Benchmarking spectrum-splitting agrivoltaics using spectrally resolved ray-tracing and crop-modelling for yield-microclimate-energy balance across arid and humid climates

The rapid expansion of utility-scale solar photovoltaics intensifies land-use competition between energy generation and agriculture. Conventional agrivoltaic (APV) systems mitigate this conflict but can reduce crop yields due to shading. This study evaluates a spectrum-splitting APV design that separates sunlight into extended photosynthetically active radiation (ePAR, 400–750 nm) transmitted for crops and near-infrared wavelengths (750–1100 nm) reflected to adjacent photovoltaic modules for energy generation. A new simulation framework was developed that integrates spectrally resolved ray tracing, canopy microclimate modeling, mechanistic crop growth simulation, and photovoltaic performance to benchmark outcomes under realistic meteorological conditions. Case studies in two contrasting U.S. climates − iceberg lettuce ( Lactuca sativa ) in arid Yuma, Arizona and determinate tomato in humid Gainesville, Florida show that this design reduced harvest delays, narrowed maturity distributions, moderated canopy microclimate and improved water use efficiency by approximately 15% compared to a conventional APV system. Electricity generation decreased by approximately 30% under the baseline geometry, although this value varies with collector density and other tunable layout parameters. Even with this reduction, both sites maintained yields within 5–12% of open-field levels while co-producing electricity, confirming viable dual-use across climates. When spatial uniformity, harvest synchrony and water-use efficiency are incorporated into benchmarking, the spectrum-splitting system outperforms fixed-tilt under grower-priority weightings despite producing less electricity.