Effects of experimental warming on floral scent, display and rewards in two subalpine herbs

Floral volatiles, visual traits and rewards mediate attraction and defence in plant-pollinator and plant-herbivore interactions, but these floral traits might be altered by global warming through direct effects of temperature or longer-term impacts on plant resources. We examined the effect of warming on floral and leaf volatile emissions, floral morphology, plant height, nectar production, and oviposition by seed predators.

We used open-top chambers that warmed plants in the field by +2-3 °C on average (+6-11 °C increase in daily maxima) for 2-4 weeks across 1-3 years at three sites in Colorado, USA. Volatiles were sampled from two closely related species of subalpine Ipomopsis with different pollinators: Ipomopsis aggregata ssp. aggregata, visited mainly by hummingbirds, and Ipomopsis tenuituba ssp. tenuituba, often visited by hawkmoths.

Although warming had no detected effects on leaf volatiles, the daytime floral volatiles of both I. aggregata and I. tenuituba responded in subtle ways to warming, with impacts that depended on the species, site and year. In addition to the long-term effect of warming, temperature at the time of sampling independently affected the floral volatile emissions of I. aggregata during the day and I. tenuituba at night. Warming had little effect on floral morphology for either species and it had no effect on nectar concentration, maximum inflorescence height or flower redness in I. aggregata. However, warming increased nectar production in I. aggregata by 41 %, a response that would attract more hummingbird visits, and it reduced oviposition by fly seed predators by ≥72 %.

Our results suggest that floral traits can show different levels of plasticity to temperature changes in subalpine environments, with potential effects on animal behaviours that help or hinder plant reproduction. They also illustrate the need for more long-term field warming studies, as shown by responses of floral volatiles in different ways to weeks of warming vs. temperature at the time of sampling.