Modelling estimation of crop and soil variables in agrivoltaic systemsin marginal areas

The present study aims to present a preliminary estimation of crop productivity in a wide range of shading conditions in agrivoltaic Systems in Italian marginal lands. The first objective of this study was the identification of marginal areas of Italy potentially suitable for agrivoltaic systems, as a baseline for technical and economic viability. Two procedures were applied, e.g.deterministic and probabilistic approaches. The deterministic approach considered the constraints which allowed us to exclude entire portions of the study area according to our purpose. Such conditions are linked to: Physical aspects: elevation and slope - Socioeconomic aspects: land-use/land-cover (LULC) - Environmental aspects: areas of natural interest. The probabilistic approach is based on the deterministic one to which a Multiple-Criteria Decision Analysis (MCDA) was added (Cervelli et al. 2024) In this work, the choice of criteria was strictly related to land marginality aimed at finding the best land allocation in terms of sustainability of energy supply, with a special reference to agrivoltaic systems. A double set of criteria was taken into account: 1) Land marginality: factors determining land marginality in terms of geomorphology (elevation, slope, etc.). environmental risk (soil erosion), socioeconomic context (land use); 2) Profitability: factors determining the maximum return of investments in the agrivoltaic systems (solar radiation elevation, slope, aspect, proximity to the electricity grid). The second objective was to build a harmonized database of yield and other crop-related data in shading conditions considering three crop functional groups, i.e cereals, grass ley, and non-food oleaginous crops, to identify the most suitable species for cultivation at different levels of shading Among the food crops, the most promising are fodder crops, limited to shading conditions of around 40%. The yields of Ricinus are highest when cultivated at shading levels of 50%, due to its apparent tolerance to shading The collected data were used to calibrate the crop parameters of the ARMOSA simulation model (Perego et al., 2013). The third objective was the development of a module that was integrated into the modular model ARMOSA to simulate a wide range of agrivoltaic plans characteristics (fix of solar tracking panels height, panel width, radiation porosity, panel density, distance between panels) to reproduce the condition that is mainly characterized by the reduction of the radiation and evapotranspiration. In three sites identified in the marginal lands of southern Italy (probabilistic approach), the yield reduction of durum wheat, barley, alfalfa, Ricinus safflower, and camelina was simulated in 15 agrivoltaic configurations in interaction with varying aspects and slopes. Evapotranspiration, nitrogen uptake, nitrogen losses and soil organic carbon sequestration were also estimated in each scenario. The average yield reduction was 10% to 40%, in agreement with the literature data. The newly implemented is a promising tool to simulate yield driven by shading and field management with a flexible model that can be set by the user.