A model library for estimating grapevine development and growth under different pedo-climatic conditions

Grapevine (Vitis vinifera L.) is a valuable fruit crop characterized by a worldwide importance. In particular, Italy, France and Spain are the biggest wine-producing countries and they play a relevant role in the world wine economy. However, the specific climate conditions of the narrow geographical areas in which grapevine is currently cultivated expose the viticulture suitability to the great risk of the climate change and extreme events. In this context, the grapevine simulation models represented useful tools for investigating the main physiological plant processes under different pedo-climatic conditions. Accordingly, the objective of this thesis is to provide a new software component for estimating grapevine growth, yield and quality in different environments. The new software component UNIFI.GrapeML is presented in Chapter 2 as an extendible model library in which the fine-granularity of the model structure allows an easier discretization and implementation of the code. In Chapter 2, UNIFI.GrapeML was tested in a specific case of study in Northeastern of Spain on a Chardonnay vineyard. A sensitivity analysis using Latin Hypercube and Sobol method was performed for evaluating the sensitivity of the model parameters on the final fruit biomass considering the environmental conditions of the study area. The results evidenced the strong impact of leaf area expansion and crop partitioning parameters on final output. The model was then calibrated on soil water content, phenology and fruit biomass data showing satisfactory results. Afterwards, UNIFI.GrapeML was implemented with a quality approach for assessing sugar concentration during ripening period according to the berry water content. This approach was evaluated in a specific case of study in Montalcino wineproducing region (Italy) for Sangiovese variety over the period 1998-2015. In this case, the phenological phases (budbreak, flowering, veraison and maturity) of Sangiovese variety were calibrated using data from Susegana and Montalcino for determining the length of grapevine cycle. Then, the model calibration and validation on observed sugar content data was performed for Montalcino site. The results showed high correlations in both calibration and validation procedures considering the yearto- year variability of the dataset. Finally, the impact of climate change and extreme events was evaluated on the phenological cycle of very early, early, middle and late varieties at European scale (Chapter 4). The effect of the mean climate change on phenology was assessed using a chilling-forcing model while the effect of extreme events was accounted through the frost events at budbreak and the effect of suboptimal temperatures at flowering stages (fruit-set impact implemented in UNIFI.GrapeML). The results highlighted an overall advance of budbreak and flowering phases of all varieties across a latitudinal and IV longitudinal geographical gradient, especially in central/eastern Europe. In particular, the climate change showed a high impact on budbreak of late compared to very early and early varieties in western European region. Moreover, a higher decrease of frost events was evidenced in western regions compared to central/eastern Europe and a more relevant effect of these events was found on very early and early compared to late varieties. On the other hand, the estimation of the temperature stress at flowering stage evidenced a lower variability between varieties and scenarios while relevant differences were showed between European regions.