Assessment of the MODIS LAI and FPAR algorithm: Retrieval quality, theoretical basis and validation

Green leaf area index (LAI) measures the amount of foliage in a vegetation canopy and strongly influences many ecological processes, including the fraction of photosynthetically active radiation (FPAR) absorbed by the canopy. Since these two variables are required in most modeling studies of vegetation and climate, they are operationally derived from measurements of the Moderate Resolution Imaging Spectroradiometer (MODIS) aboard the NASA TERRA spacecraft. The objectives of this research are to evaluate the MODIS LAI and FPAR algorithm in terms of retrieval quality, theoretical basis and validation.; Investigation of the performance of the algorithm as a function of uncertainties in input spectral reflectances indicates that knowledge of these uncertainties is critical for the retrieval of biophysical parameters of highest possible quality. Neglecting this information can cause destabilization of the retrieval process, while its use can increase the number of high quality retrievals by 10–20%.; Assessment of the parameterization of the algorithm in light of the law of energy conservation indicates that spectra of soil reflectance and single scattering albedo combined with canopy interception, transmittance and their collided portions at a fixed reference wavelength are sufficient to simulate the spectral response of a vegetation canopy to incident solar radiation. All of these are measurable and satisfy a simple system of equations. This conclusion is demonstrated by developing an approach to retrieve these parameters from hyperspectral data collected during a field campaign in Finland. This approach also provides a new way to interpret hyperspectral data.; Investigation of the relationship between field data on LAI and 30m Landsat Enhanced Thermal Mapper plus (ETM+) images indicates that comparisons at the patch level are more reliable than the pixel level. The MODIS algorithm, adjusted to fine resolution, generally overestimates LAI due to influence of understory vegetation. Comparisons at both fine and coarse resolutions indicate the need for improvements in the algorithm for needleleaf forests. An improved correlation between field measurements and the Reduced Simple Ratio (RSR) suggests that the shortwave infrared (SWIR) band may provide valuable information for needleleaf forests.