Journal list menu

Volume 106, Issue 1 e4504
ARTICLE
Open Access

Warming-induced changes in seasonal priority effects drive shifts in community composition

Emma Dawson-Glass

Corresponding Author

Emma Dawson-Glass

Research Department, Holden Arboretum, Kirtland, Ohio, USA

Department of Ecology and Evolutionary Biology, University of Michigan, Ann Arbor, Michigan, USA

Correspondence

Emma Dawson-Glass

Email: [email protected]

Search for more papers by this author
Rory Schiafo

Rory Schiafo

Research Department, Holden Arboretum, Kirtland, Ohio, USA

Negaunee Institute for Plant Conservation Science and Action, Chicago Botanic Garden, Glencoe, Illinois, USA

Program in Plant Biology and Conservation, Northwestern University, Evanston, Illinois, USA

Search for more papers by this author
Sara E. Kuebbing

Sara E. Kuebbing

The Forest School, Yale School of the Environment, Yale University, New Haven, Connecticut, USA

Yale Center for Natural Carbon Capture, Yale University, New Haven, Connecticut, USA

Search for more papers by this author
Katharine L. Stuble

Katharine L. Stuble

Research Department, Holden Arboretum, Kirtland, Ohio, USA

Search for more papers by this author
First published: 15 January 2025

Handling Editor: Tadashi Fukami

Abstract

Shifting community assembly dynamics are an underappreciated mechanism by which warming will alter plant community composition. Germination timing (which can determine the order in which seedlings emerge within a community) will likely shift unevenly across species in response to warming. In seasonal environments where communities reassemble at the beginning of each growing season, changes in germination timing could lead to changes in seasonal priority effects, and ultimately community composition. We test this expectation by assembling mesocosms of 15 species in one of two orders—“ambient” assembly order or “warmed” assembly order—based on the order in which the constituent species germinated under ambient and warmed conditions. Community composition differed significantly between mesocosms assembled in ambient versus warmed orders. The impact of assembly order on species mean biomass was largely explained by how much earlier (or later) a species arrived in the warmed-order treatment relative to the ambient-order treatment. Species whose germination phenology advanced more under warmed conditions relative to ambient conditions showed greater relative increases in biomass under the warmed assembly treatment. These findings demonstrate that warming can drive community assembly and shape community composition by reordering the relative timing of germination among species. These findings enhance our ability to predict which species are likely to benefit from warming and which may decline based on how warming may shift assembly order, ultimately informing how warming may alter plant communities.

INTRODUCTION

As temperatures rise globally, today's plant communities differ from those of the past, as well as those we will likely see in the future (Bertrand et al., 2011; Feeley et al., 2020; Liu et al., 2018; Shi et al., 2015). Plant community composition is central to determining many ecosystem functions (Loreau et al., 2001; van der Plas, 2019) making critical our ability to predict how communities shift in response to warming. That said, our understanding of the many drivers of community-level responses to warming is still developing. As ecologists increasingly come to understand the critical role of community assembly in shaping plant communities (Chase, 2003; Fukami, 2015), assembly dynamics may offer another key mechanism driving how warming alters plant community composition.

Priority effects, in which early arriving species influence the establishment and growth of later arriving species, are a key community assembly mechanism (Ejrnæs et al., 2006; Fukami, 2015; Kardol et al., 2013). Early arriving species tend to be more successful in establishment, growth, and reproduction relative to later arriving individuals (Vaughn & Young, 2015; Weidlich et al., 2021). This may be because early arrival into a community allows species greater access to limiting resources (Kardol et al., 2008, 2013), reduced competition with other species (Young et al., 2017), and/or more space and time to grow (Ross & Harper, 1972). The impacts of priority effects can result from both long-term arrival differences and arrival differences of mere days or weeks (Young et al., 2017). In seasonal environments, differences in arrival order at the beginning of the growing season can result in “seasonal priority effects” that reshape communities annually (Wainwright et al., 2012). As such, priority effects can be critical in shaping community composition (Ejrnæs et al., 2006; Körner et al., 2008).

Among the earliest and most frequently documented ecological impacts of warming are shifts in phenology (Parmesan & Yohe, 2003), with spring phenologies often advancing (Inouye, 2008; Stuble et al., 2021). However, phenological sensitivities differ from species to species (Cleland et al., 2012), meaning that while many species will advance their phenology, they will do so at different rates. This variability in phenological shifts across species can result in changes in interactions among species (Heberling et al., 2019; Kharouba et al., 2018; Renner & Zohner, 2018; Weaver & Mallinger, 2022). When considering community assembly dynamics, the timing of species arrivals can be influenced by the timing of phenophases including seed set, dispersal, and germination. However, these phenophases are all susceptible to shifts as the climate warms (Stuble et al., 2021). Thus, species that typically benefit from seasonal priority effects by releasing seed, dispersing, or germinating early during assembly under today's climatic conditions may not necessarily be the same species that benefit from seasonal priority effects in the future (Buonaiuto & Wolkovich, 2023). Such shifts in assembly dynamics, particularly seasonal priority effects, are likely critical (yet underappreciated) in reshaping communities under climate warming.

Germination is a critical phenophase in the context of community assembly, determining, in part, when a seedling will emerge within a community (Donohue et al., 2010). Germination phenology can shape species coexistence (Blackford et al., 2020; Leverett, 2017), and warming-induced changes in germination phenology have been found to induce seasonal priority effects among species pairs (Buonaiuto & Wolkovich, 2023; Wainwright & Cleland, 2013). Specifically, species with higher phenological sensitivity are more likely to advance germination timing (Buonaiuto & Wolkovich, 2023; Wainwright & Cleland, 2013; Zettlemoyer & Lau, 2021), ultimately benefiting from warming-induced seasonal priority effects. Such phenological sensitivity may vary with native status, with many non-native species expected to be more responsive to warming than natives (Stuble et al., 2021; Wolkovich & Cleland, 2011). While such findings suggest that warming-induced shifts in germination timing are likely to restructure communities via changes in assembly dynamics, this has not yet been tested experimentally in multi-species assemblages (i.e., >2 species).

As warming alters germination timing (Stuble et al., 2021), and does so unevenly across species (Cleland et al., 2012), changes in assembly dynamics will likely restructure communities (Buonaiuto & Wolkovich, 2023). Such warming-induced seasonal priority effects are likely a critical, but underappreciated mechanism reshaping communities in response to ongoing climate warming. Here, we explore whether warming-induced changes in germination phenology can influence community assembly processes, thereby shaping community composition. Specifically, we assembled mesocosms of 15 common old-field species, ordering seedling “arrival” based on the timing of germination under either present-day (“ambient”) or warmed (+3°C) conditions. We harvested the aboveground biomass from these mesocosms to ask three interrelated questions: (1) Does assembly order (ambient or warmed germination order) influence plant community composition? (2) Which species benefit from warming-driven changes in assembly order? (3) Can differences in species' performance associated with changes in assembly order be explained by relative shifts in germination timing, native status, and/or functional type?

METHODS

Study system

We assembled greenhouse mesocosms using 15 old-field plant species common to Northeast Ohio. Old fields are common throughout the eastern United States as agricultural lands are abandoned. They are populated by communities of fast-growing herbaceous species that can serve as an ideal study system to understand how warming may restructure community assembly dynamics. Species used in this mesocosm study included a mix of forbs and grasses of both native and non-native species commonly found in old-field ecosystems (Figure 1a, Appendix S1: Table S1).

Details are in the caption following the image
(a) Timeline of the arrival of each species in ambient and warmed assembly orders, as determined in the germination experiment and (b) germination day for each species in each assembly order, and number of days gained or lost under warmed assembly order relative to ambient as determined in the germination experiment.

Germination phenology

To determine the order and timing of species assembly under warmed and ambient conditions for our mesocosm experiment, we first germinated the 15 species in growth chambers simulating present day (hereafter “ambient”) and warmed (+3°C) spring conditions (increasing both photoperiod and temperature on a weekly basis; Table 1), in line with anticipated warming trends in this region (IPCC, 2013). We placed 50 seeds of each of the 15 species in sand-filled trays in the growth chambers. Sand was twice-autoclaved for 1 h to sterilize it before use in the germination trials. Trays were cleaned with BioShield (Greensboro, NC). Sand was kept moist using deionized water throughout the trials. Prior to simulating spring conditions, all seed trays (with seeds) were held in dark growth chambers at 2°C for 9 days and then at 3°C for 3 days. To generate spring conditions, we slowly increased photoperiod and temperature in the chambers, with daylength mimicking the photoperiod in northeastern Ohio from March 22 to July 12 (Table 1). Lights were off in chambers for the first 22 days of the experiment because the chambers were unable to maintain the necessary low temperatures with the lights on. Photoperiod and temperature were increased in tandem once weekly. Ambient weekly temperature settings for the growth chambers were based on long-term soil (at 2 cm) temperature monitoring at the Holden Arboretum from 2006 to 2010. Daytime and nighttime temperatures were calculated using soil temperature data collected from 8 am to 7 pm and 8 pm to 7 am, respectively.

TABLE 1. Growth chamber conditions for germination phenology trial indicating weekly photoperiod settings, as well as daytime and nighttime temperatures simulating ambient and warmed conditions.
Simulated week Photoperiod Ambient daytime temperature (°C) Ambient nighttime temperature (°C) Warmed daytime temperature (°C) Warmed nighttime temperature (°C)
22-Mar NA 4.1 3.4 7.1 6.4
29-Mar NA 5.2 4.4 8.2 7.4
5-Apr NA 6.3 5.5 9.3 8.5
12-Apr NA 7.4 6.6 10.4 9.6
19-Apr 13 h 30 min 8.5 7.7 11.5 10.7
26-Apr 13 h 48 min 9.7 8.9 12.7 11.9
3-May 14 h 5 min 10.9 10.0 13.9 13.0
10-May 14 h 21 m 12.0 11.1 15.0 14.1
17-May 14 h 35 m 13.1 12.2 16.1 15.2
24-May 14 h 47 m 14.1 13.2 17.1 16.2
31-May 14 h 58 m 15.0 14.2 18.0 17.2
7-Jun 15 h 5 m 15.9 15.0 18.9 18.0
14-Jun 15 h 10 m 16.7 15.8 19.7 18.8
21-Jun 15 h 12 m 17.3 16.5 20.3 19.5
28-Jun 15 h 10 m 17.9 17.1 20.9 20.1
5-Jul 15 h 6 m 18.4 17.6 21.4 20.6
12-Jul 14 h 58 m 18.7 18.0 21.7 21.0
  • Abbreviation: NA, not applicable.

Seeds were allowed to germinate in the growth chambers for 17 weeks. We noted the day of maximum germination (i.e., the day on which the most seeds germinated) for each species (Appendix S1: Figure S1). In instances where there were 2 days with equally high germination, we averaged the days. We then used the day of maximum germination under warmed and (separately) ambient conditions to determine the order and timing of species introduction into the mesocosm experiment. Note that one species (Cirsium arvense) did not germinate under warmed conditions. In ambient conditions, it had low germination and was the last species in the trial to germinate. As such, we conservatively added this species at the end of the experiment for both assembly treatments, in line with the maximum germination timing in ambient conditions. Similarly, Ranunculus acris did not germinate in our germination phenology trial. As such, we used the same max germination dates as Ranunculus pensylvanicus for our planting order.

Note that the goal of this germination trial was to inform assembly timing in our subsequent mesocosm experiment. We acknowledge that this germination trial cannot fully replicate the multiple germination cues that would occur under field conditions and a more robust germination experiment would be necessary to fully understand how germination timing is likely to change with warming. However, our findings can be used to inform probable shifts in relative germination timing for exploration of how such changes can translate into shifts in community assembly dynamics, and ultimately compositional changes in plant communities.

Mesocosm experiment

To understand how shifts in the timing and order of community assembly alters community compositions, we assembled communities in greenhouse-based mesocosms. We planted seedlings of all 15 old-field species into pots, based on the order that they germinated in the growth chamber (as determined by date of maximum germination, described above). Pots were assembled in one of two orders based on the timing of germination: warmed assembly order and ambient assembly order (Figure 1). Species were added to the mesocosms between June 23, 2020, and August 14, 2020 (a 53-day span). Note that sometimes a species arrived alone, while sometimes multiple species were added on a given day, as determined by the timing of germination in the germination phenology experiment (Figure 1). We established 12 replicates of each treatment (24 mesocosms in total). Each species was added to the mesocosms as a seedling (rather than seed), allowing us to precisely manipulate assembly order of plants into the system and isolate the impacts of priority effects (Stuble & Souza, 2016). Seedlings were initially established in flats in the greenhouse, with initial seed sowing staggered such that all species germinated around the same time. Seedlings were fertilized once in their germination trays with 20-20-20 all-purpose water-soluble fertilizer and treated for white flies. Seedlings added to the mesocosm experiment varied slightly in age, with somewhat younger seedlings added at the beginning of the experiment relative to those added at the end. Importantly, this difference in age was only a matter of days for each species (the species with the biggest shift in germination timing was Rumex crispus, which arrived 19 days earlier under warmed conditions), and any advantage provided by arriving slightly larger (older) when added later in the experiment would only serve to make our results more conservative by potentially muting priority effects. As such, any findings of community shifts associated with changes in assembly order should be relatively robust. We measured seedling height at the time of planting into mesocosms to account for variability in initial plant height (including potential differences associated with plant age), which we include in later analyses. Mesocosms were assembled in round bulb pans (25.4 cm diameter by 12.7 cm deep) filled with Miracle-Gro Moisture Control Potting Mix. Using shallow pot-based mesocosms allowed us to identify how warming-induced seasonal priority effects impact community dynamics in a competitive environment. We note that additional study using field-based experiments would be necessary to further validate our finding under natural conditions, but this controlled experiment allowed us to isolate the mechanisms driving changes in community composition resulting from changes in germination phenology with warming.

Mesocosms were maintained in a common greenhouse at an average of 20°C and were watered as needed (daily to every-other-day). We treated the mesocosms for white flies once in July 2020. We ended the experiment 98 days after initiation because mesocosms had grown to capacity and plants were beginning to senesce. All aboveground biomass was clipped at soil level and sorted by species. We were unable to sort belowground biomass for the 15 species. We dried all aboveground biomass at 60°C for a minimum of 72 h before weighing.

Statistical analyses

To assess if assembly order influenced the composition of plant communities, we analyzed plant community composition across mesocosms as a function of assembly treatment (ambient or warmed assembly order) using PERMANOVA (adonis2 function, vegan package; Oksanen et al., 2022). We used a Bray–Curtis dissimilarity matrix based on the final aboveground biomass of each species. Pseudo-F and p values were calculated based on 10,000 permutations of the data. We used nonmetric multidimensional scaling (NMDS) to visualize differences in community composition by assembly order.

To assess which species benefited from warming-driven changes in assembly order, we first constructed a phylogenetic tree including all 15 species (phylo.maker, V.PhyloMaker2 package; Jin & Qian, 2019, 2022; trees derived from Smith & Brown, 2018; Zanne et al., 2014). We used a phylogenetic linear mixed model to assess species-specific biomass responses to assembly treatment (Almer function; evolvability package; Bolstad, 2021), with log-transformed biomass of each species as the response variable, and assembly treatment, species, and their interaction as fixed effects in the model. We included initial seedling height and pot ID as random effects to account for variability in initial seedling heights and multiple measurements within each mesocosm across species, respectively. We used a phylogenetically corrected model as species biomass was related to phylogeny and therefore not independent based on Blomberg's K and Pagel's lambda (Appendix S1: Table S2) (phylosig function; phytools package; Medeiros et al., 2020; Revell, 2024). As such, we incorporated phylogeny as a variance–covariance correlation matrix (vcv function, ape package; Paradis, 2024) in the random effects of the above-described model. To assess how biomass varied by assembly order for each species, we then calculated a pairwise comparison of estimated marginal means using a Tukey multiple comparisons adjustment (emmeans and contrast functions; emmeans package; Lenth, 2023). To determine whether assembly order treatments influenced cumulative community biomass, we used an ANOVA with total pot biomass (the summed biomass of all species per meosocosm) as the response variable and assembly treatment as the predictor.

To assess how assembly dynamics influenced species' performance, we first calculated an assembly treatment effect size (ambient vs. warmed assembly order) on species' biomass. We calculated this effect size as the natural log of the mean biomass for each species when assembled in the warmed assembly order divided by the mean biomass of each species when assembled in the ambient assembly order. Positive effect sizes indicated that a species' biomass was greater when assembled in the warmed assembly order relative to ambient, while negative values indicated that a species' biomass was less when assembled in the warmed assembly order. We then explored how shifts in species' relative biomass (effect size) associated with assembly treatment could be explained by species' functional group (grass or forb), species' native status (native or non-native), or the difference in germination timing among assembly treatments (germination day in the ambient treatment—germination day in the warmed treatment). We first tested for phylogenetic nonindependence of each species' effect size using the methods described above. We found that effect sizes were not correlated with phylogenetic relatedness (Appendix S1: Table S2). As such, we did not account for phylogeny when analyzing species' effect sizes. We used an ANOVA with the treatment effect size (on biomass) as the response variable and difference in germination date, functional group, and native status as fixed effects.

All analyses were performed in R version 4.3.1 (R Core Team, 2023).

RESULTS

Plant community composition differed significantly between the two assembly treatments (i.e., warmed order or ambient order) (pseudo-F = 16.61, p < 0.001; Figure 2), though total community biomass did not differ significantly between the treatments (F1,22 = 1.253, p = 0.275). Individual species exhibited variable responses to the warming-driven changes in assembly order. Some species did significantly better (i.e., produced more biomass) when assembled in the warmed assembly order (viz., R. crispus, Solidago rugosa, and Vernonia gigantea; Figure 3, Appendix S1: Table S3), while some species did significantly better when assembled in the ambient assembly order (Hypericum punctatum, R. pensylvanicus, and Solidago canadensis; Figure 3, Appendix S1: Table S3), and many saw no significant difference in biomass production related to assembly order (Andropogon gerardii, Bromus inermis, C. arvense, Dactylis glomerata, Hypericum perforatum, R. acris, Rudbeckia hirta, Sorghastrum nutans, and Tridens flavus; Figure 3, Appendix S1: Table S3).

Details are in the caption following the image
Communities assembled in the ambient assembly order significantly differed in species composition from those assembled in the warmed assembly order. Each point represents a mesocosm, with blue circles representing mesocosms assembled with the ambient assembly order, while red triangles represent mesocosms assembled with the warmed assembly order. Ellipses represent 95% CLs. Black text represents each species used in our mesocosms, and gray arrows represent the relative direction and magnitude of the influence of a given species on the composition of each assembly treatment.
Details are in the caption following the image
Mean aboveground biomass of each species when assembled in ambient and warmed assembly orders. Panels are ordered by change in germination date with warming (relative to ambient), with the species in the top left panel (Rumex crispus) gaining the most days (19), and the species in the bottom right corner (Solidago canadensis) losing the most days (−5) (Figure 1, Appendix S1: Table S1). Significant differences in biomass across treatments (Appendix S1: Table S3) are denoted with asterisks (*p < 0.05; **p < 0.01; ***p < 0.001).

The responses of species' relative biomass to assembly order (effect size) were positively related with change in germination date (F1,11 = 26.33, p < 0.001) such that species that arrived earlier in the warmed assembly order produced more biomass in the warmed assembly treatment as compared to when assembled in the ambient assembly order (Figure 4). The relationship between effect size associated with assembly order did not differ significantly with native status (F1,11 = 0.01, p = 0.937) or functional group (F1,11 = 2.78, p = 0.127).

Details are in the caption following the image
Generally, greater advances in germination timing with warming generated greater increases in biomass under warmed assembly orders relative to ambient. Each point represents a species, and error bars represent the SE of the log response ratio. The black line represents the line of best fit between these two variables, and gray shading indicates 95% CIs. On the x-axis, positive values indicate earlier germination in the warmed assembly order relative to ambient assembly order, while negative values indicate the reverse. Most points fall in the positive space along the x-axis because most species advanced seed germination timing with warming. On the y-axis, positive effect size values indicate a species produced more biomass when assembled in the warmed assembly order relative to the ambient assembly order, while negative values indicate a species produced greater biomass when assembled under the ambient order of assembly relative to the warmed order of assembly.

DISCUSSION

Priority effects during community assembly are known to be an important driver of community composition (Fukami, 2015) and, as such, may provide a mechanism via which warming reshapes communities. Germination phenology is likely to shift with warming, leading to warming-induced changes in seasonal priority effects and, ultimately, changes in community composition. Here, we demonstrate that warming-driven changes in assembly order associated with changes in germination timing can, in fact, alter plant community composition (Figure 2). Notably, the number of days a species advanced its germination timing was a strongly correlated with the species' biomass response to assembly treatment (Figure 4), while non-native status and functional type showed no significant relationship between biomass response and assembly order. These findings indicate that species' phenological sensitivity (here, the ability to shift germination in response to warming) is a key predictor of plant community responses to a warming world.

Our study is not the first to provide evidence that species with higher phenological sensitivity are likely to be successful with warming. Indeed, ability to shift phenology in response to warming is associated with improved performance (Cleland et al., 2012) and abundance (Willis et al., 2008) in many plant species. However, we are among the first to document community-level responses to warming-induced shifts in assembly dynamics. Our findings lend urgency to the call to expand study of germination phenology as a key mechanism that can reshape communities (Buonaiuto & Wolkovich, 2023), a phenophase that, until now, has received limited attention in the context of warming (Iler et al., 2021; Stuble et al., 2021). Beyond predicting how natural communities are likely to shift with warming, expanded understanding of how communities will reassemble under warmed conditions can also help to better inform management strategies. For example, understanding how germination timing shapes species emergence into a community could facilitate improved restoration strategies that time species introductions with a desired compositional outcome (Stuble & Young, 2020; Weidlich et al., 2021; Young et al., 2017). As species with more plastic phenology advance germination with warming, practitioners may need to be more proactive in managing assembly order to maintain desired outcomes.

Most species arrived earlier in the warmed assembly treatment (Figure 1), and those species were (for the most part) subsequently larger in the warmed assembly treatment (relative to the ambient assembly treatment) (Figure 4). Part of the success of some species over others could be tied to the germination requirements of each species. For example, the earliest arriving species in the warmed order (B. inermis and D. glomerata; Figure 1) and one of the species that saw a more than 10-day advance in germination day with warming (S. nutans; Figure 1) all have no cold stratification requirements (USDA and NRCS, 2024a, 2024b, 2024e). Notably, however, R. crispus, which saw the greatest relative shift in germination timing with warming (Figure 1), has been found to benefit from cold stratification in its native range (Pérez-Fernández et al., 2019). Weaker chilling requirements could potentially be linked with greater responsiveness to warming (Polgar & Primack, 2011), though this hypothesis has not been extensively tested for germination requirements (but see Buonaiuto & Wolkovich, 2023). Conversely, two species (H. punctatum and S. canadensis) exhibited delayed germination in the warmed assembly treatment (relative to the ambient treatment) (Figure 1), and those species were correspondingly smaller in the warmed assembly treatment (relative to the ambient) (Figures 3 and 4). However, there is no clear rationale for the delayed phenology in these species—they do not have similar cold stratification requirements to each other (USDA and NRCS, 2024c, 2024d) and other species in the same genera as these two species advanced germination with warming (Figure 1, Appendix S1: Table S1). Temperatures above optimal conditions can result in delayed time to germination, in part, due to physiological stress and/or inhibition of normal dormancy-break requirements (Bradford, 2002). Our findings could reflect a difference in optimal temperature ranges among species, resulting in delayed germination in some species and advances in others. Our findings further emphasize the need for an enhanced understanding of germination requirements and how these requirements relate to phenological sensitivity to warming. A more comprehensive knowledge of germination phenology would aid in further predicting how community assembly dynamics are likely to change with warming.

Surprisingly, responses to assembly treatments among species were unrelated to native status. This runs counter to expectations that many non-native species in North America stand to benefit from warming due to greater phenological sensitivity than their native counterparts (Miller et al., 2023; Reeb et al., 2020). In North American old-field ecosystems, most non-natives occupy earlier phenological niches than natives, which typically flower later in the season (Reeb et al., 2020). In our study, two non-native species (B. inermus and D. glomerata) already occupied the earliest phenological niches in our germination trials (Figure 1). As such, it is possible that these non-native species were constrained in the benefits they could gain by advancing phenology. That said, not all non-native species in our study were constrained by an already early germination niche. Notably, non-native R. crispus advanced its phenology by 19 days when germinated in warmed conditions, moving from arriving on Day 20 to Day1 (Figure 1) and accumulated significantly more biomass in the warmed assembly order (Figure 3). Increased biomass in R. crispus came at the expense of the two very early non-native grasses (B. inermis and D. glomerata). Consequently, B. inermis and D. glomerata generated less biomass (albeit not significantly so) when assembled in the warmed order (Figure 3). Taken together, this could indicate that, as early arrivers in this system already, some non-natives may have less to gain in terms of assembly advantages associated with warming, particularly if warming increases direct competition with one another (Cleland et al., 2015).

Here, we based assembly order on observed shifts in germination phenology due to warming, but many additional factors could affect assembly dynamics in association with climate change. For example, climate change will also alter factors such as precipitation, CO2 concentration, and the probability of extreme weather events (IPCC, 2013). Like temperature, precipitation can also be important in shaping germination phenology (Donohue et al., 2010) and can influence priority effects (Wainwright et al., 2012). As plants are likely to experience shifts in multiple environmental factors, further study exploring multiple germination cues will be key to understanding assembly dynamics (Wolkovich & Donahue, 2021). Additionally, other plant phenophases likely to shift with warming may also affect arrival timing, such as seed dispersal (Zou & Rudolf, 2023). Warming may also create phenological mismatches between interacting species, such as plants and seed dispersers (Warren II et al., 2011), further disrupting assembly processes and complicating predictions of future assembly dynamics. Finally, our mesocosms were grown under common greenhouse conditions, rather than warmed and ambient conditions (i.e., only germination timing was manipulated with warming). Growing the mesocosms under common conditions allowed us to isolate the impacts of shifts in assembly order. However, warming during plant growth would also likely alter growth rate (Wu et al., 2012), adding an additional layer of complexity to understanding plant community responses to warming. Additional study of multiple environmental cues, phenological, and growth responses, and their implications for community assembly would be extremely valuable in further predicting plant community responses to global change (Wolkovich et al., 2022).

A final limitation of our study is that we did not measure biomass for each species when grown alone, and thus cannot determine how biomass differed when not in combination with other species. As such, we cannot mechanistically isolate if seasonal priority effects drive the dynamics observed in our study, per se. Specifically, we cannot disentangle whether species are responding to changes in species interactions driven by different assembly orders, or rather reflecting mere changes in arrival time into mesocosms (and thus more time to grow), independent of other species. However, these findings demonstrate that assembly order is important in driving species biomass, and, as assembly order shifts with warming, so too will community composition. Additional studies explicitly testing the exact mechanisms underpinning the patterns we see here (e.g., niche preemption vs. niche modification; Fukami, 2015) will help to further elucidate how seasonal priority effects will reshape communities under warming (Buonaiuto & Wolkovich, 2023).

CONCLUSION

Our findings demonstrate that warming-induced changes in assembly order driven by shifts in germination phenology can significantly change plant community composition. Understanding that changes in assembly dynamics associated with warming will alter the composition of plant communities provides critical new information to help predict how communities are being reshaped by warming. Seasonal priority effects are already known to be an important mechanism shaping community responses to factors such as restoration (Weidlich et al., 2021; Young et al., 2017) and non-native species invasions (Cleland et al., 2015; Wainwright et al., 2012). Integrating warming-induced shifts in assembly order and subsequent seasonal priority effects via changes in plant phenology into studies of the ecological impacts of warming will improve our ability to forecast community and ecosystem responses.

ACKNOWLEDGMENTS

We thank C. Hewins for her help in conducting the germination experiment. We also thank the Holden Arboretum's Research Renegades Community Science Group for their help in collecting the biomass from our mesocosm experiment. Finally, we thank N. Sanders for valuable feedback on drafts of this manuscript.

    CONFLICT OF INTEREST STATEMENT

    K. Stuble is a subject-matter editor for Ecology but did not participate in the handling of this manuscript. The authors otherwise declare no conflicts of interest.

    DATA AVAILABILITY STATEMENT

    Data (Dawson-Glass et al., 2024a) are available in Dryad at https://doi.org/10.5061/dryad.sbcc2frh4. Code (Dawson-Glass et al., 2024b) is available in Zenodo at https://doi.org/10.5281/zenodo.13988289.