Herbivory resistance in dwarf shrubs combines with simulated warming to shift phenology and decrease reproduction

Abstract

1 INTRODUCTION

In response to recent climate change, many plants have altered the timing of life-history stages (phenology), with events such as budburst, flowering and fruiting occurring much earlier than pre-industrial records (CaraDonna et al., 2014; Parmesan, 2006). Although the magnitude and direction of responses may be species specific (CaraDonna et al., 2014; Collins et al., 2021), the overwhelming pattern is cause for concern due to the unprecedented rate at which organisms at all trophic levels are changing. Differential responses may lead to asynchrony between interacting species (Both et al., 2009; Burkle et al., 2013; Hegland et al., 2009), and more research is required to enhance our understanding of the demographic impacts, such as those on reproductive success (CaraDonna et al., 2014; Forrest & Miller-Rushing, 2010; Iler et al., 2021). Furthermore, while phenological changes are typically linked in studies to bottom-up factors, such as genetics or climatic cues (Forrest & Miller-Rushing, 2010), plant vegetative and flowering phenology can be advanced or delayed by top-down stressors such as herbivory (Poveda et al., 2003; Tadey, 2020; Zhu et al., 2016). As increased herbivory and more frequent insect outbreaks have been hypothesised as consequences of a warming world (Hamann et al., 2021; Tylianakis et al., 2008), more experimental studies are needed to explore the combined effects of abiotic and biotic stressors.

Typically, research focusses on well-established climatic cues of phenological change. The accumulation of heat (e.g. degree days) is well documented as a predictor of flowering (Jackson, 1966; Miller-Rushing et al., 2007), and many temperate plants also have a winter chilling requirement that restricts emergence to springtime (Morin et al., 2009). In alpine conditions and at high latitudes, snow cover and snow melt dates provide important abiotic controls to winter survival and emergence phenology (CaraDonna et al., 2014; Iler et al., 2013). By contrast, herbivory is often considered to impact plant performance directly by reducing biomass, removing photosynthetic and/or reproductive tissue and negatively affecting fitness (Barrio et al., 2017; Bustos-Segura et al., 2021; Moreira et al., 2019; Rasmussen & Yang, 2023), although compensatory responses are also common (e.g. Lemoine et al., 2017; Poveda et al., 2003). However, additional impacts of herbivory also occur when plants under attack redirect resources from growth and reproduction to defence in an effort to improve herbivory ‘resistance’ (Benevenuto et al., 2020), and the altered physiology can substantially change both vegetative and flowering phenology (Forkner, 2014; Freeman et al., 2003; Ru & Fortune, 1999; Young et al., 1994).

Despite the top-down role of herbivory on phenology, the exact mechanism is unclear. In some studies, herbivory is considered to have an indirect effect on phenology, via physical modifications to microclimate, community composition and interspecific competition (Han et al., 2016). For example, some species-specific delays to vegetative and flowering phenology under intense large herbivore grazing (Tadey, 2020) or in controlled grazing experiments (Han et al., 2016; Zhu et al., 2016), were attributed to the impact of large herbivore presence on soil moisture. Studies investigating the impact of herbivory by insects have identified more direct phenological impacts, such as shortening flowering duration (Poveda et al., 2003; Schat & Blossey, 2005), delayed flowering (Agrawal et al., 1999; Freeman et al., 2003; Lemoine et al., 2017) and advanced flowering due to resource allocation changes (Bustos-Segura et al., 2021; Pak et al., 2009). Others have used the organic compound methyl jasmonate (MeJA) to induce defence responses similar to those exhibited during insect herbivore attack and found that the enhanced herbivory resistance caused delayed (Agrawal et al., 1999; Zhai et al., 2015), advanced (Pak et al., 2009) and neutral (Thaler, 1999) phenological effects. These conflicting results have raised questions about the role of herbivory in phenology, particularly as they either failed to disentangle the impacts of large herbivore grazing on plant physiology, or are based on laboratory conditions, annual plants and laboratory-reared insects.

Few studies have explored the combined impact of bottom-up and top-down drivers, such as warming and herbivory, on plant phenology (Lemoine et al., 2017; Sun et al., 2023). With a lack of consensus on the impact of herbivory, combined effects may counterbalance each other in the case of an herbivory-induced delay and warming advance (Lemoine et al., 2017), or lead to additive effects with responses in the same direction (Sun et al., 2023). Similarly, species effects may depend on climatic context, such as elevation, with suboptimal and relatively stressful conditions exacerbating some responses (Gimenez-Benavides et al., 2011; Hegland & Gillespie, 2024). To address these uncertainties, we conducted the first study of combined warming and herbivory resistance effects on the phenology of two long-lived dwarf shrubs in field conditions, at three elevations in open boreal forests in Western Norway. We used MeJA to simulate plants' physiological resistance responses to a single year of insect outbreaks (top-down effect) and open top chambers (OTCs) to simulate continuous summer warming (bottom-up effect). We then followed two plant species that are responsive to both treatments in different ways depending on elevation and optimal growing conditions. Vaccinium myrtillus (bilberry), an early-flowering deciduous species that thrives at mid-elevations in Norway (ca. 450 m.a.s.l.), shows typical induced defence responses to MeJA (reduced growth and herbivore damage; Benevenuto et al., 2020), although conflicting phenological responses to warming by OTCs have been reported (Anadon-Rosell et al., 2014; Prieto et al., 2009). The later-flowering and smaller evergreen shrub V. vitis-idaea (lingonberry), which is drought tolerant and performs well at low and warmer elevations, not only appears less responsive to induced defences (Hegland & Gillespie, 2024), but also advances phenology under artificial warming (Rosa et al., 2015). Of the two species, bilberry naturally tends to suffer more herbivore damage than lingonberry (e.g. Kozlov et al., 2015).

Using our combined treatments and study species, we aimed to quantify vegetative and reproductive phenological responses, and to establish consequences to plant fitness in the form of reproductive output. Our first research question was (1) What impact does the combined treatment of experimental warming and induced herbivory resistance have on vegetative and reproduction phenology in bilberry and lingonberry at three different elevations? Based on previous responses of these plants, we expected (a) advances under warming with the largest advances at higher altitudes where temperature is limiting, (b) delays due to induced resistance, with stronger responses in the year following MeJA application and (c) the combination of treatments would cancel each other out in the year after MeJA application, although warming at our highest alpine site may make plants more resistant to herbivory (Hegland & Gillespie, 2024). We also expected bilberry to be more responsive to treatments, as it is more susceptible to drought stress and insect outbreaks (Taulavuori et al., 2013). As the consequences of changing phenology to plant fitness and variables related to demography are rarely studied (Iler et al., 2021), we further posed the question (2) To what extent do shifts in phenology impact the reproductive output of bilberry and lingonberry? We answered this question with structural equation modelling and built our models on the assumption that alterations to phenology would negatively impact reproductive output due to mismatches with pollinators (Moreira et al., 2019; Schat & Blossey, 2005), but that the positive effect of warming on phenology and plant development would cancel these effects out (Lemoine et al., 2017).