Introduction
Despite the evolutionary and ecological advantage of seed reproduction for herbaceous plants, investment in seed reproductive structures, including the inflorescence, is very costly (Reekie & Bazzaz 1987; Schoen & Dubuc 1990). Therefore, as an alternative investment, polycarpic species may reduce flowering in a given year if it proves costly relative to available resources and increase non-reproductive biomass production (Fujita et al. 2014). Due to this trade-off, reproductive allocation (RA, the ratio of reproductive structure biomass to total biomass) between herbaceous species in the community can be highly variable (Thompson & Eckert 2004) and even opposite, leading to species and individual asynchrony in generative reproduction.
Plant populations can be highly synchronous in their flowering, with species flowering extensively together during one growing season, but restricting their flowering in other years (i.e. year-to-year synchrony) (Pesendorfer et al. 2021). Synchronicity in plant reproduction was identified early as a consequence of resource matching (i.e., theresource matching hypothesis , Klebs 1904). According to resource matching theory, environmentally optimal years, characterised as resource pulses, induce an increase in seed reproduction across the community (Monks et al. 2016; Pearse et al. 2016; Bogdziewicz et al. 2020). A fixed proportion of resources is consistently allocated to both reproductive and vegetative growth each year, maintaining a constant RA (\(\frac{\text{dRA}}{\text{dt}}=0\)). As such, it serves as a baseline assumption for masting, where variations in annual seed production are not considered to be indicative of adaptive strategies (Bogdziewicz et al. 2020). Conversely, significant deviations from the fixed proportion of resources allocated to reproduction may indicate an adaptive response (\(\frac{\text{dRA}}{\text{dt}}\neq 0)\). In this case, resource limitation triggers an evolutionary trade-off between reproductive and vegetative growth investments (Fernández-Martínez et al. 2019). Plants can either prioritize reproduction even at the expense of growth by increasing RA (as proposed by the resource switching hypothesis , Hacket‐Pain et al. 2018) or they can delay by decreasing RA until resources are sufficiently accumulated (as proposed by theresource budget hypothesis ; Isagi et al. 1997; Satake & Iwasa 2000; Han et al. 2014). Empirical tests of these hypotheses have yielded diverse outcomes for different limiting resources (such as light, carbohydrates, nutrients, and water; Crone et al. 2009; Miyazaki et al. 2014; Smaill et al. 2011; Montesinos et al. 2012; Sala et al. 2012; Pulido et al. 2014), highlighting the species-specific dependence of RA on resource constraints (Pearse et al. 2016).
RA studies have traditionally used single-resource models, which consider a single resource limitation and overlook the complexity of multiple simultaneous limitations that plant communities typically face. Contrasting resource requirements and competitive strategies among species are also ignored. Coexistence in a multi-species community is driven by life history trade-offs associated with different essential resources (Tilman & Pacala 1993). The most contrasting species life-history trade-offs arise from resource limitation for light at optimal sites and belowground resource limitation (moisture and nutrients) in stressful environments (Dybzinski & Tilman 2007). The competitively dominant species are less tolerant of belowground limitations, but exclude other plants by their superior competitive ability above ground. Conversely, stress-tolerant species are weaker competitors under optimal growth conditions where strong light competition prevails (Grime & Hunt 1975). This trade-off has been proposed to explain the alternation of species with different resource acquisition strategies along moisture and other resource gradients (Liancourt et al. 2005; Gross et al. 2009; Mariotte et al. 2013; Doudová & Douda 2020; Douda et al. 2021), but the multiple-resource limitation model has not yet been applied to explain species RA patterns and their synchrony.
The multiple-resource limitation model implies that plant species differ in their RA responses based on their tolerance to above- and belowground resource availability. Competitively dominant species are expected to be mainly dependent on the availability of belowground resources, and as these resources become less available, RA is reduced (Reekie & Bazzaz 1987; van Lent et al. 1995; Kettenring et al. 2011; Johnson et al. 2017). Conversely, subordinate species, whose growth is mainly limited by aboveground competition, are expected to allocate more resources to seed production under stress conditions, when competition with dominant species is less intense (Chaloupecká & Lepš, 2004). In environments where belowground resources fluctuate, competition for light may be particularly intense during resource pulses, forcing subordinate species to accelerate their reproductive efforts relative to dominant species in resource scarcity period (Sun & Frelich 2011). As a result, subordinate species may be driven to flower either before or after optimal conditions for competitively dominant species (Martínková et al. 2002; Williamson & Ickes 2002; Catorci et al. 2012). Reproductive allocation patterns of subordinate species may therefore be more consistent with the resource-matching or switching hypothesis, as they need to produce more seeds quickly in the window of opportunity when competition for light with dominant species is temporarily reduced. Under resource-limited conditions, slower growing dominant species may gradually accumulate belowground resources, delaying future flowering, consistent with the resource budget hypothesis. While this may increase RA synchrony between subordinate species, it may also increase asynchrony between subordinate and dominant species during scarcity period.
Resource limitation may have different effects on intraspecific synchrony in RA compared to the interspecific level. An increase in intraspecific synchrony is associated with a reduction in intraspecific variability in RA. This reduction could lead to increased differentiation in RA between species, potentially increasing asynchrony at the interspecific level. Furthermore, evolutionary dynamics may also play a role in the observed differences between inter- and intraspecific synchrony. In particular, plants that synchronise flowering with conspecifics may have reduced reproductive costs. Predator satiation during years of abundant seed production (Silvertown 1980; Zwolak et al. 2022), increased pollination efficiency during periods of increased flowering (Nilsson & Wastljung 1987; Kelly et al. 2001), and individual kinship (Bogdziewicz et al. 2024) are factors that favour intraspecific rather than interspecific synchrony. In particular, subordinate species may be highly intra-specifically synchronised in flowering to maximise reproductive output, as they have a limited time to reproduce when belowground resources decline and competition from dominant species decreases. Dominant species may show a more dispersed reproductive pattern in years with favourable resource conditions.
Here, we tested for the first time the multiple-resource limitation model by quantifying the contribution of belowground resource limitation and light competition to species RA and community flowering patterns (i.e. the extent of synchronous RA both inter- and intraspecifically). We established an experimental wetland plant community and exposed plants to different drought stress regimes and a dominant removal treatment. The drought stress regimes simulated either fluctuating resource levels under interannual drought by alternating between well-watered (optimal) and dry (suboptimal) conditions, or sustained drought stress over several years (Douda et al. 2018; Doudová & Douda 2020). The interannual drought allowed us to assess whether interannual variation in RA is driven by actual resource levels, as predicted by the resource switching hypothesis. The permanent drought regime assessed the ability of species to accumulate resources over several years to increase RA. Both regimes were compared to a scenario of continuous water availability throughout the experiment. The effect of interspecific interactions on three subordinate species was estimated by removing a dominant species. Lastly, we investigated whether intraspecific synchrony responded to resource levels and species interactions in a similar way to interspecific synchrony.