The adoption of winter cover crops (CCs), to be terminated in spring, is gaining popularity in temperate climate viticulture. We evaluated the effects of two novel termination practices—interrow rolling (R) and subrow mulching (SRM)—on vine performance and physiology of Barbera grapevines over 3 years, comparing them to a standard termination practice (green manuring, GM) and a control (C) consisting of alternated tilled and vegetated interrows. For each termination type, we used a cereal‐based (Humusfert, H) and a more balanced (Stratus, S) seed cocktail. Data collection included CC biomass, seasonal soil and leaf water potential, leaf assimilation (A) and stomatal conductance ( g _s ) rates, vine vigor, yield, and grape composition at harvest. R caused a 31% reduction in pruning weight compared to C. Soil water potential at 30‐cm depth, as well as leaf A and g _s , was more limited than in the other treatments and, in S‐R, ripening was delayed. R performances might suffer from increased surface evaporation and reduced precipitation effectiveness due to the thick mulch layer. Both SRM treatments ensured a thick mulch layer (over 1 kg of desiccated grass/m ² ), which effectively controlled weeds when placed under the row. SRM did not impact vine vigor and yield; however, using the S mix slightly delayed final ripening. The presence of a thick, dead mulch layer under the row improved soil water potential (Ψ _SOIL ) at 30‐ and 60‐cm depths compared to all remaining treatments that shared a mowed under‐the‐trellis strip. Seasonal dynamics of absolute Ψ _SOIL show that when Ψ _SOIL in the C plot fell below −0.2 MPa, Ψ _SOIL was significantly higher (less negative) in H‐SRM than in S‐SRM. This was also associated with average reductions of 31.3% and 37.1% in A and g _s rates in S‐SRM compared to H‐SRM, respectively. The higher fresh biomass yielded by the more balanced mix (S) may relate to increased soil water depletion, especially during the pre‐termination phase. Both C and GM treatments showed similar vine performance; however, C was significantly limited in leaf gas exchange compared to others. C does not benefit from any mulching effect, and a competitive native grass cover persists every second interrow during the dry and harsh summer season.

Global warming is challenging the performances of medium-to-late ripening red Vitis vinifera cultivars whose harvest dates might be consistently anticipated at the detriment of still insufficient phenolic maturity. A 3-year study (2021–2023) was devised to compare cv Croatina—the most grown red cultivar in the Oltrepò Pavese district—with the following five autochthonous genotypes: Croà, Moradella, Mornasca, Ughetta, and Uva Rara. Weather trends, main yield components, and grape composition parameters were recorded each year; in two out of the three trial seasons, ripening curves for total soluble solids (TSS) and titratable acidity (TA) were also derived. In terms of yield performance, all minor varieties showed a level of basal node fruitfulness (about 1.1 clusters/shoot) high enough to perform short pruning, which was not possible to achieve in Croatina. As per grape quality at harvest, Uva Rara behaved quite similarly to Croatina, whereas Ughetta, Moradella, and Croà were judged to be unsuitable due to poor berry coloration at harvest (less than 0.7 mg/kg). Conversely, Mornasca’s performance was truly interesting and promising. In 2021 and 2023, Mornasca had delayed sugar accumulation, which was strongly uncoupled with total anthocyanins, which were not limited. In 2022, a hot and dry season, Mornasca outdid Croatina because the required TSS was assured, and the color significantly improved. The conclusion is that Mornasca is less susceptible than Croatina to imbalances in pigment formation or degradation, which typically occurs with berry overheating.

Purpose: A three-year study was carried out in a steep vineyard to study inter-annual and within-season vigor variability by proximal sensing. Methods: Each season, vigor maps were obtained around veraison when a full canopy status was achieved; in 2021 and 2022, within-season vigor variability was assessed at 9 and 8 dates. Every year, main yield components, grape composition and pruning weight were determined. In 2022, dynamic in vegetative growth, canopy density, leaf gas exchange and water status, soil temperature and water status were also performed. Results: Mostly due to more rapid soil warming, vines growing in low vigor (LV) plots had fostered leaf area development and shoot growth over the first three dates of measurements before high vigor (HV) fully made up at pre-flowering. From flowering onward, vine growth in LV was totally checked and the gap versus the HV kept widening. Over summer, in LV, leaf water potential assessed pre-dawn and midday reached the quite limiting thresholds of −0.6 and −1.4 MPa, whereas in HV water stress was rather moderate. Gas exchange limitations in LV essentially mirrored the above with stomatal conductance decreased to 0.03 mmol m ⁻² s ⁻¹ at pre-veraison. Final vine size and yield were severely curtailed in LV (−49% and −62% as compared to HV, respectively) which, on the other hand, achieved much better and complete grape maturity. Conclusion: This trial confirms that even within a tiny and steep vineyard, spatial variability between vigor zones evaluated at full canopy tends to be maintained over years, thus justifying, for instance, a selective harvesting approach. However, it is quite meaningful that, in both 2021 and 2022, the ultimately defined LV vigor initially developed bigger canopy, balanced with HV at pre-flowering and then re-entered in their standard classification of “low vigor” plots. This behavior is relevant anytime a variable management strategy is planned. Indeed, early season fertilization should be adjusted with a logic opposite to that applicable to a “low vigor” level while, real-time canopy spray is recommended.

Accurate yield mapping is essential for precision viticulture, but direct yield measurement during mechanical harvesting remains difficult because of the heterogeneous harvested biomass. This study evaluates the use of images acquired during harvesting to estimate grape yield through features extracted from pre-trained deep learning models. Images were collected over two seasons in a commercial Merlot vineyard using action cameras mounted on a grape harvester. Five pre-trained models, including one convolutional neural network and four Vision Transformer architectures, were compared. The extracted features were related to measured yield using regularized regression models. The results show that pre-trained visual features can reliably estimate grape yield despite the presence of berries, grape clusters, leaves, shoots and motion blur. The convolutional neural network achieved prediction accuracy comparable to the Vision Transformer models, indicating that local image features remain highly informative for biomass-flow estimation. A sparse fusion approach identified more stable features, supporting improved model stability under the investigated conditions. This work demonstrates that general-purpose pre-trained visual embeddings provide an effective basis for grape yield estimation without task-specific model training. The proposed method, not depending on dedicated sensors, specific calibration or model tuning, offers a practical and scalable solution for integration into commercial grape harvesters, including aftermarket installations.