Validation Of Collection 5 Modis Lai Product By Scaling-Up Method Using Field Measurements

Leaf area index (LAI) is very often a critical parameter in process-based models of vegetation canopy response to global environment change. This paper made an assessment of the Collection 5 MODIS LAI product (MCD15A2) using field sample data in cropland areas. One of the major problems for validating MODIS LAI product using field measurements is the scale mismatch between ground "point" measurements and the MODIS resolutions. In heterogeneous areas, we need to transform field measurements to the scale of MODIS due to scale effect caused by the heterogeneity of land surface. In this study, we performed the scale transformation through fractal method. LAI was measured with the LAI-2000 plant canopy analyzer. The LAI-2000 measurements were multiplied by a clumping index to get true LAI values. The field data was related to 30-m resolution TM images using empirical methods to create reference LAI map. Fractal dimensions for each MODIS pixel were calculated using a triangular prism method based on reference LAI map. Then the field LAI values were upscaled to 1km spatial resolution using the fractal dimension theory. The MODIS LAI product validation results shown that, MODIS LAI are lower than the ground measurements without scale effect correction, but quite close to the upscaled field measured LAI. The conclusion is the fractal dimension theory can be used to solve the scale problem caused by spatial heterogeneity in LAI products validation.