B41H-05
Validation, Mapping and Application of MODIS fAPARchl for GPP Modeling

Thursday, 17 December 2015: 09:00
2006 (Moscone West)
Qingyuan Zhang, Universities Space Research Association Greenbelt, Greenbelt, MD, United States and Tian Yao, NASA Goddard Space Flight Center, Greenbelt, MD, United States
Abstract:
The climate is affected by the land surface through regulating the exchange of mass and energy with the atmosphere. The energy that reaches the land surface has three pathways: (1) reflected into atmosphere; (2) absorbed for photosynthesis; and (3) discarded as latent and sensible heat or emitted as fluorescence. Vegetation removes CO2 from the atmosphere during the process of photosynthesis, but also releases CO2 back into the atmosphere through the process of respiration. The complex set of vegetation-soil-atmosphere interactions requires that a realistic land-surface parameterization be included in any climate model or general circulation model (GCM) to accurately simulate canopy photosynthesis and stomatal conductance.
In this paper, we present a new MODIS product, fraction of PAR absorbed by chlorophyll throughout a canopy (fAPARchl), in three aspects: validation, mapping and application. The fAPARchl is retrieved through the coupled canopy-leaf radiative transfer model PROSPECT2 with surface reflectance of MODIS bands 1-7 at spatial resolution of 500 m. In order to validate the fAPARchl product, we design a new gridding approach that define the center of a selected site or field as the center of the grid and the MODIS observations of the grid is weighted by all overlapped observations from each orbit, respectively. Selected sites include evergreen needleleaf forests, deciduous broadleaf forests, mixed forests, crops, shrublands, and grasslands under various climate conditions. The fAPARchl product is mapped at site/pixel, local and regional scales across the United States and Canada. The fAPARchl is utilized to estimate GPP, and compared to tower flux GPP for validation. The GPP estimation performance with fAPARchl is also compared with the GPP estimation performance with MOD15A2 FPAR. The fAPARchl product is further implemented into ecological models and land-surface models to simulate vegetation GPP. This study shows the importance of the new gridding approach for validation, and concludes that fAPARchl is superior to MOD15A2 FPAR in GPP estimation. This study will be extended to more sites in other regions under various climate conditions in the future for continuing validation and applications of the MODIS fAPARchl product.