The potential of optical high resolution data for the assessment of leaf area index in East African rain forest ecosystems

Abstract

Operational standard products of biophysical variables, such as leaf area index (LAI) derived from satellite observations (e.g. Moderate Resolution Imaging Spectroradiometer (MODIS), Medium Resolution Imaging Spectrometer (MERIS) or VEGETATION), serve the need for adequate input data for quantitative modelling of ecosystem dynamics on regional to global scales. For the validation of those products and for local applications, higher resolution LAI maps based on field and high spatial resolution satellite data are crucial. As validation sites in tropical rainforests are rare, a study was undertaken in a moist semi-deciduous tropical rainforest in East Africa. In situ LAI measurements were carried out with digital hemispherical photography on 30 test sites. Spectral vegetation indices (SVIs) and texture measures calculated from top-of-canopy Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) and Satellite Pour l’Observation de la Terre (SPOT)-4 High Resolution Visible and Infrared (HRVIR) reflectance data were compared with respect to their ability to predict effective LAI. Regression analyses showed that ASTER data were generally better suited, with simple ratio (SR) performing best for early and intermediate forest stages and texture measures (grey level co-occurrence matrix (GLCM) variance) derived from SWIR information rendering superior results in later forest stages. Finally, a high resolution LAI map could be constructed with an accuracy of 0.39.