However, competitive suppression of native seedlings by exotics eclipsed the feedback effect, highlighting the need for weed control in the restoration process.Understanding long-term plant community dynamics has been a key challenge in ecology , made even more challenging by the increased firequency of plant invasion and novel environmental conditions. A number of studies have demonstrated that new climatic extremes, such as severe and prolonged drought, may have long term legacies on plant communities by affecting resource availability and altering dominant species . In systems dominated by long-lived perennial plants, novel disturbances can disrupt generally predictable successional change . Understanding the impacts of novel disturbances on vegetation dynamics is even more challenging in annual-dominated plant communities, where composition can reset each year . Annual systems have high turnover in species composition due to multiple biotic and abiotic drivers with high inter-annual variability . Drivers such as the timing and amount of precipitation and temperature , fungal pathogens , and herbivores act on seed production, seed survival and germination, seedlings, and mature plants . In addition, vegetation can be influenced by the quantity of litter from the previous season . Thus, species composition in annual systems is difficult to predict and manage . These controlling variables are often strongly impacted by previous species composition, and thus predicting community changes in annual systems may benefit from considering priority effects. Priority effects are when the timing and order of species arrival during assembly determines which species can later establish in the community, leading to alternative stable states, vertical grow rack alternative transient states, or compositional cycles . Priority effects have been observed to affect composition and diversity in multiple systems .
In perennial systems, a long-lived species can physically hold space against competitors over time, occupying this niche to the exclusion of later arrivals . In fact, planting of perennial grasses is firequently used as a way to suppress annual weeds . Among annual plants, priority effects may occur through mechanisms such as litter build up, seed production and faster germination, and plant-soil feedbacks related to changes in pathogen and symbiont communities . Priority effects are stronger among species with high overlap in resource use, resulting in greater niche co-option and exclusion of later arrivals of similar function and phenology . Assessing vegetation community dynamics in California’s annual grasslands may thus benefit from defining functional groups in terms of phenology , which dictates when California grasses compete for the limited resource of soil moisture. California’s grasslands are dominated by early-season exotic annual species that are now so entrenched in the landscape they are considered naturalized; they generally outcompete native grass and forb seedlings, as they germinate and grow faster, depleting shallow soil moisture and creating light limitation . Several priority experiments have shown that native perennial grasses in this system benefit from a two-week head start over naturalized exotic species . A newer set of invading exotic annual species, including the grasses Elymus caputmedusae and Aegilops triuncialis, are classified as noxious weeds and are of particular concern because they produce thick layers of thatch that is slow to decompose and prevents germination of other species . These noxious weeds germinate at the same time as the early-season exotics, but their above-ground growth primarily occurs after rains have ended and the earlier season grasses have senesced .
In drier years, early-season naturalized exotic growth can limit noxious weeds by utilizing all soil moisture, but cannot prevent noxious weed growth if late-season moisture is present . Native perennial grasses, which are active during the late-season, may compete with noxious weeds for the dwindling late-season soil moisture. However, how long priority effects last on all three functional groups is unknown. We established field plots consisting of early-phenology naturalized exotic annual species, late-phenology annual noxious weeds, and native species commonly used in restoration mixes . Each group was seeded alone or simultaneously in mixtures with other groups to compare to assess the importance of being seeded without initial competition. All plots were then allowed to be naturally colonized by non-seeded species in the experiment, but with a subset of plots receiving a one-year weeding treatment to provide additional priority. We assessed composition over a twelve-year period that included an extreme multi-year drought followed by a historic wet year, and then another severe drought year. Such extreme weather events in other systems have been shown to disrupt plant community dynamics, reducing cover of resident species and increasing invasion . This experiment is particularly valuable because the majority of priority effect studies take place in the greenhouse and for less than a year in duration, and it is widely recognized that more field and long-term experiments are needed to determine the strength of priority effects in varying conditions . Of the few long-term studies, the current literature has shown both that priority effects either persist or disappear depending on the system and the species involved .
We hypothesized that all three functional groups of species, when seeded alone, will: have greater cover than when grown in competition, and limit recruitment of another naturally colonizing functional group. We also predict that the multi-year drought will weaken priority effects during and beyond the drought in favor of the early-season naturalized exotics, as they have will have first access to soil moisture.Plots were located in UC Davis Campbell Tract Experimental Site in Davis, CA . Prior to plot establishment, the land was used for agricultural purposes and then lay fallow for twenty-two years. The site was primarily on Reiff series soil, with a sandier lens of Brentwood soil series on 25% of the site . Under a Mediterranean climate, the site experienced hot, dry summers and wet, cold winters that correspond with the growing season of cool season grasses.Rainfall has high inter- and intra-annual variability, with mean annual precipitation of 445 mm . During the experiment, California experienced a historic drought between the 2012-2014 water years , followed by 2 years with precipitation levels that were substantially closer to the 30-year precipitation average, and then followed by one of the wettest years on record in 2017, then a dry year, followed by another historic wet year . Prior to initial seeding at the start of the experiment in fall 2007, the seed bank and resident vegetation were minimized by disking the soil, irrigating to germinate the seedbank, and spraying germinating seedlings with herbicide . Irrigation and herbicide treatments were performed twice before planting. Plots were seeded with one of the following vegetation treatments in a randomized block design : native perennial grasses and annual forbs, exotic annual noxious grasses, naturalized exotic annual grasses and forbs, and all possible combinations of these three groups , at a rate of 139 g seed/plot. Plots were 1.5 x 1.5 m with a 1m buffer between them . For each single functional group mix, equal proportion of seeds of each species were added. For functional group mixes, an equal proportion of each functional group seed was added, commercial vertical hydroponic systems with equal proportion of individual species within each functional group. Given that the common design of seeding equivalent weights across all treatments design results in the multi-group treatments having less seed per species group , short-term conclusions may be influenced by seed limitation as well as priority effects. However, the effects of seed limitation should be short-lived given the annual species in this experiment are highly prolific and able to rebound from extremely low to high cover from one year to the next, as well as naturally self-thinning . While natives are seed limited, low density can still result in high cover , and so changes in native cover can be attributed to changes in individual size as well as population response. Percent cover of each species present was assessed visually with a modified Daubenmire bin method of the inner 1m x 1m core.
Composition was measured twice each season annually from 2008 to 2019 to capture peak flowering species with different phenologies . The highest cover value across the two sample points each year for each species was used for analysis.There were multiple levels of priority in this experimental design. For a given functional group, the seeding treatment in which they were seeded alone in monotype is the highest level, as they accessed the space first with no competition from other functional groups. The seeding treatments in which that functional group were seeded simultaneously with another group are a secondary level of priority, as they shared first access to the space. The remaining seeding treatments are those in which the functional group in question was not seeded, thus they had no priority at all. If present later in the experiment, they recruited into the space after initial colonization by the planted functional group, and thus faced high competition for resources from established residents or high propagule pressure . Given the spatial distribution of the experimental plots, every plot was close to a seed source of another functional group from neighboring plots and the unplanted walkways between plots and blocks became quickly colonized, indicating that the species were dispersing throughout the site quickly. Superimposed on the above listed priority levels was a weeding treatment to determine the difference between one vs two years of priority. Natural colonization was either allowe dimmediately or postponed an extra year due to hand-weeding after the first year to maintain the initial species compositions of the seeding treatments . Since then, all plots were occasionally weeded for agricultural weeds that were not part of the study but species that self-recruited and are typical species in California grasslands were not weeded.Priority effects can be the outcome of the seeded species having both higher propagule pressure and biotically resisting the recruiting species and we are interested in the long-term trajectory of the initially planted communities and the extent to which they are naturally colonized by functional groups that have been planted nearby. To test our hypothesis, we performed two comparisons. First, for each year, we compared cover of the recruiting functional group in its monotypic seeding treatment, which presumably represents the best conditions for performance as it has the highest level of priority, to the three seeding treatments in which the functional group was not planted . If the recruiting functional group cover is higher in its own monotypic treatment, then it would suggest that the functional groups originally planted are providing some level of resistance. Second, we compared the recruiting functional group’s cover as it changed over time within each seeding treatment it was not planted into, such as whether recruiting species cover stays consistently low or increases with time. This took into consideration potential annual population fluctuations due to environmental conditions and provides detail on resistance not captured by the first comparison. Both comparisons are necessary to determine whether low presence of a recruiting functional group is due to bioticresistance or poor environmental conditions. We also compared cover between the one- and twoyear priority weeding treatments. Both comparisons were assessed with the same linear mixed effect model but differed in how we performed the post-hoc analysis. First, we fit linear mixed effect models with percent cover of each recruiting functional group as the dependent variable, specifying seeding treatment, priority length , year , and their 3-way interaction as fixed effects and plot and block as random effects. Significance was tested with analysis of variance . Both comparisons were assessed with post-hoc multiple comparison tests using estimated marginal means on the interaction of seeding treatment and year when significant in the ANOVA but differed in the order of terms used. The first compared cover across the different seeding treatments within each year while the second compared over across the 12 years within each seeding treatment. All statistical analyses were conducted in R 4.0.3 . Linear mixed effect model fitting was performed using “lme4” and “lmerTest” . Multiple comparison tests were conducted using “emmeans” .Our study assessed the role of priority effects in determining long-term community composition in a system that experienced high annual variability in weather, including a multiyear drought event followed by one of the wettest years on record, and then another historic dry year, and another historic wet year .