Main content area

Towards an empirical spatial model to estimate the spatial variability of the total soluble solids of grapevine at the within field scale

Verdugo-Vasquez, N., Acevedo-Opazo, C., Valdes-Gomez, H., Araya-Alman, M., Tisseyre, B.
Acta horticulturae 2017 no.1188 pp. 343-350
Vitis, cultivars, data collection, harvesting, models, statistical analysis, total soluble solids, wine grapes, Chile
The aim of this work is to validate an empirical approach to estimate spatial variability of the total soluble solids (TSS) of grapevines at the within field scale. Estimation of TSS involves a spatial model calibrated with historical TSS data. The novelty of the approach is to provide collaboration between historical data (calibrated model) and a punctual reference measurement (reference site) of TSS to estimate the TSS values over the whole field at a given date. The experiment was carried out between 2009-2015 in two fields ('Cabernet Sauvignon', 'Chardonnay'), both located in the Maule Valley, Chile. A regular sampling grid was designed within each field. On each site of the grid, TSS was measured on different dates (from veraison to harvest). The reference site on each field was chosen randomly. The ability and quality of the model to estimate TSS values was tested on an independent data set, using classical statistics (R2 and root mean square error - RMSE). For both fields, the results highlighted a high spatial variability of the TSS at the within field scale. Also, TSS estimation using the model yielded high R2 values, where the best results were observed in the 'Cabernet Sauvignon' field. This is related to the quality of historical data used to calibrate the spatial model. For both fields, RMSE temporal values were less than the standard deviation in most dates. These results highlighted the opportunity and the relevance of using the proposed spatial model to estimate the spatial variability of TSS of the whole field through a punctual measurement of a reference site. Further studies should consider other cultivars and determine if the reference site selection influences the results.