By Arnout van Raaij
Abstract:
The Leaf Area Index (LAI) is one of the basic quantities used in estimation of net primary production and modeling of terrestrial carbon cycle processes. Several methods have been developed to measure the LAI. One of the currently most widely used methods is the remote sensing based method. Because satellite systems produce large extent data and a spatially continuous representation of the land surface, this method is nowadays preferred over other methods. One of the remote sensing systems currently used is the MODIS device on board the spaceborne Terra and Aqua platforms. Until now, most analyses of MODIS LAI data were based on site or biome specific average values and the changes of these averages over time. In this research, the possibility of using a MODIS-based probability density function (PDF) to analyze temporal variations of the LAI was explored. Three biomes were selected for analysis: boreal coniferous forest, temperate coniferous forest, and temperate broadleaf and mixed forest. Results showed a large probability peak around LAI = 1 during winter months, which shifts during growing season towards higher LAI values around LAI = 2.4. A second peak appears during summer around LAI = 5.2. In between these two probability peaks, a dip fixed at LAI = 4.3 is present. This dip is present in the PDF’s of all summer months of all three studied biomes and is thought to be caused by the MODIS radiative transfer algorithm, the algorithm upon which most LAI values in the MODIS dataset are based. Although it was found that the MODIS LAI dataset describes temporal variation of LAI reasonably well, it is suggested that the algorithm should be further improved to eliminate the algorithm’s double peak behavior during summer.