Consequences on energy and water exchanges of airflow modifications in agrivoltaic systems

While solar radiation is essential for plant development, overexposure can exacerbate abiotic stresses, making soil water content the key driver affecting plant growth. Agrivoltaic (APV) systems, defined as a land sharing between agricultural and photovoltaic (PV) energy productions, reframe the balance between beneficial and excessive solar exposure. Notably, APV systems have demonstrated, in some configurations, their capability to enhance agricultural yield by better conserving water in the soil. This study aims to deepen our understanding of evapotranspiration in APV configurations to better assess their impact on plant growth. It relies on data collected from three sonic anemometers installed at the SIRTA APV power plant in France, which provide the first measurements of wind speed and turbulent fluxes within an APV system. These observations reveal consistently lower wind speeds beneath PV panel compared to control areas without panels, while turbulence levels are notably higher. To complement these measurements, Computational Fluid Dynamics simulations are performed using an implicit PV panel model and a Soil-Plant-Atmosphere Continuum representation of vegetation, both integrated into the solver code_saturne . These simulations offer valuable insights into the spatial heterogeneity of energy and water exchanges within APV systems: the gradients of the turbulent fluxes are higher than unity. These findings challenge current evapotranspiration calculation methods which assume: (1) field homogeneity, (2) well-defined relationship between wind speed and turbulence, and (3) a consistent link between measurements taken at a reference height and exchanges occurring at the canopy top. • 3D sonic anemometers reveal that PV panels markedly alter airflows. • Wind speed tends to be attenuated while turbulence is generated due to the panels. • Most evapotranspiration models rely on assumptions invalid in agrivoltaic systems. • Measurement location can cause up to a twofold deviation in evapotranspiration estimates.