Abstract
Quantifying changes in plant growth and interspecific interactions, both of which can alter dominance of plant species, will facilitate explanation and projection of the shifts in species composition and community structure in terrestrial biomes expected under global warming. We used an experimental warming treatment to examine the potential influence of global warming on plant growth and interspecific interactions in a temperate steppe in northern China.
Six dominant plant species were grown in monoculture and all 15 two-species mixtures for one growing season under ambient and elevated temperatures in the field. Temperature was manipulated with infrared radiators.
Total biomass of all the six plant species was increased by 34–63% in monocultures and 20–76% in mixtures. The magnitude of the warming effect on biomass was modified by plant interactions. Experimental warming changed the hierarchies of both competitive response and competitive effect. The competitive ability (in terms of response and effect) of one C4 grass (Pennisetum centrasiaticum) was suppressed, while the competitive abilities of one C3 forb (Artemisia capillaris) and one C3 grass (Stipa krylovii) were enhanced by experimental warming. The demonstrated alterations in growth and plant interactions may lead to changes in community structure and biodiversity in the temperate steppe in a warmer world in the future.
Introduction
Global warming has raised the Earth's surface temperature by ∼0.76°C and is predicted to increase it further by 1.8–4.0°C over this century (IPCC 2007). The unprecedented rise in temperature has well been documented to cause shifts in species composition (Harte and Shaw 1995; Klein et al. 2004; Schuur et al. 2007; Weltzin et al. 2003), changes in community structure (Alward et al. 1999; Walker et al. 2006; Weltzin et al. 2003) and even local/global extinctions of plant species (Klein et al. 2004; Memmott et al. 2007; Thomas et al. 2004) in terrestrial ecosystems. The above observations indicate increases in the dominance of some plant species or decreases in the dominance of others, which could result from differential growth responses of plant species to elevated temperature (Llorens et al. 2004; Williams et al. 2007) or changes in plant interactions under warmer conditions (Heegaard and Vandvik 2004; Klanderud 2005). However, our understanding of whether these two aspects occur separately or concurrently and whether they affect each other in natural communities is limited.
Plant interactions are the key drivers in structuring terrestrial plant communities (Callaway and Walker 1997; Fowler 1986). Attempts to predict the responses of plant communities to climate change may fail if plant interactions are not explicitly considered (Davis et al. 1998). However, it remains controversial whether and to what extent global warming changes plant interactions in terrestrial plant communities. For example, climate warming modifies the effect of species removal in an alpine community (Klanderud 2005) but not in an Arctic tundra (Hobbie et al. 1999). The contrasting results might be caused by the complexity and difficulty in studying plant interactions in natural ecosystems where many species co-occur. Species removal experiments can only address the effects of one species on all other co-occurring species or the impacts of all other co-occurring species on that species, in which direct interactions between any two species are largely neglected.
The competitive ability of one species is a combination of competitive effect (i.e. the ability of a species to affect other species) and competitive response (i.e. the ability of a species to avoid being affected) (Goldberg and Landa 1991; Miller and Werner 1987). In natural ecosystems, plant species can be ranked by competitive hierarchy (an ordered ranking from competitive dominant to competitive subordinate) based on either competitive effect or response (Keddy 2001). Shifts in the position of plant species in the hierarchies of competitive effect or response imply changes in their competitive ability, with consequent alterations in plant interactions, dominance and community structure. Previous studies have shown that competitive ranking of plants can be affected by nutrient conditions (Austin et al. 1985; Tilman 1990). Here we show for the first time that competitive hierarchies can also change under the influence of climate warming.
The temperate steppe in northern China represents the regional vegetation in the vast area across the Eurasian continent and is sensitive to climate change (Christensen et al. 2004; Niu et al. 2008b). With six dominant species planted in monoculture and 15 two-species mixture, this study was conducted to examine how experimental warming and plant interaction affect growth of different plant species in the temperate steppe in northern China, and to investigate if experimental warming changed plant interactions and thus the relative positions of species in the hierarchies of competitive response and effect.
Materials and Methods
Study site
The study was conducted in Duolun County (N 42°2′29″, E 116°17′20″), a semi-arid area located in the southeast part of Inner Mongolia Autonomous Region, China. Averaged over the past several decades (1953–2004), mean annual precipitation in this area is 385.5 mm, with peaks in July and August. Mean monthly air temperatures range from −17.5°C in January to 18.9°C in July, with mean annual temperature of 2.1°C. The soil is a chestnut soil according to the Chinese classification or Haplic Calcisols according to the FAO classification, with 62.75±0.04% sand, 20.3±0.01% silt and 16.95±0.01% clay. Mean bulk density is 1.31 g cm−3 and pH 6.84±0.07. The predominant species are Stipa krylovii Roshev, Agropyron cristatum (L.) Gaertn, Cleistogenes squarrosa (Trin.) Keng, Potentilla acaulis (L.) and Artemisia capillaris Thunb. These species grow in mixed stands with a mean density of 200–300 individuals per square meter and annual aboveground biomass of 120–200 g m-2.
Experimental design
Two 3×4 m plots were dug to a depth of 0.45 m and backfilled after installation of PVC tubes that were not closed at the bottom. One plot was heated continuously using an infrared radiator and the other was not heated (control). One 1.65×0.15 m infrared radiator (Kalglo Electronics Inc, Bethlehem, Pennsylvania) was suspended 2.25 m above the ground in the warmed plot. Soil temperatures were spatially uniform in the warmed plots in this study. Previous studies also demonstrated that infrared radiators induced spatially uniform changes in soil temperature (Wan et al. 2002a) and increased competition of Ambrosia psilostochya (Wan et al. 2002b) in a tallgrass prairie in USA. In the control plot, one ‘dummy’ heater with the same shape and size as the infrared radiator was suspended 2.25 m above the ground to simulate the shading effects of the radiator. The distance between the control and the warmed plot was ∼5 m to avoid heating the control plot by the infrared radiator. Soil temperature and moisture were monitored at three points in each plot. Soil temperature was measured by a Longstem Thermometer 6310 (Spectrum Technologies Inc, Plainfield, USA) and soil volumetric moisture was measured by a Diviner 2000 Portable Soil Moisture Probe (Sentek Pty Ltd., Balmain, Australia).
Plant material
Six dominant perennial species, including two C3 grasses (S. krylovii, A. cristatum), two C3 forbs (A. capillaris, P. acaulis) and two C4 grasses (Pennisetum centrasiaticum, C. squarrosa) that co-occur in the steppe of northern China were selected. Seedlings with similar size were transplanted from adjacent areas to the PVC tubes in mid May. The PVC tubes (11 cm in internal diameter and 50 cm in length) were buried in the two plots with the top end of the tubes 5 cm above the ground. The tubes were filled with sieved and mixed soil from the dug place. Species were planted in monoculture (two individuals of one species per tube) or mixture (two individuals including two different species per tube) in the tubes. There were totally 21 species combinations (each with 20 replicates) including six monocultures (one for each species) and 15 two-species mixtures. The tubes with different species combinations were arranged randomly in each plot. Two weeks after transplanting (1 June 2005), the warmed plot was heated continuously (24 h day−1) until the end of the growing season (late September).
Biomass measurements
All the plants were destructively harvested by the end of September 2005. The soil was carefully removed from the root system. The roots were thoroughly rinsed. Aboveground and belowground parts were oven-dried at 65°C for 48 h and weighed.
Calculating competitive response and competitive effect
The RYP values form a matrix where columns consist of target species (i), and the rows consist of neighbor species (j) with which the target species are grown. Column means correspond to the average competitive responses of target species, whereas row means correspond to average competitive effects of neighbor species (Bossdorf et al. 2004; Goldberg and Landa 1991; Keddy 2001; Wilson and Keddy 1986). A species with a high value for competitive response or a low value for competitive effect has a high competitive ability and should become dominant in a mixture, whereas a species with a low value for competitive response or a high value for competitive effect has a low competitive ability and is predicted to be subdominant in a mixture. Because competitive responses and effects may not be strongly negatively correlated among species, we ranked the six species both in terms of competitive response and competitive effect.
Data analysis
The effects of warming on soil moisture, soil temperature, plant biomass and the effects of competition on biomass, as well as the comparison of competitive response/effect among the six species under ambient or elevated temperature, were all analysed by one-way analysis of variance (Duncan test). All the statistics were performed by SPSS 11.0 for windows (USA). Because there was only one plot each for control and warming (confounding of treatment and plot), reported warming effects include potential differences between plots.
Results
Plant biomass
Over the entire experimental period, mean soil temperature (10 cm in depth) was 1.04°C higher in the plot with the infrared radiator than in the control plot (P < 0.001). Soil moisture at the depth of 0–10 cm was slightly (P = 0.462) lower in the warmed plot (12.5% V/V) than in the control plot (13.7% V/V) (Fig. 1). Warming significantly increased root biomass in P. centrasiaticum, A. capillaris, A. cristatum, C. squarrosa, and S. krylovii, aboveground biomass in A. cristatum and total biomass in A. capillaris, A. cristatum, C. squarrosa, and S. krylovii (P < 0.05). Root (P = 0.056), aboveground (P = 0.466), and total biomass (P = 0.079) of P. acaulis were not affected by experimental warming. In monoculture, root, aboveground, and total biomass of the six plant species were 40–96%, 12–64%, and 34–63% greater under elevated than under ambient temperature, respectively (Fig. 2).
Changes in soil temperature (10 cm depth) and moisture (0–10 cm depth) induced by an infrared heater over the experimental period.
Aboveground and root biomass per plant (mean ± 1 SE) under ambient and elevated temperatures and in monoculture and two-species mixture for the six target species in competition with the same six species as neighbors (Competitor). Abbreviations for species names: P.c (Pennisetum centrasiaticum), Ar.c (Artemisia capillaris), A.c (Agropyron cristatum), P.a (Potentilla acaulis), C.s (Cleistogenes squarrosa), S.k (Stipa krylovii).
Competition changed the magnitude of warming effect on plant biomass. For example, when averaged across all the two-species mixtures, warming-induced increases in total biomass of A. capillaris and C. squarrosa were much higher in mixture (77 and 61%) than in monoculture (34 and 46%, Fig. 2), whereas warming-induced increments in total biomass of P. centrasiaticum and S. krylovii were much lower in mixture (20 and 21%) than in monoculture (45 and 37%).
Competition significantly affected root and total biomass in A. cristatum and C. squarrosa and aboveground biomass in A. cristatum (P < 0.05) but had no effect on biomass of the other four species (P > 0.05). Under ambient temperature, A. cristatum and C. squarrosa had 2% and 28% higher total biomass, respectively, in mixture than in monoculture. Furthermore, there was a clear difference in species behavior in different two-species mixtures as assessed by biomass. For example, total biomass of P. acaulis was 0.44 g per plant when grown with A. cristatum, but was 1.18 g per plant when grown with S. krylovii under ambient temperature (P < 0.05, Fig. 2).
Warming changed the magnitudes and directions of plant interactive effects on biomass in some species. In comparison with its monoculture, total biomass of A. capillaris in mixture with A. cristatum was decreased by 40% (P < 0.05) under ambient temperature but increased by 48% under elevated temperature. In contrast, competition between P. centrasiaticum and P. acaulis increased the total biomass of P. centrasiaticum by 57% in the control plot but decreased it by 15% in the warmed plot (Fig. 2).
Competitive response
Significant differences (P < 0.05) in competitive responses were observed only between P. acaulis and C. squarrosa. Cleistogenes squarrosa had competitive response values above 1 (P < 0.05) and P. acaulis had values below 1 (P = 0.06 in the control and P < 0.05 in the warmed plots) under both ambient and elevated temperatures (Fig. 3), suggesting positive responses of C. squarrosa and negative responses of P. acaulis to the presence of other species. Under ambient temperature, the hierarchy of competitive responses among the six species was in the order: P. acaulis < A. capillaris < S. krylovii < P. centrasiaticum < A. cristatum < C. squarrosa.
Hierarchies of competitive response (mean ± 1 SE) among the six plant species under ambient (upper panel) and elevated temperatures (lower panel). A large value indicates high competitive ability. See Fig. 2 for species abbreviations. Different letters indicate significant (P < 0.05) differences among species.
Experimental warming did not change the competitive response values of the six species (P > 0.05) but increased the differences among species. Competitive response of C. squarrosa was significantly higher than those of P. acaulis, P. centrasiaticum and S. krylovii, while the competitive response of P. acaulis was also significantly lower than that of A. capillaris and A. cristatum (P < 0.05; Fig. 3). In addition, competitive response of P. centrasiaticum decreased most (0.194), whereas that of A. capillaris increased most (0.290) among the six species in response to experimental warming, leading to exchange of their position in the hierarchy of competitive responses. The other four species retained their positions in the hierarchy of competitive responses (P. acaulis < P. centrasiaticum < S. krylovii < A. capillaris < A. cristatum < C. squarrosa) under experimental warming (Fig. 3, lower panel).
Competitive effect
Stipa krylovii and P. acaulis had competitive effect values above 1 (P < 0.05), which means positive impacts of these species on their co-occurring species under ambient temperature. The other four species had competitive effect values similar to 1 (P > 0.05; Fig. 4, upper panel). The hierarchy of competitive effects under ambient temperature was in the order: A. capillaris < P. centrasiaticum < A. cristatum < C. squarrosa < S. krylovii < P. acaulis, among which P. acaulis was significantly higher than A. capillaris, P. centrasiaticum, A. cristatum and C. squarrosa (P < 0.05). Experimental warming reduced the competitive effect from 1.19 to 0.96 in S. krylovii (P < 0.05). An increase of 0.144 under warming was observed for the competitive effect of P. centrasiaticum, even though this change was statistically insignificant. As a consequence, S. krylovii and P. centrasiaticum switched their positions in the hierarchy of competitive effects under warming as compared with control. Warming did not alter the positions of the other four species in the hierarchy of competitive effects (Fig. 4, lower panel).
Hierarchies of competitive effect (mean ± 1 SE) among the six plant species under ambient (upper panel) and elevated temperatures (lower panel). A low value indicates high competitive ability. See Fig. 2 for species abbreviations. Different letters indicate significant (P < 0.05) differences among species.
Discussion
Effects of warming and plant interactions on biomass
Plant productivity has been reported to increase (Hartley et al. 1999; Rustad et al. 2001; Wan et al. 2002b, 2005) or decrease (Hobbie et al. 1999; Olszyk et al. 1998) by experimental warming. In our study, warming increased biomass production of all the six species (not significantly in P. acaulis, Fig. 2). The warming-induced increases in plant productivity could have resulted from enhanced plant photosynthesis due to higher temperature or longer time period suitable for plant growth activities. In the same experiment, seasonal mean photosynthetic assimilation of the two species (A. capillaris and S. krylovii) was 15 and 19% higher under elevated than ambient temperature (Niu et al. 2008a), contributing to the stimulation of growth and biomass production of these two species observed in this study. The indirect effect of greater nutrient availability, resulting from increased N mineralization (Rustad et al. 2001; Wan et al. 2005) might also have contributed to the biomass increases. Both direct and indirect effects of warming could be particularly important in the temperate steppe of northern China, which tend to be both temperature and nutrient-limited (Christensen et al. 2004; Yuan et al. 2005).
In comparison to monocultures, plant interactions in mixtures can exert positive or negative impacts on the growth of some species (Bossdorf et al. 2004; Callaway et al. 2002; Howard and Goldberg 2001; Silletti et al. 2004), which in turn may influence plant community structure and plant responses to environmental change. In the temperate steppe of northern China, we observed reduced biomass of A. capillaris when grown with C. squarrosa or A. cristatum and increased biomass of P. centrasiaticum when grown with P. acaulis (P < 0.05, Fig. 2). These results suggest potentially large impacts of changed plant interactions on community structure. The assessment of plant interactions between species could help to understand how species coexist in the field and how natural vegetation is structured. In our study, the warming-induced changes in biomass of target species varied depending on the specific neighbor species (Fig. 2), suggesting that these interspecific interactions may change the growth responses of the six tested plant species to elevated temperature.
Warming changes species competitive hierarchy
Species competitive ability could be ordered into competitive hierarchy, which has been reported in the previous studies (Keddy et al. 2000; Miller and Werner 1987). Positions of plant species in the hierarchies of competitive effect or competitive response are the outcome of interspecific competition of co-occurring species and can be used to predict long-term community dynamics (Goldberg 1990). Experimental warming changed the position of A. capillaris in the hierarchy of competitive response from the second under ambient temperature to the fourth under elevated temperature, whereas no change in its position in the hierarchy of competitive effect (Figs 3 and 4). The raised position of A. capillaris in the hierarchy of competitive response suggests that the growth and dominance of this species will be strengthened in the natural community in a warmer world in the future. The largest reductions in competitive effect value and a consequently lowered position of the C3 grass species S. krylovii in the hierarchy of competitive effect (Fig. 4) also implies a strengthening of its competitive ability in the community under elevated temperature. Experimental warming decreased the response value (Fig. 3), but increased the effect value (Fig. 4) of the C4 grass P. centrasiaticum, thus reducing its position in the competitive response hierarchy but raising its position in the competitive effect hierarchy. The opposite directions of the shifted positions of P. centrasiaticum in the hierarchies of competitive response and effect strongly suggest weakened competitive ability of this species in response to elevated temperature.
The direct effects of warming via differential stimulation of growth and production could have, at least partially, contributed to the altered plant competition (Fig. 2). In addition, the indirect effects of warming through changing soil water and nutrient regimes could also have resulted in the altered plant competition. Reduced soil moisture content under warming (Fig. 1) might have exacerbated the water limitation in this area, with differential impacts on the six species that have different water use efficiency (WUE). Among the five (A. capillaris, P. centrasiaticum, S. krylovii, P. acaulis and A. cristatum) of the six species for which photosynthesis were monitored in the same experiment, warming induced the greatest increase in WUE in A. capillaris. Improved WUE could partially explain the enhanced competitive ability of A. capillaris under experimental warming.
Pennisetum centrasiaticum is an early-successional species in the temperate steppe and usually dominates in disturbed habitats (Zhang et al. 2004). Under changed environment induced by experimental warming, the competitive advantage of late-successional species over pioneer species could have been enhanced (Morin and Chuine 2006). The weakened competitive ability of P. centrasiaticum could lead to accelerated replacement of this species by other five late-successional species under warmer conditions.
Both dependence (Austin et al. 1985; Tilman 1990) and independence (Fowler 1982; Keddy et al. 2000, 2002) on the environment of species competitive hierarchy have been reported previously. For example, Keddy et al. (1994) found that environmental factors affected the hierarchy of competitive response but had no effect on the hierarchy of competitive effect. Observations in this study showed that competitive hierarchies of both competitive effect and response could be changed by experimental warming. Alteration of the hierarchies of competitive effect or response under elevated temperature in the temperate steppe in northern China could potentially lead to differential changes in the growth, biomass production, coverage and dominance of different plant species, with consequent shifts in species composition and community structure under climate warming. Weakened competitive ability of P. centrasiaticum under elevated temperature indicates a probable decrease in dominance of this species in the temperate steppe in northern China under global warming.
Limitations of experiment
Although our experiment allowed us to quantify competitive abilities under experimental warming, there were some limitations of our study. The short time period used in this study limits the potential for extrapolation to longer time scales. For practical reasons, this experiment, like many others (Goldberg and Barton 1992; Keddy et al. 1994) ran for one growing season and this may have led to biased estimates of competitive performance of slower growing species. Long-term experiments are needed in the future to investigate whether and how the competitive rankings of effect and response change through time. In addition, there was only one plot each for control and warming (confounding of treatment and plot) in this study. Therefore, warming effects might include potential differences between plots.
Conclusions
As part of the Eurasia grassland biome and located in semi-arid regions, the temperate steppe in northern China is sensitive to climate change (Christensen et al. 2004, Niu et al. 2008b). Given its large area, changes in species composition in the temperate steppe could substantially impact carbon cycling at both ecosystem and regional scales. Irrespective of the short time period, our experiment clearly showed changed species competition in terms of competitive effect and response in the temperate steppe in response to experimental warming. With strengthened competitive ability of some species (A. capillaris and S. krylovii), weakened competitive ability in other species (P. centrasiaticum) and consequent changes in competitive hierarchy, we expect possible changes in species dominance, community structure and biodiversity in this biome in northern China under a warmer world in the future.
Supplementary Data
Supplementary data is available at Journal of Plant Ecology online.
Authors thank three anonymous reviewers and Bernhard Schmid for valuable suggestions to improve the original manuscript. We thank Yanfang Zhang, Zhixiong Li, Mingyu Wu, Jianyang Xia, Wenjing Cheng and Shihuan Song for the setting up of infrared heaters and field assistance. This study was finally supported by the National Natural Science Foundation of China (90511006) and Chinese Academy of Sciences (Hundred Talents Program).
