Identification of grazed and mown grasslands using a time series of high-spatial-resolution remote sensing images

Abstract

In many regions, a decrease of grasslands and change in their management can be observed with agriculture intensification. Hence, the evaluation of grassland status and management in farming systems is a key-issue for sustainable agriculture. However, inventory of grassland surfaces in agricultural areas is very incomplete and the spatiotemporal distribution of their management is still largely unknown. The objective of this study is to identify mown and grazed grasslands from a time series of high spatial resolution images acquired in 2006 on an experimental watershed located in Brittany, France. The coupling of two radiative transfer models (PROSPECT+SAIL) has been applied to the remote sensing images to derive biophysical variables, in order to identify grassland management. Then, based on training samples, the classification of the temporal profiles extracted from the images was performed using three different methods with increasing automation: a knowledge-based classification, a k-nearest neighborhood and a decision tree procedure.