Estimation of vegetated land surface properties in BOREAS area using daily MAIAC-based MODIS images

We recently developed a framework to estimate vegetated land surface properties in BOREAS area. The properties include leaf area index (LAI), leaf chlorophyll content, leaf water, leaf dry matter, the fraction of photosynthetically active radiation (PAR) absorbed by chlorophyll (FAPARchl) of a canopy, light use efficiency (LUE), gross primary productivity (GPP) and phenology. We used daily MODerate resolution Imaging Spectroradiometer (MODIS) image produced through an advanced atmospheric correction algorithm, i.e., the Multi-Angle Implementation of Atmospheric Correction (MAIAC) algorithm. MAIAC uses gridded MODIS multitemporal images to mask clouds and simultaneously retrieve atmospheric aerosol and surface bi- directional reflectance. Conceptually, only PAR absorbed by chlorophyll is used for photosynthesis, and therefore, there is a need to quantify FAPARchl. A coupled leaf-canopy radiative transfer model (PROPSECT+SAIL2, PROSAIL-2) with daily MODIS data as input was inverted with a Markov Chain Monte Carlo (MCMC) method (the Metropolis algorithm), which provides posterior distributions of retrieved variables. We ran the PROSAIL-2 model for ten sites in BOREAS area. Our study shows that it is important to make sure the input data of the PROSAIL-2 model are covering same target area. One way is to scale up the MODIS acquisitions. The leaf optics changes during plant growing season and should not be assumed as constants for a whole growing season. We found that biogeochemical models that use fraction of PAR absorbed by canopy (FAPARcanopy) and fraction of PAR absorbed by leaf (FAPARleaf) in estimating gross and net primary production are likely to overestimate the amount of PAR used in photosynthesis process, which represents one large source of error and uncertainty of estimation of productivity. This study also investigated the potential of using MODIS ocean band reflectance to estimate LUE over the ten sites. We may use only meteorological PAR and MODIS images to estimate GPP.