Validation of MODIS F/sub PAR/ products in boreal forests of Alaska

Moderate Resolution Imaging Spectroradiometer (MODIS) fraction of photosynthetically active radiation absorbed by vegetation (F_PAR ) products covering the boreal forest of interior Alaska were analyzed and compared with field measurements of canopy light harvesting over a multiyear period (2001 to 2004), as well as to high-resolution F _PAR maps derived from IKONOS and Landsat ETM+ imagery. The spatial variability of F_PAR within the MODIS products was examined by incorporating the field measurements and aggregating the high-resolution F_PAR maps to the MODIS scale. Characterization of the temporal accuracy of the MODIS F_PAR products was conducted through comparisons with continuously operating canopy light interception measurements. The MODIS product tended to overestimate F_PAR relative to both ground-based measurements and Landsat-derived estimates of F_PAR, particularly in the more sparsely vegetated burn scars on which we focus related research, but adequately captured seasonal variability associated with vegetation phenology. A combination of canopy closure and ground cover vegetation was identified as the source of most of the discrepancies between the MODIS estimated and field measured F_PAR values. Neither the field measurements nor the high resolution image F_PAR maps based on the field measurements characterized the light environment of the ground cover vegetation (i.e., <10 cm height); thus, absolute validation of the MODIS products was incomplete - despite the extensive spatial and temporal characterization of F_PAR dynamics in the study region. We discuss these results, explore some other possible sources for observed differences between the MODIS, field, and high-resolution F_PAR maps, and consider possible ways to address these issues in future work