In the same way that species diversity can positively influence ecosystem processes and functions, genetic diversity within species can also improve the stability and functioning of populations, particularly during stressful events (Reusch et al., 2005; Hughes et al., 2008; Salo and Gustafsson, 2016). Increasing the number of genotypes within a population can enhance productivity (Crutsinger et al., 2006; Aguirre and Marshall, 2012) and increase resilience (Massa et al., 2013), with benefits that cascade to the wider ecosystem (Crawford and Rudgers, 2013). It is therefore expected that a population with a wider range of genotypes and corresponding phenotypes will have better ‘insurance’ against the effects of environmental stress (Foster Huenneke, 1991).

In relatively stable environments where biotic and/or abiotic conditions remain constant, differences in the success of particular genotypes can lead to the dominance of one or a few clones, and thus an overall decrease in the number of genotypes over time (McLellan et al., 1997). Should environmental conditions change, however, low diversity populations are considered at a greater risk of extinction compared to genetically diverse populations. Genotypes that are well-adapted to local conditions will not necessarily perform well in altered conditions. When genotypic diversity is high, it is statistically more likely that one or more individuals will have a suitable genotype that will thrive under the new conditions. The mechanism by which this occurs is known as complementarity, whereby the inclusion of a variety of genotypes, and thus phenotypes, allows the population access to different pools of resources, limiting competition between individuals while at the same time improving the likelihood of population success under changing conditions (Hughes et al., 2008). Ultimately, increasing the pool of genetic diversity among clonal genotypes improves evolutionary potential and adaptive capacity, particularly when faced with environmental disturbance.

Organisms that utilize predominantly clonal growth strategies are usually expected to have naturally low levels of genetic variation (Ellstrand and Roose, 1987, although there are exceptions, e.g., Sánchez-Vilas et al., 2010) and the responses of individuals to stress, including rapidly changing conditions, often involves phenotypic plasticity – the ability of an organism to modify its phenotype in response to their environment (Nicotra et al., 2010). The expression of phenotypic plasticity can vary among genotypes (Bradshaw, 2006; Lande, 2009). For example, a plastic response to environmental change can be as simple as stunted growth when nutrients are lacking, but the degree to which growth is limited by nutrients may differ among genotypes, representing genetic differences in their tolerance to stress (Schlichting, 1989). Predicting the responses of populations to environmental stress thus requires an understanding of the genetic variation in stress tolerance (Munday et al., 2013), commonly assessed by experimental designs in which the performance of individual genotypes is quantified across environments (e.g., Clark et al., 2013; Foo et al., 2014). If genotypes vary in their plastic responses (indicated by a genotype by environment interaction) then greater genetic diversity should lead to a higher likelihood of populations including tolerant genotypes.

Some of the most compelling demonstrations that genotypic diversity influences ecosystem functioning have come from experimental manipulations of diversity in seagrass ecosystems (Procaccini et al., 2007). Seagrasses provide an ideal model ecosystem to test ideas of genotypic diversity because they reproduce both sexually through seed and pollen dispersal and clonally via the vegetative growth of rhizomes, generating variation in genetic diversity within and among meadows (e.g., Ort et al., 2012; Evans et al., 2014). Over the past decade, experiments have shown that increasing the number of genotypes within an experimental plot can increase resistance to grazing (Hughes and Stachowicz, 2004), influence grazer biomass (Hughes et al., 2010), enhance resistance and resilience to invasion (Massa et al., 2013), and result in greater shoot densities and biomass compared to monocultures during disturbance events (Reusch et al., 2005; Hughes and Stachowicz, 2011). There is also evidence to suggest that rehabilitation and restoration efforts can be significantly enhanced by including a range of different genotypes sourced from multiple donor locations (Procaccini and Piazzi, 2001; Williams, 2001; Reynolds et al., 2012).

The relationship between genotypic diversity and resilience in seagrass meadows is particularly important in plant responses to extreme events, such as heat waves (Reusch et al., 2005), unusually intense grazing (Hughes and Stachowicz, 2004) and eutrophication (Arnaud-Haond et al., 2010). Less well understood is whether genotypic diversity alone can buffer against the effects of more typical, short-term environmental stresses that are not necessarily extreme and do not always result in high shoot mortality (Salo et al., 2015).

In this study, we examined how individual genotypes vary in their response to short-term stress in the threatened seagrass Posidonia australis. We use this predominantly clonal species to contrast how individual genotypes, and meadows of varying genetic diversity, perform under a simulated stress due to increased shading. Reduced light availability due to turbidity is one of the most common stressors facing seagrass meadows globally (Waycott et al., 2009). Rapid increases in turbidity are often caused by anthropogenic inputs to coastal waters, particularly via fine sediment re-suspension, nutrient over-enrichment and pollution (Ralph et al., 2006). Such turbidity increases are most often derived from industrial and agricultural run-off into estuaries and bays, commercial dredging, coastal construction, stormwater, and sewage (Pergent-Martini and Pergent, 1996; Short et al., 1996).

By utilizing carbohydrates stored within the rhizome, seagrasses can endure short periods of light reduction (for example during ‘pulsed’ turbidity events such as those following storms or heavy rain; (Longstaff and Dennison, 1999). However, extended periods of shading can severely impair growth rates and productivity (e.g., Collier et al., 2012), resulting in short, stunted growth forms (Walker et al., 1999), reduced shoot densities and biomass, and ultimately widespread decline (Short and Wyllie-Echeverria, 1996).

The genotypic diversity of P. australis, as measured using microsatellite DNA markers, varies greatly among meadows in south-eastern Australia (Evans et al., 2014), and this diversity reflects phenotypic variation in functionally relevant traits such as leaf morphology and productivity (Evans et al., 2016). Slow-growing and large seagrass species such as P. australis and P. oceanica are seemingly more tolerant to extended periods of shade than faster growing species such as Zostera marina and Halophila ovalis, which generally show rapid responses to shading (reduced shoot densities, high shoot mortality, and rapid recovery; (Ochieng et al., 2010; Yaakub et al., 2014). Currently, for slow-growing species that are poor colonizers, such as P. australis, one of the only definitive indicators that the seagrass is under stress is meadow dieback (Gobert et al., 2006). For conservation purposes, this is obviously not an acceptable indicator of stress, as by then the plants have already reached a critical state.

Shading studies conducted in seagrass meadows often have a long duration of months to years, in which seagrass recovery following the removal of shades is also recorded (e.g., Fitzpatrick and Kirkman, 1995; Collier et al., 2009; McMahon et al., 2011, but see Salo et al., 2015). More often than not, records of P. australis meadows that have experienced shoot fatality/decreased density have not shown any signs of recovery (Walker and McComb, 1992; Fitzpatrick and Kirkman, 1995), or recovery has been extremely slow (Meehan and West, 2004). In the context of rehabilitation and conservation efforts, it is therefore most important to determine the susceptibility of seagrasses to shading stress prior to fatality, considering that shoot mortality is the endpoint we wish to avoid. Understanding what factors can predict resilience and tolerance to stress are therefore required to underpin the future monitoring and management of these meadows. Consequently, we performed a short-term shading experiment in situ to answer whether individual genotypes within meadows varied in their response to short-term light reduction, using leaf growth rates and photosynthetic efficiency as indicators of performance. We then used these genotype-level results compiled across six meadows to model the likelihood of including a high-performing genotype within meadows of differing genotypic diversity levels.

The use of shade cloth significantly reduced the levels of photosynthetically active radiation (PAR, 400–700 nm) in contrast to the mesh and unshaded control treatments (F_2,24 = 8.93, P = 0.001). PAR under the shades averaged 20.2 ± 2.4 μmol photons m^-2 s^-1, compared to PAR adjacent to controls, which averaged 237 ± 19.7 μmol photons m^-2 s^-1. The reduction in light created by the shaded treatments was comparable to that caused by ‘high’ natural turbidity levels in coastal benthic habitats (Longstaff and Dennison, 1999).

The growth rates of P. australis varied among meadows, and the effects of shading varied among the meadows (Figure 2A, significant meadow × treatment interaction, Table 2). This was due to a significant reduction in growth rates due to the shaded treatments in two meadows (Lake Macquarie and Jervis Bay; Figure 2A). Both of these meadows are considered to have very low genetic diversity, with only three genotypes present (Evans et al., 2014). Shading treatments did not significantly impact the growth rates of the three ‘higher’ genotypic diversity meadows (Botany Bay, Port Stephens, and St. Georges Basin), nor the meadow with the lowest genotypic diversity, Wallis Lake, which supports only two distinct genotypes (Table 2). The effects of shading did not vary among individual genotypes within a meadow (genotype × treatment interaction non-significant, Table 2), visualized by the slope and elevation of their reaction norms across treatments (Figure 3A).

Values of F_V/F_M under the shaded treatments significantly differed among genotypes, but not among meadows (Figures 2B, 3B and Table 3). To model the effect of meadow genotypic richness on sensitivity to light reduction, we first compiled the mean values of F_V/F_M for each sampled genotype across all meadows. The values of F_V/F_M among shaded treatments varied between 0.68 and 0.81 and we considered plants showing F_V/F_M values of 0.8 or greater to be ‘high performers,’ or the plants best acclimated to light reduction. F_V/F_M responses from all individual genotypes (two or three per meadow) were then combined to create a theoretical pool of regional genotype responses. Repeated resampling of the pooled data generated the likelihood that a meadow with a given number of genotypes (from 2 to 16) would produce one or more high performers. Modeled simulations showed that for a theoretical meadow from our sampled region to have a 90% or greater likelihood of having one or more genotypes performing at ≥0.8, the presence of at least four genotypes would be required (Figure 4). A 90% chance of including a genotype with the higher performance of 0.81 would require approximately eight genotypes.

Our results show that the majority of Posidonia australis meadows are resilient to short-term shading stress. However, a subset of meadows in two low diversity estuaries were particularly vulnerable to the early effects of shading. We found that different genotypes within meadows varied in their physiological responses to shading, and estimate that approximately four or more genotypes are required to ensure a high probability of including a ‘high-performing’ (i.e., higher photosynthetic efficiency) genotype in a theoretical meadow consisting of random genotypes from the range of meadows sampled.

The vast majority of P. australis meadows on the east coast of Australia grow in shallow (1–5 m) waters just below the low tide mark (West, 1990) and are thus subjected to rapid changes in light from turbidity. Changes in light availability can trigger a wide range of physiological responses in seagrasses (Kumar et al., 2016) and our results suggest that in the short-term, P. australis meadows are generally very resilient to reduced light. For the majority of meadows, shading did not significantly impact seagrass growth rates over the course of the experiment. This contrasts with the findings of a similar study performed in Zostera marina, where shading over a similarly short time frame resulted in suppressed eelgrass growth rates for a range of genotypes (Salo et al., 2015). This may be explained by the fact that Z. marina is a faster-growing species with a smaller rhizome and reduced likelihood of translocation of stored resources.

We found that the degree to which shading affected growth varied among meadows as a result of two meadows with low genotypic richness, Lake Macquarie and Jervis Bay, both experiencing a ∼60% reduction in mean growth rates per meadow under the shaded treatments compared to controls. The growth of P. australis varied among genotypes, but there was no interaction between genotype and shading treatment, indicative of variation in stress tolerance among genotypes within a meadow. While these two meadows share a low genotypic richness, we cannot discount other environmental variables that were not measured and are likely to vary on large scales among all our sampling sites (e.g., salinity and nutrient supply). Lake Macquarie and Jervis Bay, however, are very different estuaries. Lake Macquarie is characterized by industrial and residential development, with a small P. australis meadow that was listed as threatened under the Australian Environment Protection of Biodiversity Act 1999 in 2015. Contrastingly, Jervis Bay is an open embayment, where some of the largest P. australis meadows on the east coast of Australia are protected within a marine park (Creese et al., 2009). This suggests that the overall vulnerability of Lake Macquarie and Jervis Bay P. australis meadows to short-term stress is at least partly due to the existence of particular genotypic individuals within the meadows that are less resilient to light reduction than others rather than shared environmental conditions.

The mechanisms behind which genotypic diversity can enhance performance under stress include both complementarity and dominance (Hughes and Stachowicz, 2009). For strong dominance to occur, a small number of genotypes will become very abundant within a population; a common occurrence in slow-growing, predominantly clonal seagrass meadows such as P. australis (Evans et al., 2014). Meadows of P. australis on the east coast of Australia are presumably long-established (following the stabilization of sea levels after the last glacial maximum ∼6,500 years. As these meadows are geographically fragmented and unlikely to experience any contemporary gene flow via seed or pollen dispersal (Evans et al., 2014), the dominant clones existing within these meadows are assumed to be capable of withstanding local environmental changes by having highly plastic phenotypes. Indeed, our results show that different genotypes within meadows vary widely in their physiological and morphological response to reduced light (Figures 2B, 3). A change in growth rates as a plastic response to shading may initially assist in shoot survival, but may not necessarily benefit the plant in the long-term. Prolonged exposure to shading can induce photoacclimation to the lower light regime, which is associated with a drop in respiration rates, resulting in shorter, stunted growth forms and lower productivity (Collier et al., 2012). The rapidly reduced growth rates of P. australis in Lake Macquarie and Jervis Bay, suggests these vulnerable clonal plants are at risk of loss or damage under rapid environmental change (e.g., reduced light availability due to increased turbidity), regardless of the plasticity of existing genotypes.

The variation among seagrass genotypes in their tolerance to shading adds to the recent literature that quantifies the degree of intraspecific variation in tolerance to a range of environmental stressors in marine organisms (e.g., to ocean warming in macroalgae, Clark et al., 2013; to warming and acidification in sea urchins, Foo et al., 2014). Thus, complementarity via a greater variety of genotypes is still considered most beneficial in providing an ‘insurance’ effect against disturbance. We modeled how increased numbers of genotypes could be advantageous during environmental disturbance using the photosynthetic efficiency data (F_V/F_M), in which higher values represent greater acclimation to reduced light. Terrestrial research suggests that an optimal value of F_V/F_M for over 70 different plants under non-stressed conditions is 0.83 (Björkman and Demmig, 1987), with values lower than this indicative of plant stress or possible damage to photosystem II (Roháček and Barták, 1999). Healthy seagrass plants may record values slightly lower than this (e.g., 0.75 for Halophila ovalis; Beer et al., 2014). As our study recorded values ranging from 0.68 to 0.81, we chose a ‘high-performance’ threshold of 0.8, with shoots obtaining values equal to or greater than this considered those best acclimated to low-light conditions.

Repeated resampling of the pooled photosynthetic performance data using varying numbers of genotypes in each sample revealed that there is a 90% or greater likelihood of including a high performing genotype when four or more genotypes are included within a meadow (i.e., those with photosynthetic efficiency values of 0.8 or greater; Figure 4). To go one step further, obtaining a 90% likelihood of including a genotype with an average F_V/F_M of 0.81 or greater (the highest recorded value in this study) would require eight or more genotypes. It should be noted that this is a very simplified model of a theoretical meadow consisting of randomly selected genotypes from all those we sampled, although it does utilize realistic data from the 17 genotypes sampled in situ. Additionally, while photosynthetic efficiency as measured by F_V/F_M is commonly used as a measure of plant stress and performance, it does represent only a snapshot of efficiency at the instant the measurement is made. Thus, it is important to note that assumptions are made when interpreting the overall ‘performance’ of plants based on F_V/F_M data, and morphological measures such as growth are perhaps a more accurate representation of overall performance. Nonetheless, the results of the simulation support our original hypothesis that including more genotypes is always advantageous in terms of adaptive capacity, no matter how well-adapted one particular genotype is to the existing local conditions. This has significant implications for P. australis meadows that are almost entirely made up of only one to three genotypic clones, such as those found on the south-eastern coastline of Australia, close to the species range edge (Evans et al., 2014).

As individual genotypes that exhibit high photosynthetic performance under reduced light are increasingly likely to be found in meadows with four or more genotypic clones, we can hypothesize that under longer durations of reduced light (months to years), meadows with a high number of genotypes will also be better able to maintain high productivity. This is an obvious advantage, particularly given the unprecedented rate of environmental change experienced by seagrass meadows in the 21st century (Unsworth et al., 2014). An ever increasing human population size has the potential to further impact already vulnerable seagrasses in estuarine environments (Orth et al., 2006): coastal housing developments, land reclamation and commercial dredging have all been directly linked to seagrass decline from reduced light as a result of increased turbidity (Short et al., 1996; Lotze et al., 2006; Ralph et al., 2006; Walker et al., 2006; Yaakub et al., 2014). Past exposure to prolonged turbidity does not always predict future performance under additional stress, with seagrasses either performing better (Maxwell et al., 2014) or worse (Yaakub et al., 2014) than those from more ‘pristine’ meadows. It will also be important to investigate how stress due to changes in light availability may interact with additional stressors such as temperature (e.g., York et al., 2013).

With P. australis already listed as near-threatened on the IUCN Red List, effective management of future declines and impacts is a necessity, and must utilize information on individual meadow and genotype performance, particularly under stressful conditions. This research highlights the importance of incorporating genotypic information when investigating differences in plant stress response. While most P. australis meadows were generally resilient to short-term shading stress, variation in stress response is likely driven by the performance of individual genotypes. Although low diversity meadows may contain genotypes that perform well under stress, increasing the number of genotypes within a seagrass meadow greatly increases the likelihood of including a high performing genotype that will be more resilient to a future changes in light conditions. These results have significant implications for restoration and rehabilitation efforts, particularly for the sourcing of appropriate donor meadows and genotypes for transplantation.

SE, AV, and AP conceived and designed the study, SE performed the experiments, SE and AP analyzed the data. SE, AV, and AP wrote the paper.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.