Natural selection on floral volatiles and other traits can change with snowmelt timing and summer precipitation

Introduction

Global climate change is causing rapid changes in environmental conditions, such as increased average temperatures, more frequent extreme temperatures, and alterations of precipitation patterns (Pörtner et al., 2022). Those environmental changes have the potential to alter traits of organisms in ways that may influence species interactions. Average trait expression in a population can respond to the environment either directly or through evolutionary change. The former mechanism is phenotypic plasticity, in which the phenotype associated with a particular genotype responds directly to the environmental conditions (Bradshaw, 1965). In the latter mechanism of evolutionary change, the environmental change alters natural selection on the trait (Siepielski et al., 2009, 2017; Bemmels & Anderson, 2019), or on the ability of the trait to respond plastically, leading to an evolutionary change if trait variation is at least partly heritable (Gomulkiewicz & Shaw, 2013; Carlson et al., 2014).

In plants, floral traits play crucial roles in interactions with animals, and like other traits may be affected by climate change. Pollinators are thought to be the main source of natural selection on floral traits (review in Harder & Johnson, 2009), and traits can also influence interactions with natural enemies such as florivores and seed predators (Galen & Cuba, 2001; Frey, 2004; Sletvold et al., 2015). Some floral traits, such as floral size, show relatively consistent plastic responses to drought (review in Kuppler & Kotowska, 2021) or other environmental changes expected under climate change. In addition to trait expression, natural selection on floral morphology can change with climatic factors (Campbell & Powers, 2015). A change in selection with adverse abiotic conditions could happen in several ways. Increased resource limitations on seed production can weaken selection mediated by pollinators, as suggested for Ipomopsis with earlier snowmelt (Campbell & Powers, 2015). A drop in pollinator availability at a new time of flowering can strengthen pollen limitation and selection for attractive traits. Selection can shift due to changing pollinator preferences in response to plastic changes in floral traits (Dorey & Schiestl, 2022) or the availability of nectar or pollen resources.

Along with flower size, reward production, and petal color, floral scent emissions are also intimately involved in interactions with animals (Raguso, 2008). Flowers often emit a complex blend of many volatile organic compounds (hereafter volatiles; Dudareva et al., 2013), and a variety of insects not only detect these compounds but also show preferences or avoidance of particular volatiles or mixtures (Galen et al., 2011; Kessler et al., 2013; Byers et al., 2014). However, as pointed out by a recent meta-analysis (Kuppler & Kotowska, 2021), we know less about how floral volatiles respond to environmental conditions including drought (Burkle & Runyon, 2016, 2017; Glenny et al., 2018; Rering et al., 2020; Descamps et al., 2021; Dötterl & Gershenzon, 2023) than we do for other floral traits. Furthermore, natural selection on floral volatiles is rarely measured, possibly due to the difficulty of doing so in the field (Parachnowitsch et al., 2012; Kessler et al., 2013; Gross et al., 2016; Chapurlat et al., 2019; Gfrerer et al., 2021; Campbell et al., 2022a). To our knowledge, there are no studies of how natural selection on floral volatiles is altered by the environmental change expected under climate change. We also know little about whether plasticity in floral volatiles is adaptive. Concordant changes in volatile emissions and fitness in a new environment would indicate adaptive plasticity (Caruso et al., 2006).

We focused on phenotypic plasticity of floral volatiles, and how selection on volatiles, morphology, and rewards changes in response to two features of climate change: earlier snowmelt in the spring and changes in precipitation during the growing season. Our study system was the well-studied Ipomopsis aggregata, pollinated primarily by hummingbirds, but also visited by insects (Price et al., 2005). This species emits > 50 identifiable floral VOCs (Irwin & Dorsett, 2002; Bischoff et al., 2014), a few of which have been shown to influence fitness components in this species or a close relative. Emission of the floral volatiles α-pinene and β-pinene influences seed initiation (Campbell et al., 2022a), suggesting pollinator-mediated selection on scent along with floral morphology and color (Campbell, 1989; Campbell et al., 1996; Meléndez-Ackerman & Campbell, 1998). Pinene emissions also influence the proportion of fruits escaping seed predation by a fly, Delia (Anthomyiidae), that attacks seeds before dispersal (Campbell et al., 2022a), and may affect fly oviposition (Irwin & Dorsett, 2002). Pollinator-mediated selection on the floral volatile indole emitted by Ipomopsis tenuituba, a close congener of Ipomopsis aggregata, is supported by behavioral attraction of hawkmoth visitors to that compound (Bischoff et al., 2015). Similar preferences of mutualists and antagonists for Ipomopsis floral traits could create conflicting selection pressures, so we investigated reproductive success due to both pollination and avoidance of seed predation.

In many snow-dominated ecosystems at high elevations or latitudes, warmer spring temperatures are accelerating snowmelt, at the same time that summer droughts are changing in duration and severity (Clow, 2010; Pederson et al., 2011; Klein et al., 2016; Wadgymar et al., 2018). Earlier snowmelt produces a longer drought period in early summer before the onset of the summer monsoon rains (Sloat et al., 2015). That change in snowmelt has weakened selection on some aspects of floral morphology in I. aggregata (Campbell & Powers, 2015), likely due to increased water limitation preventing high seed production even when pollination is increased by a favored flower trait. Dry summer periods without precipitation are also becoming longer in the region (Zhang et al., 2021). A controlled drydown experiment showed that the amounts and composition of I. aggregata floral volatiles respond to reductions in soil moisture over 5–13 d (Campbell et al., 2019), but plasticity over longer time periods under natural conditions has not been characterized. Using field manipulations of snowmelt timing and summer precipitation also used by Powers et al. (2021) to demonstrate plasticity in floral morphology and rewards, we asked the following questions.

1. How does an advancement in timing of snowmelt and an increase or decrease in summer precipitation influence emissions of floral volatiles?
2. To what extent are effects on floral volatiles driven by changes in soil moisture?
3. Does natural selection on volatiles and other floral traits by pollinators and seed predators vary with changes to snowmelt and precipitation? If water limitation weakens selection based on seed production, we would expect weaker selection with early snowmelt and reduced precipitation.
4. Are plastic responses of traits to different environmental conditions adaptive?