Estimación de la fAPAR sobre la Península Ibérica a partir de la inversión del modelo de transferencia radiativa 4SAIL2

Abstract

This work aims to the estimation of fAPAR over the Iberian Peninsula using MODIS data. First, the 4SAIL2 and PROSPECT radiative transfer models have been used to simulate a data set of reflectance and fAPAR. Second, an artificial neuronal network (ANN) has been trained using the simulated data and finally, it has been inverted to derive fAPAR estimates over the Iberian Peninsula from MODIS reflectances images. Moreover, the impact that the observation and illumination configuration have on the fAPAR estimates has been assessed. The fAPAR estimates from MODIS have been compared with other validated fAPAR products. The results confirm an overall error around the user requirements (0.1) when the fAPAR estimated from the (PROSPECT+4SAIL2+Nadir) combination is compared with the selected products. This combination is proposed as an alternative to estimate fAPAR over the Iberian Peninsula due to the ability to characterize different land cover types as well as the high intra-annual variability of particular canopies.