Regresar

Pasture Monitoring Applying Normalized Difference Vegetation Index (NDVI) Time Series with Sentinel-2 and Landsat 8 Images, to Improve Milk Production at Santa Mónica Farm, Imbabura, Ecuador

Abstract:

The soils have had an increasing pressure due to the intensification of their use for agriculture, forestry, grazing and urbanization. In this way, the implementation of good practices for sustainable soil management are essential to reverse their tendency to degradation as preventive measures and so, guarantee food security and protect the provision of different ecosystem services associated with the soil. The advent of the Sentinel and Landsat satellite programs provide free data sets with good spatial and temporal resolution that can be a valuable source of information for monitoring pasture resources. In order to evaluate this type of techniques, a time series (TS) was generated with images of the Landsat 8 (L8) OLI (Operational Land Imager) sensor and a time series with images of the Sentinel-2 (S2), MSI (Multispectral Imager) sensor to determine the best results in the quantification of changes in the coverage of pastures at the Santa Mónica farm. In this study, pastures were analyzed using the normalized difference vegetation index (NDVI) time series obtained from median quarterly mosaics obtained in 2019. Different samples were drawn that represent the change trend throughout the time series and were classified according to their degree of change and persistence in the series. The results indicate that the densification of the time series allows to provide better results in the quantification of the changes and dynamics of the coverage. The established methodology represents a great advance on the generation of images and the monitoring and detection of changes in coverage through time series [22]. Hence, it is one the first studies carried out in the country that incorporate this type of process. It was concluded that the determination of spectral signatures with the index used together with the near infrared (NIR) and short wave infrared (SWIR 1) spectral bands, allow to extract values and intervals where the change produced by pastures is identified with an acceptable level of accuracy.