Critical levels of nitrogen in leaves of grapevine grown on sandy soils in southern Brazil using Bayesian models
Sandy soils in southern Brazil have low levels of soil organic matter, and consequently, low nitrogen (N) availability to grapevines. Therefore, it is necessary to apply N fertilizers to achieve satisfactory yield. However, the N critical levels (CL) in leaves of grapevines grown in subtropical regions is not sufficiently known. The study aimed to determine the critical levels and sufficiency ranges (SR) of N in leaves for grapevine cultivars grown on sandy soils in southern Brazil using Bayesian models. Leaf data from different long-term calibration experiments in southern Brazil were used. The grapevine cultivars sampled were Chardonnay, Pinot Noir, Cabernet Sauvignon and Alicante Bouschet, which were grafted onto 1103 Paulsen rootstock. About 600 observations of N concentration in leaves collected during flowering and veraison and the respective grape yield were used. The relative production was calculated from the highest yield obtained for each cultivar and each crop season and from that, the other proportions were established. The concentration of total soluble solids (TSS) and anthocyanins was evaluated at harvest. Linear regression with a plateau at the 95% percentile was implemented to quantify the relationship between N concentration in leaves and grape yield, TSS, and anthocyanins. The critical concentration of N in leaves was assumed to be the point at which the adjusted line reaches the plateau. The CL and SR of N in leaves collected during flowering and veraison for yield were 19.2, 18.8 to 19.5 g N kg‑1 and 25.4, 24.3 to 27.2 g N g kg‑1, respectively. The CL for the TSS and anthocyanins were 18.4 and 25.2 g N kg‑1 at flowering, and 25.2 and 26.2 g N kg‑1 at veraison. The proposition CL and SR of N in leaves allow to accurately verify the potential yield and the grape quality.
Stefanello, L.O., Tassinari, A., Schwalbert, R., Schwalbert, R.A., Hindersmann, J., Moura-Bueno, J.M., Carranca, C.L.V.A.F., Toselli, M. and Brunetto, G. (2022). Critical levels of nitrogen in leaves of grapevine grown on sandy soils in southern Brazil using Bayesian models. Acta Hortic. 1333, 153-160
grape yield, leaf diagnosis, vineyards, regression models, machine learning