- Top of page
- Materials and methods
1. It has been suggested that much of the elevated CO2 effect on plant productivity and N cycling in semi-arid grasslands is related to a CO2-induced increase in soil moisture, but the relative importance of moisture-mediated and direct effects of CO2 remain unclear.
2. We grew five grassland species common to the semi-arid grasslands of northern Colorado, USA, as monocultures and as mixtures of all five species in pots. We examined the effects of atmospheric CO2 concentration (ambient vs. 780 p.p.m.) and soil moisture (15 vs. 20% m/m) on plant biomass and plant N uptake. Our objective was to separate CO2 effects not related to water from water-mediated CO2 effects by frequently watering the pots, thereby eliminating most of the elevated CO2 effects on soil moisture, and including a water treatment similar in magnitude to the water-savings effect of CO2.
3. Biomass of the C3 grasses Hesperostipa comata and Pascopyrum smithii increased under elevated CO2, biomass of the C4 grass Bouteloua gracilis increased with increased soil moisture, while biomass of the forbs Artemisia frigida and Linaria dalmatica had no or mixed responses. Increased plant N uptake contributed to the increase in plant biomass with increased soil moisture while the increase in plant biomass with CO2 enrichment was mostly a result of increased N use efficiency (NUE). Species-specific responses to elevated CO2 and increased soil moisture differed between monocultures and mixtures. Both under elevated CO2 and with increased soil moisture, certain species gained N in mixtures at the expense of species that lost N, but elevated CO2 led to a different set of winners and losers than did increased water.
4. Elevated CO2 can directly increase plant productivity of semi-arid grasslands through increased NUE, while a CO2-induced increase in soil moisture stimulating net N mineralization could further enhance plant productivity through increased N uptake. Our results further indicate that the largest positive and negative effects of elevated CO2 and increased soil moisture on plant productivity occur with interspecific competition. Responses of this grassland community to elevated CO2 and water may be both contingent upon and accentuated by competition.
- Top of page
- Materials and methods
Both empirical and modelling studies indicate that semi-arid grasslands show some of the largest increases in plant productivity in response to atmospheric CO2 enrichment (Melillo et al. 1993; Morgan et al. 2004b). Not all plant species respond in the same way to elevated CO2. For instance, the C3 grass Hesperostipa comata and the sub-shrub Artemisia frigida showed strong increases in above-ground biomass with elevated CO2 in a 5-year open-top-chamber experiment at the shortgrass steppe in Colorado, while above-ground biomass of the C3 grass Pascopyrum smithii and the C4 grass Bouteloua gracilis was not affected (Morgan et al. 2004a, 2007). Differences in photosynthetic pathways between C3 and C4 plants or the ability to fix N are important species traits that affect how plant species respond to elevated CO2 (Johnson, Polley & Mayeux 1993; Lüscher & Nösberger 1997; Reich et al. 2001b). Because soil resources such as water and nitrogen (N) are affected by elevated CO2, the ability to compete for these resources is another factor that could cause variation in plant species growth responses to elevated CO2 (Berntson, Rajakaruna & Bazzaz 1998; Derner et al. 2003; Maestre, Bradford & Reynolds 2005). Soil moisture in particular is an important resource in semi-arid grasslands that could be critical for species-specific responses to elevated CO2. Indeed, it was suggested that the increased growth of certain semi-arid grassland species under elevated CO2 was a result of improved soil moisture conditions (because of decreased stomatal conductance) more than direct effects of elevated CO2 on photosynthesis (Lecain et al. 2003; Morgan et al. 2007).
Increased plant growth under elevated CO2 coincides with increased plant N uptake as well as increased N use efficiency (NUE, Soussana et al. 2005; Norby & Iversen 2006; Finzi et al. 2007). The extent to which increased plant growth under elevated CO2 involves changes in NUE or plant N uptake depends on how much N is available in the soil for plant growth, which itself is influenced by CO2. Elevated CO2 could reduce soil N availability because of increased microbial immobilization (Díaz et al. 1993; Gill et al. 2002). Initial increases in plant N uptake could reduce soil N availability in the long-term because of increased storage of N in long-lived plant biomass and soil organic matter (Luo et al. 2004; Reich, Hungate & Luo 2006). In systems where soil N availability is reduced by elevated CO2, increases in plant growth under elevated CO2 may therefore only be possible when plants increase their NUE. On the other hand, in dry ecosystems elevated CO2 can significantly improve soil moisture conditions, thereby increasing N mineralization and plant N uptake (Hungate et al. 1997; Dijkstra et al. 2008). Therefore, increased plant N uptake under elevated CO2 may be more important for increased plant growth in dry than in wet ecosystems.
Here we studied the effects of atmospheric CO2 (ambient vs. 780 p.p.m.) and soil moisture (15 vs. 20% m/m) on plant growth and plant N uptake of five species common to the semi-arid grasslands in northern Colorado, in an environmentally controlled greenhouse experiment. We tried to keep soil moisture levels constant throughout the experiment to separate soil moisture effects from direct effects of elevated CO2 not related to soil moisture, such as effects on photosynthesis and rhizosphere processes affecting nutrient cycling (Dijkstra & Cheng 2008). Many studies using greenhouse and growth chambers to test elevated CO2 effects on plant growth and plant N uptake have been done by growing plants as monocultures or in isolation as single plants (e.g., Morgan et al. 1994, 1998; Dijkstra & Cheng 2008). However, plant growth responses to elevated CO2 grown in isolation or as monocultures may be very different from plant growth responses when grown in mixtures (Navas 1998; Poorter & Navas 2003). Large variation in species-specific plant growth responses to elevated CO2 could change competitive interactions within plant communities (Bazzaz & McConnaughay 1992; Körner & Bazzaz 1996). Indeed, above-ground plant biomass of the C3 grass P. smithii and the C4 grass B. gracilis significantly increased under elevated CO2 when grown as monocultures in growth chambers (Morgan et al. 1994, 1998; Hunt et al. 1996), whereas the same species showed no or very little response to elevated CO2 when growing in a natural plant community (Morgan et al. 2004a). To elucidate the role of inter- and intraspecific competition for resources, we compared CO2 and soil moisture treatment effects on plant species grown as monocultures with their effects on the same plant species grown in mixtures.
We asked the following questions.
1. Is the stimulatory effect of elevated CO2 on the growth of five semi-arid grassland species caused by improved water conditions, or also by other CO2 effects?
2. What are the roles of increased plant N uptake and increased NUE in the stimulatory effects of elevated CO2 and increased soil moisture?
3. Do plant growth and N uptake responses to elevated CO2 and increased water availability differ between inter- and intraspecific competitive interactions among plants?
Materials and methods
- Top of page
- Materials and methods
The soil we used for our experiment came from a semi-arid grassland at the USDA-ARS Central Plains Experimental Range (CPER), northeastern Colorado (lat. 40°50′, long. 104°47′). The soil is a sandy loam of the Ascalon series (Aridic Argiustolls). The top 20-cm of the soil was scraped from the surface with a backhoe and dumped on a large metal sieve (mesh size 4 mm) to remove large plant parts and to homogenize the soil. The soil had 0·95% C and 0·09% N, and a pH of 6·6. We filled 120 polyvinyl chloride (PVC) pots (diam. 20 cm, height 40 cm) with sieved soil (c. 14 kg of air-dry soil per pot). The pots were capped at the bottom and no leaching occurred during the experiment. The initial inorganic N content (NH4+ + NO3−) of the soil was 23 mg N kg−1 soil or ∼0·3 g N pot−1. The pots were then watered to field capacity or 30% m/m. We transplanted seedlings of the perennial grasses Bouteloua gracilis (BOGR, C4 grass), Hesperostipa comata (HECO, C3 grass), and Pascopyrum smithii (PASM, C3 grass), the sub-shrub Artemisia frigida (ARFR), and the invasive forb Linaria dalmatica (LIDA) as monocultures (five seedlings per pot, 20 pots per species). In the other 20 pots we transplanted all five species as mixtures (one seedling of each species per pot).
We grew the plants in two greenhouses located at the USDA-ARS Crops Research Laboratory, Fort Collins, CO, USA. Half of all the pots (10 replicates of each monoculture/mixture or species composition) were placed in one greenhouse that was kept under ambient atmospheric CO2 (400 ± 40 p.p.m, average ± standard deviation), and the other half in a greenhouse kept under elevated CO2 (780 ± 50 p.p.m.). The CO2 concentration was continuously monitored and the CO2 supply was computer-controlled (Argus Control Systems Ltd, White Rock, BC1). The added CO2 entered the greenhouse through a ventilation system ensuring uniform distribution of the CO2 concentration inside the greenhouse. Air temperature in both greenhouses was kept between 27 and 29 °C during the day and between 16 and 18 °C during the night. Temperature was regulated by computer-controlled air conditioners and heaters (York International, York, PA). Both greenhouses were equipped with 600 W lights (P.L. Light Systems, Beamsville, ON) that were on during the day for 12 h. During the day the light intensity in each greenhouse was ∼200 W m−2. The relative humidity in each greenhouse was 24 ± 5% during the day and 30 ± 5% during the night. To reduce greenhouse effects not related to the CO2 treatment, we swapped the pots once a week between the two greenhouses during our experiment (12 weeks, Heijmans et al. 2002; Goverde & Erhardt 2003). The CO2 treatment was swapped concurrently so that the same plants received the same CO2 treatment throughout the experiment.
During the first week of the experiment the pots were watered frequently to maintain soil water content near 30% to enhance seedling growth. After that, watering was discontinued until half of the pots (five replicates for each monoculture/mixture and each CO2 treatment) dried down to 15% gravimetric soil moisture (low water) and the other half to 20% gravimetric soil moisture (high water). The 15 and 20% soil moisture contents correspond to 50 and 67% of field capacity respectively. The relative difference between the two water treatments is 33%. At CPER, elevated CO2 (720 p.p.m.) increased soil moisture on average from 11·4 to 12·9% (increase of 14% compared with ambient CO2) in the upper metre, while the relative difference between ambient and elevated CO2 was sometimes as much as 45% (Lecain et al. 2003). Thus, the magnitude of our water treatment was not unrealistic compared with the water savings effect of elevated CO2 under field conditions. We maintained the low and high soil water levels by watering the pots three times per week with DI water. Once a week, the pots were weighed and watered to their target soil moisture levels, while during the other two times of the week, the amount of water that was added was estimated based on previous water loss from each pot. Pots inside each greenhouse were placed in five blocks of twelve pots (one replicate of each of the six species composition and two water treatments).
With our frequent watering we tried to maintain constant soil moisture levels during the experiment, thereby eliminating potential CO2 effects on soil water content. However, between watering periods, pots under ambient CO2 dried out faster than pots under elevated CO2 (Fig. 1). On average, soil moisture of the low water treatment was 12·6% and 13·1% under ambient and elevated CO2 respectively, and soil moisture of the high water treatment was 16·9% and 17·7% under ambient and elevated CO2 respectively (averaged for 25–85 days after transplanting).
Figure 1. Average gravimetric soil moisture content during the experiment for each of the CO2 and water treatments (averaged across species identity and species number, aCO2 = ambient CO2, eCO2 = elevated CO2).
Download figure to PowerPoint
We harvested all pots 85 days after transplanting. Plants were separated into shoots and roots, dried (65 °C) and weighed. The plant material was then ground and analysed for N on a mass spectrometer (20–20 Stable Isotope Analyzer, Europa Scientific, Cheshire, UK). We were unable to separate root biomass in the mixtures by species and the data reported are for the combined roots from all species. The soil in each pot was thoroughly mixed and a 25 g subsample was extracted with 60 ml 2 M KCl, filtered (using pre-cleaned Whatman No. 1 filter paper) and frozen until analyses for NH4+ and NO3− on a flow injection analyzer (QuickChem FIA+, Lachat Instruments, Milwaukee, WI). We assumed that the difference between the final soil inorganic N amount (NH4+ and NO3−) and the initial amount at the start of the experiment was taken up by the plant. Note that this is a potential amount, and that the amount of initially available N that was actually taken up was somewhat lower because some of the initial inorganic N was lost as gaseous N during the experiment (Dijkstra et al. 2010). We then compared this amount to the total amount of N in plant biomass to deduce plant N supply through decomposition during the experiment.
We calculated the absolute change in shoot biomass in response to elevated CO2 and high water for each species grown in monoculture and in mixture. Because there was only one plant for each species in the mixtures, but five plants in each of the monoculture pots, we multiplied the absolute responses in the mixtures by five for plant density-independent comparison with the monocultures. We also calculated the Shoot Biomass Enhancement Ratio (BER) and shoot N uptake Enhancement Ratio (NER, Berntson, Rajakaruna & Bazzaz 1998; Poorter & Navas 2003) to elevated CO2 and high water for each species in monoculture and mixture. BERCO2 was calculated as the ratio of the average shoot biomass of the elevated CO2 treatment divided by the average shoot biomass of the ambient CO2 treatment, while BERwater was calculated as the ratio of the average shoot biomass of the high water treatment divided by the average shoot biomass of the low water treatment. NERCO2 and NERwater were calculated similarly, but using shoot N content (in g pot−1) rather than shoot biomass. BER and NER values greater than one indicate positive effects of elevated CO2 or high water on shoot biomass and shoot N content (increased N uptake). Further, if BER and NER are the same, then the positive effect of elevated CO2 or high water on shoot biomass is accompanied by increased N uptake alone, but not by increased N Use Efficiency (NUE, shoot biomass/shoot N content). If BER is higher than NER, then the positive effect of elevated CO2 or high water on shoot biomass involves increased NUE. We further defined the N Use Efficiency Enhancement Ratio (NUE-ER) as the ratio of the NUE of the elevated CO2 or high water treatments divided by the NUE of the ambient CO2 or low water treatments (NUE-ERCO2 and NUE-ERwater respectively).
For the monocultures we used anova to test for main effects of CO2 (ambient and elevated CO2), water (low and high water), and species (ARFR, LIDA, BOGR, HECO, and PASM), as well as their interactions, on shoot, root, and total biomass and their N contents. For each species we used the Tukey’s HSD test to compare the means of the four CO2 by water treatment combinations. We did the same analyses with the mixtures, but then only for shoot biomass and N content (we were unable to separate root biomass by species in the mixtures). For root and total biomass and their N contents in the mixtures we left the factor species out of the anova, and only tested for CO2, water, and CO2 × water effects. Using all pots, we tested for main effects of CO2, water, and species number (monocultures and mixtures), and their interactions on shoot, root, and total biomass and their N contents. For these last analyses we first averaged the five monoculture species in each block to create equal sample sizes compared with the mixtures. We also used Tukey’s HSD tests to compare the means of the four CO2 by water treatment combinations for monocultures and mixtures separately. Finally we used anova to test for main effects of CO2, water, species, species number, and all their interactions, on shoot biomass, shoot N content, and shoot NUE. In all anova s we included block as a random effect. We log-transformed data when necessary to reduce heteroscedasticity. All statistical analyses were done with JMP (version 4.0.4; SAS Institute, Cary, North Carolina, USA).
- Top of page
- Materials and methods
Total plant biomass in the monocultures increased under elevated CO2 (by 6·8% averaged across the water treatment) and with high water (by 10·4% averaged across the CO2 treatment, Table 1, Fig. 2a and b). In the mixtures, elevated CO2 and high water effects on total plant biomass were similar in magnitude (average increase of 9·8% under elevated CO2 and 6·1% with high water), but less significant for the CO2 and not significant for the water treatment (Table 1). Individual species in monoculture showed different responses to elevated CO2 and high water. Total biomass of the C3 grasses HECO and PASM increased with elevated CO2, although only significantly so in combination with high water (Fig. 2a). Total biomass of the C4 grass BOGR was not affected by elevated CO2, but increased with high water. The sub-shrub ARFR did not respond to elevated CO2 or water, while the invasive forb LIDA responded positively to elevated CO2 with low water but negatively with high water.
Table 1. Summary of anova results (P-values) for the effects of CO2 (ambient and elevated), water (low and high), and species identity (ARFR, LIDA, BOGR, HECO, and PASM) in the monocultures only and in the mixtures only, and for the effects of CO2, water, and species number (monoculture and mixture) in all pots (ns = not significant, P > 0·1)
|Effect||Shoot biomass||Root biomass||Total biomass||Shoot N||Root N||Total N|
|CO2 × water||ns||0·02||ns||ns||ns||ns|
|CO2 × sp||ns||ns||ns||ns||ns||ns|
|Water × sp||0·02||0·01||0·001||0·005||ns||0·08|
|CO2 × water × sp||ns||<0·0001||0·002||ns||0·03||ns|
|CO2 × water||0·05||ns||ns||0·10||ns||0·06|
|CO2 × sp||0·06||–||–||0·10||–||–|
|Water × sp||0·09||–||–||ns||–||–|
|CO2 × water × sp||ns||–||–||ns||–||–|
|CO2 × water||0·05||ns||ns||0·01||ns||0·03|
|CO2 × sp#||ns||0·04||ns||ns||ns||ns|
|Water × sp#||ns||ns||ns||ns||ns||ns|
|CO2 × water × sp#||ns||ns||ns||ns||ns||ns|
Figure 2. Average total biomass (a and b) and total plant N content (c and d) for each of the CO2, water, and species identity treatments within the monocultures (a and c), and for each of the CO2, water and species number treatments for all pots (b and d, aCO2 = ambient CO2, eCO2 = elevated CO2, L = low water, H = high water, ARFR = A. frigida, LIDA = L. dalmatica, BOGR = B. gracilis, HECO = H. comata, PASM = P. smithii, MONO = monocultures, MIX = mixtures, error bars indicate 1 SE). Panel C and D also include the change in soil inorganic N between the end and start of the experiment. Different letters above bars indicate significant differences among the CO2 and water treatments for each species or species number separately (P < 0·05, Tukey’s HSD test).
Download figure to PowerPoint
In contrast to total plant biomass, total plant N (in g pot−1) was not affected by elevated CO2 in the monocultures (Table 1, Fig. 2c). On the other hand, total plant N, averaged across all species and CO2 levels, increased significantly with high water (on average by 10·4%, Table 1). Although similar in magnitude (average increase of 9·3%), the high water treatment effect on total plant N was not significant in the mixtures (Table 1, Fig. 2d). Within the monocultures, total species-specific plant N responses to elevated CO2 and high water were similar to species-specific plant biomass responses, with the largest increases in total plant N for HECO, only an increase with high water for BOGR, and no elevated CO2 or high water effects for ARFR and LIDA. Unlike total plant biomass, total plant N of PASM did not respond to elevated CO2 or high water (Fig. 2c). Soil inorganic N was depleted from 0·32 g N pot−1 at the beginning of the experiment to very low concentrations in all treatments at the end of the experiment (on average to 0·017 g N pot−1, Fig. 2c and d). As a result, changes in soil inorganic N during the time frame of the experiment were very similar among treatments. Thus, treatment effects on total plant N were not due to differences in plant uptake of soil inorganic N that was available at the start of the experiment, but most likely because of differences in N supply (i.e., net N mineralization, and possibly organic N uptake).
In the monocultures, effects of elevated CO2 and high water were slightly larger for shoot biomass than for total biomass (average increase of 14·1% under elevated CO2 and 14·0% with high water). Also, the increase in shoot N in response to high water was slightly larger than for total plant N (average increase of 14·3%). While larger in magnitude, individual species shoot biomass and shoot N responses to elevated CO2 and high water in monoculture showed a similar pattern as individual species total biomass and total plant N responses, with the exception that LIDA shoot biomass did not respond to elevated CO2 or high water and that BOGR shoot biomass increased under elevated CO2 with high water (Fig. 3a). When all five species were grown in mixtures, elevated CO2 and high water had no effect on shoot biomass, while shoot N significantly decreased under elevated CO2 (on average by 16·8%) and increased with high water (on average by 8·7%, Table 1, Fig. 3b and d). Also, responses of the individual species to elevated CO2 and high water changed compared with their responses in monoculture (Fig. 3b and d). For instance, when grown in mixtures, shoot biomass and shoot N of LIDA was negatively affected by elevated CO2, particularly with low water, while shoot biomass of BOGR was negatively affected by high water under ambient CO2. The CO2 × sp × sp# and Water × sp × sp# effects on shoot biomass and shoot N were marginally significant (Table 2). Absolute differences in shoot biomass responses to elevated CO2 and high water (net change in shoot biomass) for monocultures and mixtures are shown in Fig. 4. To compare net changes in shoot biomass between monocultures and mixtures we multiplied the net changes in the mixtures by five (see Methods). Here it becomes particularly clear that species responses to elevated CO2 and high water depended on whether these species were grown in monoculture or in mixture. In particular, LIDA responses to elevated CO2 were positive in monoculture but negative in mixtures, and BOGR responses to high water were positive in monoculture but negative in mixtures (particularly under ambient CO2).
Figure 3. Average shoot biomass (a and b) and shoot N content (c and d) for each of the CO2, water, and species identity treatments in the monocultures (a and c) and in the mixtures (b and d, for explanation of abbreviations see Fig. 2, error bars indicate 1 SE). Different letters above bars indicate significant differences among the CO2 and water treatments for each species separately (P < 0·05, Tukey’s HSD test).
Download figure to PowerPoint
Table 2. Summary of anova results (P-values) for the effects of CO2 (ambient and elevated), water (low and high), species (ARFR, LIDA, BOGR, HECO, and PASM), species number (monocultures and mixtures), and their interactions on shoot biomass, shoot N pool, and shoot NUE (ns = not significant, P > 0·1)
|Effect||Shoot biomass||Shoot N||Shoot NUE|
|CO2 × water||ns||0·01||0·06|
|CO2 × sp||0·04||0·09||<0·0001|
|CO2 × sp#||ns||ns||ns|
|Water × sp||ns||ns||ns|
|Water × sp#||ns||ns||ns|
|Sp × sp#||<0·0001||<0·0001||<0·0001|
|CO2 × water × sp||ns||ns||ns|
|CO2 × water × sp#||ns||ns||ns|
|CO2 × sp × sp#||0·07||0·10||ns|
|Water × sp × sp#||0·04||0·08||ns|
|CO2 × water × sp × sp#||ns||ns||ns|
Figure 4. Net change in shoot biomass in response to elevated CO2 (a) and increased soil moisture (b) for each species and water treatment (a) or each species and CO2 treatment (b) grown in monoculture and in mixture (for explanation of abbreviations see Fig. 2, and for CO2, water, species, and species number treatment effects on shoot biomass see Table 2).
Download figure to PowerPoint
The N uptake Enhancement Ratio (NER) was plotted as a function of the Biomass Enhancement Ratio (BER) to investigate the association of shoot biomass responses to elevated CO2 and high water with N uptake (expressed by NER), and to evaluate changes in NUE (expressed by the deviation from the 1 : 1 line in Fig. 5). BERCO2 values (BER in response to elevated CO2, Fig. 5a) were generally greater than 1, except for some species grown in mixtures. On the other hand, NERCO2 values (NER in response to elevated CO2) were mostly smaller than 1, except for some species under high water. Further, NERCO2 values were always lower than BERCO2 values, indicating that the NUE increased for all treatments under elevated CO2 (P < 0·0001, Table 2). The NUE Enhancement Ratios in response to elevated CO2 (NUE-ERCO2) ranged between 1·02 and 1·42. Most of the NERwater values were greater than 1, and NERwater values were sometimes lower and sometimes higher than BERwater (Fig. 5b). The NUE-ERwater ranged between 0·78 and 1·16 and on average, the increase in NUE with high water was only marginally significant (P = 0·09, Table 2). Most of the treatments that had a NUE-ERwater smaller than 1 were under elevated CO2. There was also a marginally significant CO2 × water interaction for NUE (P = 0·06, Table 2).
Figure 5. The shoot N uptake enhancement ratio as a function of shoot biomass enhancement ratio in response to elevated CO2 (a, NERCO2 and BERCO2 respectively) and in response to increased soil moisture (b, NERwater and BERWater respectively) for each species and water treatment (a) and for each species and CO2 treatment (b) grown in monoculture and mixture (for explanation of abbreviations see Fig. 2). The black solid 1:1 line indicates a NUE enhancement ratio (NUE-ER) of 1. Data points above this line have a NUE-ER <1 and below this line have a NUE-ER >1. The grey dashed lines indicate different values of NUE-ER. Note that BER and NER are presented on a natural log scale.
Download figure to PowerPoint