Here, we experimentally examined how an episodic lack of sunshine (ELS), as an extreme weather event, would affect rice productivity under warming with elevated [CO2 ]. In 2009 and 2010, rice plants were grown at two levels of [CO2 ] (ca. 390 and 650 μl l(-1) ) and three levels of warming (≈ambient, +1.2 °C, and +2.2/2.4 °C) in six independent temperature gradient field chambers (three each for ambient and elevated [CO2 ]). At panicle initiation (PI), booting (BT), or flowering (FL), rice plants were exposed to ELS (ca. 18% of full sunlight) for 10-14 days consecutively. As expected, ELS elicited a significant reduction in aboveground biomass (AGB) and yields. However, elevated [CO2 ] had the potential to relieve the ELS-induced reduction in AGB and yield, whereas warming had the reverse effect for yields, without a significant warming × [CO2 ] interaction. When ELS applied at PI, BT, and FL, the extents to which warming-reduced yields (averaged across [CO2 ] levels) ranged from 9 to 25%, 7 to 14, and 10 to 18% at +1.2 °C, and ranged from 24 to 56%, 22 to 55%, and 18 to 46% at +2.2/2.4 °C across two seasons, respectively. Meanwhile, under normal sunshine they ranged from 1 to 3% at +1.2 °C and 7 to 21% at +2.2/2.4 °C. Warming predisposed rice plants that had experienced ELS to be more sensitive to spikelet sterility and spikelet number per panicle, accounting for most of the yield reductions. These findings provide evidence that an expected warming could further exacerbate rice productivity if ELS occurs simultaneously during reproductive stages. Our results collectively suggest that it might be critically important to consider extreme events for a holistic evaluation of the potential impact of warming and [CO2 ] on crop productivity, when considering changing climate.
