Use of Sentinel-2 images for biomass assessment in extensive pastures in the Apennines (Central Italy)

The monitoring of forage production is one of the most important activities for establishing correct pasture management. Since above-ground biomass estimation with in-field traditional methods is costly in time and money, remote sensing techniques have been largely utilized and improved over the last decades to monitor grass growth and forage production. In this trial, we tested the potential of satellitebased vegetation indices (NDVI, SAVI, PVR, GLI, TVI, VARIGreen) in detecting actual pasture production, in terms of fresh biomass and dry matter over extensive grazing systems. Biomass was harvested from plots of 1m2-surface wide (8 replicates per each study site) in two areas of Tuscany (Central Italy). The vegetation indices were elaborated from Sentinel-2 satellite images, acquired in correspondence of the sampling dates, and then correlated to pasture production measured from ground surveys. Best results in estimating fresh and dry forage biomass were achieved respectively with NDVI (R2 = 0.59) and SAVI (R2 =0.49) but also other indices calculated from bands within the visible spectrum showed similar results in above ground biomass estimation.