HR: 15:00h
AN: B22C-06 INVITED     [PDF]
TI: MODIS LAI and FPAR Product on Global, Regional and Local Scales
AU: * Knyazikhin, Y
EM: jknjazi@crsa.bu.edu
AF: Boston University, 675 Commonwealth Ave., #457, Boston, MA 02134 United States
AU: Yang, W
EM: ywze@crsa.bu.edu
AF: Boston University, 675 Commonwealth Ave., #457, Boston, MA 02134 United States
AU: Dong, H
EM: dh@crsa.bu.edu
AF: Boston University, 675 Commonwealth Ave., #457, Boston, MA 02134 United States
AU: Bin, T
EM: tanbin@crsa.bu.edu
AF: Boston University, 675 Commonwealth Ave., #457, Boston, MA 02134 United States
AU: Shabanov, N
EM:
AF: Boston University, 675 Commonwealth Ave., #457, Boston, MA 02134 United States
AU: Myneni, R
EM: rmyneni@crsa.bu.edu
AF: Boston University, 675 Commonwealth Ave., #457, Boston, MA 02134 United States
AB: An algorithm based on physics of radiative transfer in vegetation canopies for the retrieval of vegetation green leaf area index (LAI) and fraction of absorbed photosynthetically active radiation (FPAR) from MODIS surface reflectance data was developed, prototyped and is in operational production at NASA computing facilities since June 2000. This presentation is focused on the analysis of the of the LAI and FPAR retrievals as a function of time and spatial scale as detailed below. First theme covers analysis of the global MODIS LAI and FPAR products from July 2000 to December 2002, collections 1 and 3. About 70% of the total retrievals are obtained with the main radiative transfer based algorithm. Temporal compositing from 8-day to monthly further increases the frequency of main algorithm retrievals. The retrieved LAI and FPAR fields display expected features when analyzed by biomes and latitudes. The main algorithm fails as expected when input surface reflectance data have high uncertainties, especially under snow and cloud conditions. The analysis presented here reinforces the need for examining product quality flags accompanying the LAI and FPAR product before using these products in application studies. Second theme covers analysis of the LAI product at regional and local scales. We highlight the statistical nature of MODIS LAI and FPAR products arising from the relation between uncertainties in algorithm inputs and outputs, using Collection 3 MODIS LAI product. Two random variables impact the quality of retrieved LAI and FPAR fields at local scale- uncertainties in biome classification and surface reflectance measurements. To decrease impact of input uncertainties, averaging of LAI and FPAR product over an extended area is required to accumulate a sufficient number of pixels with high quality input. Further improvements in LAI and FPAR retrieval coverage and quality will require a better consistency between observed and simulated reflectances in spectral space.
DE: 0400 Biogeosciences
SC: Biogeosciences [B]
MN: 2003 Fall Meeting