New approaches to retrieval and application of lidar-measured vegetation vertical canopy profiles.

摘要

Forests play an important role in the exchange of heat, momentum, water, and carbon between the land surface and the atmosphere. The rates of exchange mostly depend on the three-dimensional distribution of canopy material. Large-footprint waveform-recording laser altimeters (lidars) have demonstrated a potential to significantly improve remote estimates of vertical forest structure. It is a relatively new field of science, with a number of opportunities and challenges.; Presently available methods to analyze lidar data are based on models with simplifying assumptions, which lead to rather large errors in the retrieved variables. Standard assumptions include neglecting the effects of multiple scattering, unrealistic representation of canopy structure, and ignoring foliage clumping. Moreover, the use of lidar-recorded waveforms is mostly limited to aboveground biomass estimation studies. The utility of this data to other fields of science is not well-investigated and poorly understood.; The goal of the research presented in this dissertation is to improve current methods used for lidar waveform processing and to develop new applications. For lidar data, this research utilized data from SLICER (Scanning Lidar Imager of Canopies by Echo Recovery), an air-borne instrument extensively used in 1994--1997. The research consisted of three separate tasks. First, a physical model describing the propagation of lidar signals through vegetation was developed on the basis of time-dependent radiative transfer theory. This model was used to evaluate the effects of multiple scattering, which were shown to cause a significant magnification of the amplitude of the reflected signal, especially that originating from the lower portion of the canopy. Second, a new application for lidar data was developed. In particular, canopy height profiles (CHPs) retrieved from lidar waveforms were used for modeling gross primary production (GPP) of deciduous forests. Comparison with other photosynthesis models showed that accounting for a vertical canopy profile might lead to more accurate estimates of GPP. Third, the current algorithm for retrieval of CHPs of deciduous forests was modified to be applicable to coniferous canopies. The modifications account for clumping of needles into shoots, shoot orientation and effects of multiple scattering, and result in the retrieval of more realistic CHPs.