Performance of the MISR LAI and FPAR algorithm: a case study in Africa

The Multi-angle Imaging SpectroRadiometer (MISR) instrument is designed to provide global imagery at nine discrete viewing angles and four visible/near-infrared spectral bands. The MISR standard products include green leaf area index (LAI) of vegetation and fraction of photosynthetically active radiation absorbed by vegetation (FPAR). These parameters are being routinely processed from MISR data at the Langley Atmospheric Sciences Data Center (ASDC) since October 2002. This paper describes the research basis for transitioning the MISR LAI/FPAR product from beta to provisional status. The quality and spatial coverage of MISR land surface reflectances that are input to the algorithm determine the quality and spatial coverage of the LAI and FPAR products. Therefore, considerable efforts have been expended to analyze the performance of the algorithm as a function of uncertainties of MISR surface reflectances and to establish the convergence property of the MISR LAI/FPAR algorithm, namely, that the reliability and accuracy of the retrievals increase with increased input information content and accuracy. An additional objective of the MISR LAI/FPAR algorithm is classification of global vegetation into biome types—information that is usually an input to remote sensing algorithms that use single-angle observations. An upper limit of uncertainties of MISR surface reflectances that allows discrimination between biomes, minimizes the impact of biome misidentification on LAI retrievals, and maximizes the spatial coverage of retrievals was estimated. Algorithm performance evaluated on a limited set of MISR data from Africa suggests valid LAI retrievals and correct biome identification in about 20% of the pixels, on an average, given the current level of uncertainties in the MISR surface reflectance data. The other 80% of the LAI values are retrieved using incorrect information about the type of biome. However, the use of multi-angle data minimizes the impact of biome misidentification on LAI retrievals; that is, with a probability of about 70%, uncertainties in LAI retrievals due to biome misclassification do not exceed uncertainties in the observations. We also discuss in depth the parameters that characterize LAI/FPAR product quality—such as quality assessment (QA) that is available to the users along with the product. The analysis of the MISR LAI/FPAR product presented here demonstrates the physical basis of the radiative transfer algorithm used in the retrievals and, importantly, that the reliability and accuracy of the retrievals increase with increased input information content and accuracy. Further improvements in the quality of MISR surface reflectances are therefore expected to lead to LAI and FPAR retrievals of increasing quality. D 2003 Elsevier Inc. All rights reserved.