Tomorrow's plant communities: different, but how?


tel +1 617 287 6614;
fax +1 617 287 6650;

Passively, plants everywhere sense the human influence on the environment. Most directly, immediately, and universally, plants respond to the rapid (> 0.5%) annual increase in the concentration of their carbon supply. But they also respond to changing climate patterns, deposition of acid and reactive nitrogen, the arrival and activity of new herbivores or competing species, and many other factors. With the environment changing so quickly, and in so many ways, there is an increasing demand for predictions of how ecosystems will look in the future, and an increasing challenge in making just those predictions.

Our understanding of how species will respond to environmental change has moved through several stages. Initial hypotheses based on first principles often gained credence from growth chamber studies that examined responses of single plants in pots to single environmental changes such as elevated CO2. However, some of these hypotheses were increasingly contradicted as researchers used more realistic settings: growing plants in nutrient-limited conditions or with competition, conducting studies outdoors in natural ecosystems, and discarding chamber-based studies for designs with less imposing infrastructure. Recently, several studies have examined responses of natural and managed ecosystems to multiple global changes, seeking to more closely and comprehensively simulate future conditions (Beier, 2004). In this issue, Williams et al. (pp. 365–374) provide an example of this ‘multifactor’ environmental change research, presenting the most comprehensive analysis to date of plant demographic responses in a natural ecosystem.

‘results from these and similar studies suggest that photosynthetic pathways are not certain predictors of competitive outcomes in a future atmosphere’

Winners and losers: testing an initial prediction for elevated CO2

How can we best predict ‘winners’ and ‘losers’ in a future environment? One approach has concentrated on grouping species by traits such as photosynthetic pathway. Decades ago, a straightforward working hypothesis suggested that C3 plants, known to have strongly CO2-limited photosynthetic rates, would more commonly outcompete CO2-concentrating C4 plants in a future, more CO2-rich atmosphere (e.g. Patterson & Flint, 1980). Growth chamber and glasshouse studies (typically measuring individual plants in pots) supported this hypothesis, as C3 plants increased shoot mass by an average of 45% under elevated CO2, vs a smaller (but still positive) 12% increase in C4 plants (Poorter & Navas, 2003). A strong C3 advantage also emerged in competition experiments (again, typically in pots) that provided ample nutrients to the plants. However, C3 species gained no advantage from elevated CO2 in competition experiments with limited nutrient availability (Poorter & Navas, 2003). Relatively few experiments have studied responses of C3 and C4 plants in natural ecosystems. In those cases, however, the straightforward working hypothesis rarely triumphs. Two cases in which CO2 preferentially benefited C3 species were identified in eastern North America. In Maryland salt marshes, Erickson et al. (2007) found that elevated CO2 increased growth of a C3 sedge, and accelerated its replacement of a C4 grass in a mixed C3–C4 community. In the understory of a planted Tennessee forest, an invasive C3 vine sometimes benefited from CO2 enrichment while an invasive C4 grass sometimes declined in these conditions (Belote et al., 2003). The story is not always so straightforward; in North American shortgrass steppe, recruitment increased in one C3 grass species and thus it produced more biomass in response to elevated CO2, while another C3 grass and a C4 grass did not respond (Morgan et al., 2004). In the tallgrass prairie of the same continent, Owensby et al. (1993, 1999) found that C4 plants sometimes benefited most from CO2 enrichment, presumably via the water savings created when plants growing in elevated CO2 narrow their stomatal openings. Together, results from these and similar studies suggest that photosynthetic pathways are not certain predictors of competitive outcomes in a future atmosphere – ecological processes such as competition for resources can complicate matters.

The changing climate also complicates predictions of C3–C4 competition, as photosynthetic rates of C4 plants typically respond more positively to warming. Before the experiment at the Tasmanian free-air CO2 enrichment (TasFACE) facility described by Williams et al., no field studies had tested how competition among plants of these different photosynthetic types responds to both warming and CO2. Unlike previous studies that have examined responses of plant growth or (less often) reproduction, Williams et al. examined the entire life cycle of four common species in a temperate grassland on the Australian island state of Tasmania. While the dominant C4 grass species was unaffected by CO2 or warming, CO2 unexpectedly reduced seed production of the dominant C3 grass species, reducing population growth over 3 yr. Together, CO2 and warming also reduced germination and establishment of the C3 grass, triggering population decline. In addition to the responses of the two grass species, Williams et al. examined responses of two invasive C3 weeds – both perennial members of the Asteraceae – with the intent of elucidating whether global changes will increase the abundance of these species over time.

Will future conditions bring more ‘unwelcome’ plants?

A variety of ‘unwelcome’ species – such as agricultural weeds, plants with irritating oils or allergens, and invasive plants in wildland areas – benefit from CO2 enrichment when grown in isolation in growth chambers or glasshouses (e.g. Ziska, 2003; McPeek & Wang, 2007; Ziska et al., 2007). Although these studies can be informative on several fronts, the predictive value of plant responses in these settings is essentially zero; biomass responses of isolated plants are not correlated with responses of those same species in the context of a diverse plant community (Poorter & Navas, 2003). Furthermore, in these settings, the CO2 responses of invasive species do not differ from those of noninvasive species (Dukes, 2000). Responses of species grown in monocultures better predict responses in mixed communities, but are far from perfect (Poorter & Navas, 2003). One example: above-ground biomass of the invasive annual forb Centaurea solstitialis increased by 70% in response to elevated CO2 in monocultures, and by an almost identical amount when grown in competition with a native Californian serpentine grassland community – but, importantly, the response in diverse communities was not statistically significant (Dukes, 2002).

Relatively few experiments have examined the responses of problematic species to CO2 or warming in more natural settings that include competition and realistic soil and environmental conditions. Results from these few studies highlight several cases where problem species benefit from elevated CO2. For instance, seedlings of an invasive broad-leaved evergreen tree grew faster in response to CO2 in a Swiss forest, while seedlings of a similar native species did not, suggesting that the invasive may become more common over time (Hättenschwiler & Körner, 2003). Growth and population biomass of poison ivy (Toxicodendron radicans), a native vine that produces an irritant to human skin, increased dramatically in North American forest exposed to elevated CO2 (Mohan et al., 2006). Adding to the problem, poison ivy grown in elevated CO2 produced a more potent form of the irritant. Other studies suggest some caveats regarding the responses of invasive species. For instance, the aggressively invasive grass Bromus madritensis responded strongly to elevated CO2 in a North American desert, but only in a wet year (Smith et al., 2000). In the understory of an eastern North American forest, an invasive vine benefited from CO2, but an invasive C4 grass declined – and each response was significant in only one of two years (Belote et al., 2003).

The TasFACE study described by Williams et al. is one of the first to examine the responses of invasive species to two or more types of environmental change at a time. So what happened in the grasslands of Tasmania? Good news: population growth rates of the invasive forbs did not respond to elevated CO2, and were strongly suppressed by warming (via reduced germination), suggesting that these species may disappear from the grassland over time.

Predicting community change

The study by Williams et al. highlights the utility of two approaches: first, a multifactor approach to the study of global change, which allows more comprehensive understanding of community and ecosystem responses, and secondly, a demographic approach to community change, in which responses of all life cycle stages are quantified for important species. This allows a more detailed and mechanistic understanding of why plant communities may change in future conditions, and may permit further extrapolation. For instance, if researchers determine individual life cycle effects of several separate factors, these may be combined to predict future population changes. However, one major obstacle remains: interactive effects.

Although Williams et al. identified cases where the response of a life cycle stage to combined CO2 and warming did not match expectations based on the single treatments, they did not attempt to explain the mechanisms behind these interactive effects. A better understanding of when and why such interactions arise would greatly benefit the field of global change research, and may grow out of studies similar to this one.

As studies such as that of Williams et al. bring us closer to predicting the future composition of a plant community, we must keep in mind the bigger picture. Global change responses of diseases, herbivores, mutualists, and other species may have important effects that alter the community trajectory from that suggested by plot-scale global change experiments, with important consequences for the eventual functioning of the ecosystem. The valuable approaches of Williams et al. are likely to be adopted by others, and should provide important clues to the future of other plant communities – but important challenges will remain.