Summary Results from:

Investigation of product accuracy as a function of input and model uncertainities: Case study with SeaWiFS and MODIS LAI/FPAR algorithm
As they relate to the validation of MOD15

Authors: Wang, Y., Tian, Y., Zhang, Y., El-Saleous, N., Knyazikhin, Y., Vermote, E., Myneni, R.B.

Source: Remote Sensing of Environment, 78: 296-311, 2001

Link to: Access Publication


The derivation of vegetation leaf area index (LAI) and the fraction of photosynthetically active radiation (FPAR) absorbed by vegetation globally from the Sea-Viewing Wide Field-of-view Sensor (SeaWiFS) multispectral surface reflectances using the algorithm developed for the moderate resolution imaging spectroradiometer (MODIS) instrument is discussed here, with special emphasis on the quality of the retrieved fields. Uncertainties in the land surface reflectance and model used in the algorithm determine the quality of the retrieved LAI/FPAR fields. The in-orbit radiances measured by space-borne sensors require corrections for calibration and atmospheric effects, and this introduces uncertainty in the surface reflectance products. The model uncertainty characterizes the accuracy of a vegetation radiation interaction model to approximate the observed variability in surface reflectances. When the amount of spectral information input to the retrieval technique is increased, not only does this increase the overall information content but also decreases the summary accuracy in the data. The former enhances quality of the retrievals, while the latter suppresses it. The quality of the retrievals can be influenced by the use of uncertainty information in the retrieval technique. We introduce a stabilized uncertainty, which is basic information to the retrieval technique required to establish its convergence; that is, the more the measured information and the more accurate this information is, the more reliable and accurate the algorithm output will be. The quality of retrieval is a function of the stabilized uncertainty whose accurate specification is critical for deriving biophysical surface parameters of the highest quality possible using multispectral land surface data. The global LAI and FPAR maps derived from SeaWiFS multispectral surface reflectances and uncertainty information, as well as an analysis of these products is presented here.