|
The aim of this work is to analyse grapevine temporal signatures over a viticultural landscape with remote sensing data in order to evaluate the impact of the perennial practices on optical and SAR signals. For this, the effects of different combinations of vineyard organisations on multi-temporal high and very high spatial resolution SAR and optical data have been analysed. The study area is located in Mendoza (Argentina). All the ground data were acquired during an experimental campaign performed in 2014-2015 over 153 vineyards (around 400ha). The data of the perennial practices were recorded over each of the 153 vineyards. The schedule of the annual practices were also provided over the all site. The grapevine phenological cycle was finally monitored on 14 fields through GLAI (Green Leaf Area index) estimation using hemispherical images and grapevine canopy height measurements. The satellite images dataset is composed of 6 Dual-Pol TerraSAR-X images, 5 Pleiades images, and 6 Landsat-8 images. After calibrating the images, backscattering coefficients and polarimetric parameters were extracted from Terrasar-X images (entropy and alpha angle) and the NDVI from optical images. The analysis of the temporal signatures regarding perennial practices revealed a strong impact of the inter-row spacing management. Ranges of values of backscattering coefficients differ with the type of management but their temporal variation seems to be linked to soil moisture changes. Depending on the roughness of the inter-row spacing the alpha angle evolution is linked to grapevine growth in the case of a chemically weeded inter-row spacing. But it remains at high values when the inter-row spacing is ploughed or grassed. In those cases, the entropy increases along the growth cycle. The vegetation indexes series are increasing with the grapevine growth depending on canopy width and soil management. Thus this work highlighted the variability of the SAR and optical signals due to heterogeneities in vineyard organisation and management.
|
