Disentangling the sources of variation in stomatal regulation in field-grown cultivar-rootstock combinations

Abstract

The inherent variability of Nature poses challenges for researchers to draw clear conclusions from field experiments. Identifying and assessing adaptations to climate change requires agronomic field trials. These are always complex to interpret, given the numerous and complex interactions between genotypes, environmental conditions, and cultivation practices. The existing diversity of plant material makes the selection of cultivars and rootstocks a feasible lever of adaptation to climate change, and one that is already being adopted by growers. Here, we assessed the main sources of variation in stomatal regulation (gs) across a large cohort of cultivar-rootstock combinations in our completely randomized GreffAdapt vineyard (Cabernet Sauvignon, Grenache, and Syrah each grafted onto 5BB, 1103P, 110R, 140Ru, and SO4). Using the LI-600 porometer/fluorometers we assembled an incredibly large and robust dataset of gs values across the season, comprising over 6000 gs and chlorophyll fluorescence measurements. Our goal was to disentangle sources of variation in gs in order to determine the contribution of plant material. Classical analysis of variance showed that the environment and date of measurement accounted for the overwhelming proportion portion of the variance (~90%), which obscures potential cultivar-rootstock effects over the season. Therefore, we integrated machine learning, spatiotemporal normalization of the gs response, and the use of linear mixed models to disentangle environmental (e.g. VPD, PAR, and leaf temperature), plant material, and vigour-related (e.g. yield and pruning weight) variables. The resulting model demonstrated that cultivar and cultivar-rootstock interactions, taken together, explained over half of the variation in gs. The rootstock by itself had a low contribution to the variation in gs, particularly in Grenache, which exhibited the most conservative stomatal regulation across nearly all rootstocks. The 5BB rootstock promoted high stomatal conductance on both Cab. Sauvignon and Syrah, while 1103P showed a more conservative behavior in Syrah. This study highlights that because of the incredibly high variation due to environment across space and time, field studies cannot rely on just a few measurements taken at a few times. Broader and more comprehensive datasets, paired with state-of-the-art statistical modeling, are likely important for understanding the contribution of specific factors to water-use regulation in vineyards.