Dissertation: B. Koetz

Vegetation controls important ecosystem processes responsible for energy and mass exchanges within the terrestrial biosphere. A comprehensive characterization of the vegetation canopy is thus required to monitor the heterogeneous and dynamic terrestrial biosphere. Although Earth Observation provides detailed measurements of the Earth surface, it has been a challenge to produce reliable data sets of the Earth vegetation condition in its spatial distribution and change over time. Relative to the heterogeneous and three-dimensional nature of vegetation the generally available amount of independent Earth Observation measurements is limited. The problem of deriving vegetation characteristics from Earth Observation becomes consequently underdetermined. For an improved retrieval of vegetation characteristics the number of independent information sources needs to be increased. However, detailed measurements of the Earth Observation systems imaging spectrometry and LIDAR provide such independent information. The information dimensions observed by the two sensor systems contain data relevant to different aspects for a comprehensive characterization of vegetation canopies. The information dimension observed by LIDAR provides direct measurements of the vertical canopy structure including the canopy height. Whereas, the spectral information dimension provided by imaging spectrometers contains information about biophysical as well as biochemical canopy properties. The presented dissertation focuses on the combined exploitation of the independent Earth Observation measurements of imaging spectrometry and LIDAR based on radiative transfer modeling. The approach of radiative transfer modeling explicitly describes the relationship between the remotely sensed signal and vegetation characteristics. A Radiative Transfer Model (RTM) considers the incident radiation, sensor specifications and physical processes that govern the radiative transfer within the canopy. Two such physically based RTM are employed to separately describe the signals of an imaging spectrometer and a large footprint LIDAR. The combined inversion of these RTM’s presents an efficient methodology for a synergistic exploitation of the independent information dimension obtained by the two Earth Observations systems. The developed methodology ensures a retrieval algorithm of increased robustness, but also provides an enhanced vegetation canopy characterization. Results present reliable and quantitative estimates of canopy characteristics including the horizontal and vertical canopy structure as well as the foliage biochemistry over coniferous forest stands. The dissertation shows thus the potential of independent information dimensions of Earth Observation and RTM inversion for the retrieval of quantitative vegetation characteristics.