Resilience Analysis in Viticulture: an Expert Knowledge-Based Approach

The Importance of Resilience in Viticulture

The concept of resilience has evolved across various fields, with Holling (1973) defining it in ecology as the ability of ecosystems to absorb disturbances and continue functioning.

Resilience is understood through three core capacities: absorptive, adaptive and transformative. Absorptive capacity allows systems to endure shocks, adaptive capacity enables gradual adjustments and transformative capacity facilitates radical shifts when systems are unsustainable (Béné et al., 2012). These capacities are linked to disturbance levels, with less intense shocks more likely to be absorbed without affecting the system’s structure (Béné et al., 2016). In agriculture, resilience has become crucial in addressing climate change and environmental stressors (Jasná et al., 2014). Both short-term and long-term management strategies are necessary to build resilient farming systems (Martins et al., 2024). In viticulture, resilience requires moving beyond traditional methods and embracing innovation to better cope with climate variability and sustain production (Tscholl et al., 2024).

The Agrofood Research Hub at the University of Brescia launched a research initiative aiming at identifying adaptive strategies in viticulture, focusing on vineyard management to mitigate environmental stressors such as weather and pests. A conceptual model was developed, emphasizing how managing the vine’s physiological responses can reduce the gap between actual and target yields. This research introduces the concept of Technical Management Routes (TMRs), inspired by the concept introduced by Renaud-Gentié et al. (2014). These routes represent the interaction of different practices operating at short, medium, and long-term scales. Understanding the impact of different TMRs on both yield and quality is quite complex due to the variety of interacting practices that influence the final outcome.

This research addresses that complexity by focusing on the short-term effects (within a single-growing season) of various management routes across different meteorological and phytopathological conditions. Data used for this research were collected using an expert knowledge-based approach, by investigating how different soil and canopy management practices and their integration operate at a short-, medium-, and long-term scale. Given the complexity of various practices interacting to affect yield and quality, the research focuses on the short-term impact (within one growing season) of different TMRs under varying weather and pest conditions.

Expert Knowledge Elicitation: the case study of Franciacorta wine growing area

Experts from the Franciacorta wine region were called in to investigate how different vineyard management strategies and their interactions affect grape yield and quality in Franciacorta wine production. This was achieved using Expert Knowledge Elicitation (EKE) techniques (O’Hagan, 2019), a method that collects expert insights and converts them into quantitative data such as values and probability distribution parameters.

Four scenarios of vintage were developed by adjusting water availability (optimal, low and high) and factoring in the presence or absence of phytopathological risks. Eight management scenarios were planned for each of the vintage scenarios, by customizing, when appropriate, the management scenarios based on the specific characteristics of the year.

In developing the management scenarios, the following management categories and strategies were considered:

• Fertilization: no fertilization (no fertilizers) [Ref.*]; Organic fertilization only (application of organic fertilizer only, in fall or spring); Mineral fertilization only (mineral fertilizer only at budbreak or fruit set);
• Inter-row Management: permanent grass cover (no inter-row tillage) [Ref.]; Single subsoiling (one subsoiling pass in fall or spring in alternating rows); Subsoiling + subsequent harrowing (one subsoiling pass in fall or spring, in alternating rows + harrowing on the same row);
• Sub-row Management: grassed (under-row area maintained with grass cover) [Ref.]; Tilled (under-row area mechanically tilled);
• Defoliation: no defoliation (no defoliation performed) [Ref.]; Early (mechanical defoliation at flowering); Late (manual defoliation between cluster closure and harvest);
• Shoot Thinning: no thinning (no shoot thinning applied) [Ref.]; Yes (light shoot thinning on cane head and bends only);
• Topping Height: fixed (topping at the height set by the trellising system) [Ref.]; Variable (tipping height adjusted based on phenological stage)
• Topping at Flowering: none (no topping at flowering) [Ref.]; some (topping performed at the flowering period).

*The label “Ref.” in square brackets indicate the reference strategy used for comparison: relative effects in Tables 1–4 were calculated with respect to these reference strategies.

A total of 32 unique scenario combinations were created by incorporating practices related to soil and vegetation management.

In order to assist experts in identifying the effects of management practices based on the specific characteristics of the year, a typical context of Franciacorta viticulture has been described: Chardonnay vineyards cultivated using the Guyot training system on morainic soils that maintain balanced fertility.

With the aim to make the scenarios more realistic, emergency irrigation was included in cases of low water availability, following the Franciacorta DOCG regulations (Italian acronym for ‘Denominazione di Origine Controllata e Garantita’). Phytosanitary measures were assumed to be uniformly effective throughout all scenarios.

The elicitation process consisted of two stages. The first stage involved individual interviews with 11 experts, generating 352,000 data points on how the scenarios impacted yield (t/ha), potential alcohol content (% vol) and total acidity (g/l). In the second round, 7 more experts contributed an additional 96,000 data points.

For each of the three parameters, a linear regression model was built by incorporating two explanatory factors: year-specific variables (water availability and disease risk) and seven management practices. A follow-up analysis focused solely on management variables was performed separately for each vintage.

Both models isolated the direct influence of each variable by removing any confounding effects from other factors. 

Key Findings

The results of this research highlight, first and foremost, that experts clearly distinguished between different vintages. Notably, low water availability scenarios yielded less production, more potential alcohol and reduced total acidity – despite the use of emergency irrigation.

The assessment of management scenarios helps identify key elements to consider when defining the TMR (Table 1-4). In general, fertilization and soil management practices have been found to have the most significant impact on both yield and grape quality compared to canopy management practices.

• Fertilization: mineral and organic fertilization were both associated with increased yield in all scenarios. However, their effect was less pronounced under drought, likely due to limited nitrogen uptake. In optimal water conditions, mineral fertilization had a greater yield effect, most likely because of the slower nutrient release from organic matter. Under optimal water availability, fertilization also resulted in lower sugar content and better acidity retention at harvest, suggesting delayed ripening. The acidity-retaining effect was more evident in the absence of disease pressure, and, under high water availability, was only observed with organic fertilization. In drought years, fertilization had minimal, though statistically significant effects on quality and slightly reduced acidity on average.
• Soil management: inter-row subsoiling and under-row tillage increased yield in all scenarios, though to varying extents. Both practices were also generally associated with improved acidity retention, except under drought where under-row tillage slightly reduced acidity. Their effects on potential alcohol content varied depending on conditions. In drought years, continuous inter-row harrowing (the only scenario where it was included) significantly increased yield and helped reduce sugar accumulation while preserving acidity, again suggesting delayed maturation.
• Defoliation: early defoliation significantly increased yield under optimal water and low disease pressure, and modestly improved acidity retention. Under high water availability, only late defoliation had a notable effect, increasing yield but reducing sugar accumulation and acidity.
• Shoot thinning: defined here as slight intervention on head and curve shoots, this practice was only significant under optimal water availability. It increased yield when disease risk was present. Quality effects were mild but consistent across scenarios.
• Topping: height and timing (e.g., around flowering) were significant only under specific conditions. Topping around flowering mattered under high water availability, possibly due to vigorous vegetative growth. Topping at variable height had a weak and mostly negative effect on yield, and limited qualitative impact.