recorded from the different canopy levels. The other two methods make use of the intensities of the echoes to estimate the transmittance of the canopy. The plant area density is assessed based on the Beer-Lambert law for light attenuation. Hence an exponentially falling correlation between the PAI and the number of ground echoes can be observed. Furthermore, it can be seen that the PAI results from the three methods differ significantly for locations predominantly covered with beech trees. For coniferous forest areas, the PAI results are more similar. A differentiated approach depending on the forest type is necessary. Hence, the variation of the PAI between the methods is implemented as an indicator for the identification of beech forests. The comparison of the three methods to each other, to photographs, and measurement data shows that the first return method used in most papers misses important information given with the intermediate echoes. In forests with homogeneous, dense canopies, first echoes can only be recorded from the crown. The two other methods which include all returns and their intensities achieve to describe the full density profile over the whole forest height. However, the PAI levels are significantly lower, especially in beech tree areas. This could cause an underestimation of the drag relevant for wind flow simulations. Deciduous forests have a higher density during summer. Since ALS flights are usually performed during leaf-off season, a method to derive the leaf-on PAI is developed. For the forest during winter, the method based on first returns is selected, because of the higher PAI level. Above that, previous studies showed that assuming a drag coefficient of 0.2, this approach gives reliable results. For the forest during summer, the PAI of the areas defined to be beech forest are raised by a leaf factor of 1.8 in order to account for the PAI increase during frondescence.