Parameterization and verification of the STICS model for grapevines in the region of Croatia

Many studies have shown how changes in air temperature affect the dates of occurrence of phenological phases. Grapevine plays a significant socio-economic role, both in Croatia and throughout the entire Mediterranean region. 'Graševina' and 'Plavac mali' are the most represented varieties in Croatia, so the emphasis in this research is on them at four different locations: Daruvar, Križevci, Hvar, and Lastovo. In addition to 'Graševina' and 'Plavac mali,' international varieties such as 'Chardonnay' and 'Merlot' were also observed.

STICS is a plant development model that simulates the soil-plant-atmosphere relationships based on the water, carbon, and nitrogen balance required for the growth of different crops, whether annual and/or perennial, herbaceous and/or woody.

The optimization of parameters was carried out using the implemented Nelder-Mead simplex method. Although it is desirable to optimize as many parameters as possible, in reality, increasing the number of parameters complicates the model. Therefore, it is necessary to find a compromise and determine the minimum number of parameters sufficient for quality optimization. In this work, 10 parameters were selected. These are parameters that are common to all predefined varieties and had different values.

During measurements, the STICS model showed satisfactory results at three out of four observed stations. The model verification demonstrated success with correlations primarily between r=0.6 and r=0.8. The performance of the model can be visually represented in various ways. In this work, two methods were employed: QQ plot and Taylor diagram.

Phenological and meteorological data from the Croatian Meteorological and Hydrological Service were used.