Retrieval of vegetation biochemical and biophysical parameters using radiative transfer models and RapidEye imageries in different biomes

Paper 1728 - Session title: Land Posters.

Mapping and monitoring vegetation biophysical and biochemical parameters is valuable for biodiversity modeling, but also for a diversity of other applications including agricultural, ecological, and meteorological studies. The spatial and temporal distribution of vegetation parameters are used as inputs in models quantifying the exchange of energy and matter between the land surface and the atmosphere. Among the many vegetation parameters, leaf area index (LAI) and chlorophyll are of prime importance as they provide vital information for biodiversity assessment and have been proposed as Essential Biodiversity Variables (EBVs). Significant efforts to estimate and quantify these parameters using Radiative Transfer Models (RTM) have been carried out in the last two decades. However, most of these studies rely on hyperspectral measurements. Although hyperspectral data have demonstrated to be accurate and suitable for retrieval of vegetation parameters, they are usually expensive and not frequently available. New generation of multispectral sensors with high spatial resolutions such as Sentinel-2 and RapidEye which provide spectral information at the red-edge region have provided further opportunities for estimating these parameters. As part of the ESA: DUE, Innovators-III funded project RS4EBV (Remote Sensing for Essential Biodiversity Variables), the retrieval of LAI and chlorophyll is investigated utilizing RapidEye images and different RTMs. Several RapidEye images were acquired between March and September 2015, for each of the three study sites of the project, i.e. 1) the Bavarian Forest National Park, 2) the National Park of Schiermonnikoog in the north of the Netherlands, and 3) the North Wyke Farm Platform in Devon, UK. In situ measurements of a large number of biophysical and biochemical parameters including LAI and chlorophyll were collected concomitant with the time of image acquisitions. The widely used canopy radiative transfer models: PROSAIL (SAILH and the PROSPECT) and INFORM (Invertible Forest Reflectance Model), were investigated for retrieval of LAI and chlorophyll in grasslands (including the saltmarsh) and forest study sites, respectively. The two RTMs were first parameterized based on the spectral band settings of RapidEye. Consequently, large look-up tables (LUTs) were generated for each study sites accounting for available prior information related to the distribution and range of the vegetation characteristics in each site. The LUTs were then inverted using the spectral reflectance obtained from the images. To assess the performance of the model inversion and analyze the suitability of the models, the normalized RMSE and R2 between independent in situ measurements and estimated parameters were used. Our results will demonstrate the potential and drawbacks of model inversion for estimating vegetation biophysical and biochemical parameters at three different types of biomes in Europe using new generation of multispectral satellite data.