We suggest that a combination of photovoltaic panels (PVPs) and food crops on the same land unit should be possible. In these agrivoltaic Systems (AV), an upper layer of PVPs partially shade crops at ground level. PVPs are arranged in East-West orientated arrays, as in conventional solar plants but with a clearance that allow crop mechanisation. Our aim is to design AV with optimal land productivity. We tested the effect of PVPs cover at different density on different crops.

We managed a 820m² prototype, comparing standard full PVPs density (FD), and half density (HD), to full sun (control). Light pattern and plant development were measured in each treatment. Numerical simulations of the system were also achieved by coupling a radiation model with the STICS crop model. We monitored two species in the field : lettuce (L.capitata and L.acephala), bean, and ran simulations for wheat.

Radiation simulations and pyranometers records agreed that the annual available radiation at crop level was 46 and 71% in FD and HD respectively. We measured and modelled surprisingly high relative yields for plants in FD (durum wheat: 0.73; capitata: 0.53; L. acephala: 0.65; beans: 0.41) and HD (0.83, 0.74, 0.88 and 0.94 respectively). It appeared that the two varieties of lettuce have different sensibility to shade, resulting from different strategies for light interception and conversion. For acephala, an increase of the percentage of light interception (control: 34%; FD: 36%; HD: 38%) and of light use efficiency (respectively 17.5, 18.5 and 17.7 g of dry matter.MJ^-1.m²) were calculated. Conversely, for capitata, interception decreased, and conversion only slightly increased in HD. For both cultivars, increasing light interception appears to the main adaptation lever to compensate reduced light resource.