Agrivoltaics model scores farmworker heat stress next to crop light and water

Erfan Hosseini and colleagues have published a model in the Journal of Advances in Modeling Earth Systems that simulates the microclimate under solar panels and puts worker heat stress in the same accounting as crop light, soil and water ^{[1] [2]}. In a simulated tomato farm run on weather from a hot, humid day in Princeton, New Jersey, the average perceived temperature for humans during working hours fell by 4.46 C compared with open field ^{[3] [4]}, which reframes row spacing and panel height as an occupational-safety specification rather than only an energy-versus-yield trade.

The model tracks energy, momentum and mass through the whole system: panels, crops, soil, air and water movement, and carbon dioxide uptake ^[2]. That is the point of it. Existing agrivoltaic work, according to the authors, tends to take one variable at a time, light availability or crop growth, rather than the interaction of microclimate, crop type, light and panel type ^[5]. The team checked outputs against field data, including leaf temperatures from Davis, California and soil temperatures from Chicago City, Minnesota ^[6].

The crop results are the familiar shading story with numbers attached. Shaded tomato leaves ran 1.84 C cooler across the day and up to 7.56 C cooler in the peak afternoon, and evapotranspiration fell 22.4% ^[7]. Incident sunlight dropped 47% but carbon uptake fell only 31% ^[8], which works out to roughly 30% more carbon fixed per unit of light reaching the canopy ^[13]. Set the same two figures against water instead of light and the picture is less flattering: carbon uptake fell further than water loss, implying about 11% less carbon per unit of water transpired ^[14]. Shade buys thermal relief and water savings; it does not buy free photosynthesis.

The hardware benefits too. Panels over the cropped surface were 5.6 C cooler in daytime than panels over bare soil, recovering about 15% of the efficiency that heat costs them ^[9].

Two cautions. Both write-ups of this study trace to the same Eos research spotlight ^[10], and the headline worker number comes from a hypothetical farm on one hot day in one location ^[3], not from measurements on labourers. And the summaries report the 4.46 C figure as a drop in "average perceived temperature for humans" without naming the index or the baseline it fell from ^[11]. A 4.46 C reduction matters very differently at a wet-bulb globe temperature near a work-rest threshold than well below one, and heat rules are written against thresholds.

Watch for the underlying paper's definition of that perceived-temperature term and whether it maps onto the indices used in state heat standards, since that determines whether the result can be cited in a permit or a safety plan. Watch, too, for validation against instrumented workers rather than leaves and soil ^[6], and for runs across the other climates and crop combinations the authors say the model can handle ^[12]. If the effect holds, layouts optimised only for the light-versus-generation curve are being designed without one of their consequences priced in.