| Quantitative estimates of forest chlorophyll content from hyperspectral remote sensing are of great use for terrestrial carbon cycle modeling and sustainable forest management. Open forest canopies present a big challenge for the separation of the effects from canopy structure and leaf optical properties, and thus the retrieval of biochemical parameters. Process-based algorithms were developed to estimate the chlorophyll content of broadleaves and needleleaves from hyperspectral measurements.; Field experiments were conducted from 2003 to 2004 near Sudbury and Haliburton, Ontario, to collect canopy structural, leaf biophysical and biochemical data. Experiments show that optical properties and biochemical contents of broadleaves change with the growing season and canopy height. Needleleaves from different sites, age classes, and branch orientations demonstrate different visible optical properties in relation to their chlorophyll contents. A process-based radiative transfer model PROSPECT was modified to retrieve leaf chlorophyll content from measured leaf spectra. For broadleaves, leaf thickness was introduced to consider the seasonal and canopy-gradient variation in light absorption. The accuracy of chlorophyll retrieval is increased from 67% to 91%. For needleleaves, the effects of needleleaf width and thickness, and geometrical effects of leaf-holding devices on spectra measurements were taken into account. These modifications improve the accuracy of chlorophyll retrieval from 31% to 59%.; Correct exposure for digital hemispherical photographs is crucial for estimating canopy structural parameters. A photographic exposure theory was tested for different forest types with various canopy closures and under different sky conditions. The exposure method improves the estimates of leaf area index by 40% in comparison with commonly used automatic exposure.; The effects of canopy structure on optical remote sensing signals were investigated using the geometrical-optical model 4-Scale. A look-up-table approach was developed to provide the probabilities of viewing sunlit foliage and background components, and a spectral multiple scattering factor as functions of LAI, and solar and view zenith angle. Leaf reflectance spectra and chlorophyll content were retrieved from the hyperspectral Compact Airborne Spectrographic Imager (CASI) images with a root mean square error of 4.34 mug/cm 2 for needleleaf species. LAI was retrieved and chlorophyll content was mapped using CASI imagery. |
