| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Research Institute for Agrobiology and Soil Fertility (AB), Dep. of Soil Chemistry and Soil Ecology, P.O. Box 14, 6700 AA Wageningen, The Netherlands
a.gorissen{at}ab.wag-ur.nl
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and methodsResultsDiscussionREFERENCES |
|---|
278 kg N ha-1, both labeled with 15N. The effects of elevated CO2 on 15N and N uptake and dynamics in the plant–soil systems were determined after 32 and 55 d, with close attention to the rhizosphere. Total N uptake by the plants was not affected by elevated CO2, compared with ambient CO2, independent of N treatment and grass species. A clear decrease from 1.77 at ambient CO2 to 1.25 at elevated CO2 was observed in the shoot/root (S/R) ratio of N, resulting from a significant decrease of the N concentration in shoots, and an unchanged root N concentration. At 700 µL L-1 CO2, N concentration in the shoots decreased from 12.9 to 9.9 g kg-1, even at the low N supply, whereas the slight decrease in root N concentration for plants grown at elevated CO2 (7.9 vs. 7.3 g kg-1) was not significantly different. The relative increase of 15N found in the rhizosphere soil microbial biomass (SMB) and the rhizosphere soil residue under elevated CO2 was too small to affect plant growth, even in the low N treatment. The total amount of 15N recovered in the plants was not affected by the CO2 treatment. Although at the second harvest slightly more 15N was found in the plants than at the first harvest, probably due to turnover of the SMB, no interaction with CO2 was observed. This shows that the fertilizer 15N had not been immobilized to a larger extent or for a longer time by the SMB at elevated CO2 than under ambient CO2, even independent of N level and grass species. No evidence was found that under elevated CO2 substantial amounts of N had been immobilized by the SMB, nor that mineralization of native soil organic matter (SOM) had been stimulated by an increased supply of substrate to the SMB. We conclude that elevated CO2 has the potential to induce significant changes in plant N nutrition, modifying N allocation and tissue quality within perennial grasses, but that these effects appear to be independent of the SMB.
Abbreviations: HN, high N treatment • LN, low N treatment • SMB, soil microbial biomass • SOM, soil organic matter • S/R, shoot/root [ratio of N]
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionREFERENCES |
|---|
Not only is the response of the SMB important in nutrient dynamics, but nutrient cycling may also be strongly influenced by the species composition of the vegetation (Berendse et al., 1989). Wedin and Tilman (1990) showed that mineralization in initially identical soils was strongly diverging when planted with different grass species, probably due to differences in quality and quantity of belowground litter. The production of belowground litter and its quality are likely to be affected by an increase in atmospheric CO2, with subsequent changes in decomposition rates (Cotrufo and Ineson, 1995; Gorissen et al., 1995; Van Ginkel et al., 1996). A shift in species composition, favoring the superior N competitors in a vegetation under elevated CO2 (Wedin and Tilman, 1990) may induce further changes in the litter decomposition rates of a community (O'Neill, 1994).
In this study, we tested whether the effect of elevated CO2 on plant and microbial N uptake depends on soil N supply, using three grass species that varied in growth rate: L. perenne (high), A. capillaris (medium), and F. ovina (low). We applied two N levels, a very low one to induce serious depletion symptoms and a higher one that was expected to be sufficient for the time span of the experiment. The N fertilizer was labeled with 15N to follow the dynamics, within the plant–soil systems, of fertilizer N that was applied at the start of the experiment. The effects of CO2, N level, and grass species on dry weights and C dynamics in the plant–soil system have been discussed by Cotrufo and Gorissen (1997). Here, the effects of elevated CO2 on N dynamics as affected by N availability and grass species are presented, with emphasis on the rhizosphere.
|
Materials and methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionREFERENCES |
|---|
60% of field capacity (14%, w/w) by addition of deionized water. Inert gravel was spread onto the soil surface to prevent excessive water loss and algal growth. Thirty-two pots were used for each grass species. All pots received 28 mg P and 50 mg K kg-1 dry soil as KH2PO4 and K2SO4 solution. Half of the pots were amended with 29.6 mg N per pot (corresponding with about 135 kg N ha-1) added as a solution of NH4NO3. These fertilizers were given twice during the experiment, on Day 1 and 32, with a total N supply corresponding with 270 kg N ha-1. In addition, all 32 pots received one application of 1.8 mg 15N-NO3 (0.967atom %) at the start of the experiment (corresponding with 8 kg N ha-1). This resulted in a low N (LN) treatment (8 kg N ha-1) and a high N (HN) treatment (278 kg N ha-1).
On Day 1, pots were transferred to the Experimental Soil Plant Atmosphere System (ESPAS) phytotrons (Gorissen et al., 1996) where half of them, for each plant species and N treatment, were exposed to 350 µL L-1 CO2 and the other half to 700 µL L-1 CO2 in a continuously 14C-labeled atmosphere (specific activity 2.1 kBq mg-1 C). Temperatures in the growth chambers (18/14°C day/night) were controlled by a platinum resistance thermometer Pt100, relative humidity (65/70% day/night) using a capacitive humidity sensor, and irradiation (300 µmol m-2 s-1 with a day/night rhythm of 16/8 h) by means of a photosynthetically active radiation meter. All environmental variables were checked with a third independent meter to assure identical conditions in both chambers.
Gravimetric soil water content was kept at an average of 14% throughout the experiment by periodically adding deionized water to a predetermined weight. Although elevated CO2 may alter soil water content in grasses and subsequently mineralization processes (Rice et al., 1994), we have chosen to equalize soil water content in the different treatments in order to isolate one of the key mechanisms.
Harvest
Two harvests took place during the experiment, on Day 32 and 55, with four replicated pots being harvested each time. Plants were removed from pots with bulk and rhizosphere soil collected separately and sieved to pass 5- and 2-mm mesh, respectively. Soil closely adhering to the roots was considered as "rhizosphere soil" and the remaining soil as "bulk soil". The roots were carefully washed in a known amount of deionized water, which was successively centrifuged. After centrifugation, the deposited soil was collected and mixed with the previously collected rhizosphere soil. During harvest, a soil subsample was taken from both bulk and rhizosphere soils from each pot and was oven dried (70°C) for the measurement of water content, 15N content, and total N. After washing, plants were separated into shoots and roots before oven drying (70°C). The remaining fresh soils were stored in plastic bags at 2°C until further analysis.
Measurements
For the measurement of total N and 15N in the plants, 4 mg of finely ground shoot and root material was weighed into tin cups. The total N content and the 15N isotope ratio (atom % 15N) in shoots, roots, rhizosphere soil solutions, rhizosphere soils, and bulk soils were determined using an ANCA-MS (Europa Scientific Ltd., Crewe, UK) after combustion of the sample at 1000°C and passage of the gases through a reduction column at 600°C (Roboprep-CN, Europa Scientific Ltd.) in an isotope ratio mass spectrometer (Tracermass, Europa Scientific Ltd.).
The SMB N flush (SMB N), and 15N flush (SMB 15N) were measured in rhizosphere soils using the fumigation–centrifugation method (Van Ginkel et al., 1994). After centrifugation, total N and 15N content in rhizosphere soil solution of unfumigated and fumigated soil samples was measured using a modified mass spectrometer method (Bruulsema and Duxbury, 1996). Two milliliters of soil solution was evaporated from tin cups at 35°C to avoid possible loss of volatile N compounds before subsequent analysis with the ANCA-MS.
Calculations
The percentage of N recovered in the different plant–soil compartments that originated from the N fertilization applied at the start of the experiment was calculated by: [(atom % 15N sample - atom % 15N background)/(atom % 15N fertilizer - atom % 15N background)]100 (Edwards, 1973).
Statistics
The experiment consisted of two CO2 levels, two N levels, three species, and two harvests. Each treatment combination had four replicates. Within each growth chamber all treatment combinations were completely randomized between pots. A total of 96 plant–soil systems was included in the experiment (2 by 2 by 3 by 2 by 4) of which 16 were omitted due to insufficient plant material after grinding (F. ovina, first harvest, both CO2 levels, both N treatments). Analysis of variance was carried out using regression techniques to adjust for these missing values (Genstat 5; version 3.1; IACR, Rothamsted, UK). Differences were considered significant when P values were lower than 0.05.
|
Results |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionREFERENCES |
|---|
|
|
This shift in N distribution within the plants, coupled with the increase in plant yield (Cotrufo and Gorissen, 1997) significantly affected plant "quality" factors such as N concentration and C/N ratio (Table 1; Fig. 1). These effects were more pronounced in the shoots than in the roots, since dry weight of the shoots hardly increased at elevated CO2, whereas the dry weight of the root system was much higher than at ambient CO2 (Cotrufo and Gorissen, 1997). At 700 µL L-1 CO2, N concentration in the shoots decreased from 12.9 to 9.9 g kg-1, even at LN, whereas the slight decrease in root N concentration for plants grown at elevated CO2 (0.79 vs. 0.73%) was not significantly different (Table 1). The C/N ratio in the shoots substantially increased at elevated CO2, but depended on the N supply. At LN, the C/N ratio increased from 39.6 to 53.0 (+34%) and at HN from 31.0 to 37.5 (+21%). This difference was more pronounced at the second harvest than at the first harvest. A slight but not significant increase from 38.8 to 42.0 was observed for the C/N ratio of the roots at elevated CO2, although interactions were found with N, species, and harvest (Table 1).
Nitrogen
The N treatment showed a species-related effect on total N uptake (Table 1). At the end of the experiment, total N uptake at LN was 36.0% higher in L. perenne than in F. ovina (10.01 vs. 7.36 mg), but at HN this difference was 25.8% (59.87 vs. 47.60 mg). The interaction between N and harvest was to be expected, since only HN plants received extra fertilizer N after the first harvest. The same trends in total N contents were observed for the shoots and roots separately. The S/R ratio at LN was 30% lower than at HN. At LN, the N percentage was 24 and 18% lower in the shoots and roots, respectively, than at HN. Logically, this decrease was accompanied by an increase in the C/N ratio at LN from 34.3 to 46.3 (+35%) in the shoots and from 35.1 to 46.0 (+31%) in the roots compared with HN.
Species
The mean total N uptake at the end of the experiment showed the following order: L. perenne, 34.94 mg; A. capillaris, 32.53 mg; and F. ovina 27.48 mg, although depending on the N level, as mentioned above. Lolium perenne had the lowest mean S/R ratio of N, accompanied by the lowest N concentration in the shoots and the roots and the highest C/N ratios in shoots and the roots.
Nitrogen in the Rhizosphere Soil
Carbon Dioxide
No significant changes were found in the mineral N content in the soil solution at elevated CO2, as compared with ambient CO2 (Table 2)
. Within the rhizosphere soil environment, elevated CO2 did not induce statistically significant changes in the N concentration in any of the three soil compartments, and similar values of N, expressed per gram of rhizosphere soil, were measured for the rhizosphere soil solution, SMB, and soil residue (Table 2, Fig. 1). However, since at elevated CO2 the amount of rhizosphere soil recovered from each plant–soil system was significantly greater (Table 2) due to the larger root system (+24%; Cotrufo and Gorissen, 1997), significantly more N, in absolute terms, was measured at elevated CO2 in the rhizosphere soil solution (0.37 vs. 0.28 mg N), in the SMB (5.96 vs. 4.56 mg N), and in the soil residue (203 vs. 168 mg N).
|
20% more mineral N being recovered per gram of soil in the soil solution of the fertilized systems (Table 2). The SMB did not respond significantly to N fertilization, with similar values of the N flush being measured per gram of rhizosphere soil at both levels of N fertilization (Table 2). The N concentration of soil residue was not affected by the N treatment (Table 2). However, since at high N more rhizosphere soil was recovered from each plant–soil system (Table 2; see Cotrufo and Gorissen, 1997), significantly more N, in absolute terms, was measured in the rhizosphere soil residue (226.9 vs. 144.4 mg N) and in the SMB (6.36 vs. 4.15 mg N) of the fertilized systems.
Species
The N concentration in the soil solution and in the SMB was not affected by plant species (Table 2). However, species significantly changed the concentration of N in the rhizosphere soil residue; values were 0.91 mg in L. perenne, 0.98 mg in A. capillaris, and 0.99 mg in F. ovina at the end of the experiment.
Nitrogen-15 in Plants and Soil
The distribution of 15N among different plant–soil compartments showed that elevated CO2 had decreased allocation of fertilizer-derived N to the shoots by 11% (48.9 vs. 43.5%), whereas the allocation to the roots had increased by 15% (27.7 vs. 31.9%) without interactions with the other treatments (Table 3)
. The percentages that were recovered in the rhizosphere compartments were very low. Elevated CO2 increased the percentage of 15N in the rhizosphere soil solution by 35% (0.114 vs. 0.155), but this pool only contained 0.13% of the amount of fertilizer N given at the start. The percentage of 15N in the SMB had increased by 34% as a mean, and in the rhizosphere soil residue by 57%. The percentage 15N in the bulk soil was not clearly affected by the high CO2 treatment. In some cases, interactions between CO2 and N, harvest, or species were observed, but without any clear tendency (Table 3).
|
|
Discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionREFERENCES |
|---|
Hungate et al. (1996) found no increase in N content in the SMB under nutrient-rich conditions. However, under N limitation, interactive effects were observed between CO2 and species, with a decrease in microbial N in a soil planted with L. perenne or fescue [Vulpia microstachys (Nutt.) Monro], but an increase in soils planted with other species. According to calculations by Schenck et al. (1995), the amount of N immobilized in the SMB would be negligible compared with N yield in the total plant biomass. This may indeed make sense when excess mineral N is available. However, it should be noted that they did not take into account N immobilization in microbial products, which might have been substantial. In our study, the amounts of 15N-fertilizer recovered in the SMB and the residue, which should contain the microbial products, were almost equal and in the LN treatment certainly not negligible compared with the N content of the plants. In the LN treatment initially only 5.7 mg of mineral N, including fertilizer N, was available. Successful competition by the SMB would therefore probably result in a reduced N uptake by the plants. The mean recovery in the plants was 9.2 mg, leaving 3.5 mg that must have been mineralized from SOM. If the SMB, when stimulated by an extra supply of substrate at elevated CO2 (Cotrufo and Gorissen, 1997), would have really successfully competed with the plants for N, it would be in this treatment where total N uptake by the plants would have been reduced. However, total net N uptake by the plants was not affected by CO2 at either of the two N levels, and the N amount in the microbial biomass flush and in the residue per gram of rhizosphere soil had not changed. Despite the fact that elevated CO2 increased the total amount of N recovered in the rhizosphere SMB because of an extension of the root system, the SMB in the LN treatment contained 0.3 mg N more at elevated CO2 than at ambient CO2 and at HN 2.5 mg N, both insubstantial amounts compared with the N content of the plants. In addition, the 15N data support the view that only a small fraction of the fertilizer might have been immobilized by the microbial biomass. If the fertilizer N given at the start of the experiment, which was labeled with 15N, had been immobilized by the SMB to a larger extent or for a longer period under elevated CO2 than under ambient CO2, then relatively more 15N should be found in the microbial biomass and the rhizosphere soil residue, which was indeed the case. However, the increases were too low to affect plant growth, even in the LN treatment. The absence of interaction between CO2 and harvest date further indicates that fertilizer 15N had not been substantially immobilized for a longer period in the SMB at elevated CO2, even independent of N level and grass species.
The reverse scenario for CO2 effects on soil N dynamics was put forward by Zak et al. (1993). An increase in total belowground biomass of bigtooth aspen (Populus grandidentata Michaux) seedlings under elevated CO2 was followed by an increased labile C pool and an increase in microbial C in the rhizosphere soil. Subsequent mineralization experiments, performed after harvesting the plants, showed that N mineralization was greater in the bulk soil of the elevated CO2 treatment, but not in the rhizosphere soil. Zak et al. (1993) suggested that N mineralization would increase under elevated CO2 as a result of stimulation of the bacterial populations near actively growing roots and recycling of nutrients through predation by protozoa. Our experiment confirmed the increased C input into soil under elevated CO2, which amounted to 24% (Cotrufo and Gorissen, 1997), with an increase of 24, 39, and 21% for root, rhizosphere soil, and bulk soil, respectively. Total SMB C increased by 15% and SMB 14C in the rhizosphere by 31%, but this was more or less proportional to the increase in root mass. If this increased C input would have stimulated N mineralization (Lekkerkerk et al., 1990; Zak et al., 1993), an increase in the total amount of N in the plants or an increase in total amount of soil mineral N would be expected. However, neither the total N uptake by plants nor the mineral N content in the soil was increased by elevated CO2. In addition, an increased turnover of the SMB following the extra supply of substrate during the experiment would have resulted in lower amounts of 15N in the SMB at the end of the experiment. This would probably have been accompanied by a higher 15N uptake by plants or an increased incorporation in the rhizosphere soil residue, especially in the LN treatment. Both should become visible in an interaction between CO2 and harvest date, but neither of these responses was observed. It appears that increased input of easily decomposable C compounds and N mineralization by the SMB are not tightly linked in this study. The results further support the observations by Van Ginkel et al. (1997) and Van Ginkel and Gorissen (1998) that decomposition of native SOM was not stimulated by elevated CO2 in grasses, at least in these relatively short-term experiments. The amount of C deposited in soil may not have been sufficient to alter soil N dynamics. After 55 d, this amount varied from 16 mg to 164 mg C per plant (Cotrufo and Gorissen, 1997), and this is only a small fraction compared with the 14500 mg of C present in native SOM. The substantially increased decomposition of organic matter as reported by Körner and Arnone (1992), possibly originates from differences in the quality of the organic matter. They used an artificial organic soil covered with leaf and bark compost, which was less stable than the native SOM in our mineral soil.
Although the possibility for extrapolation of this short-term process study to entire ecosystems is limited, our results may give an indication of what may be expected. The conclusion that decomposition of native SOM in grasslands is not stimulated under elevated CO2, if it holds, would have consequences for growth stimulation in the long term. As Hungate et al. (1996) pointed out, plants may be able to gather more nutrients by extending the roots or by using mycorrhizal hyphae, but only when these nutrients are present and available. In our experiment, even in the HN treatment, the mineral N pool in the soil was very small. Although plant growth was still stimulated even in the LN treatment, we believe that this effect must be transient, since the L. perenne leaves showed extreme yellowing. The only way to maintain a higher production at elevated CO2 in N-depleted systems seems to be by addition of external N either via fertilization, N deposition, or N2 fixation (Arp et al., 1997). Only one of these three possibilities can potentially be influenced by elevated CO2: N2 fixation, in mixed cultures with legumes by symbiotic N2 fixation (Soussana and Hartwig, 1996) or in pure grassland ecosystems by free-living N2 fixers, which are usually C-limited (Killham, 1994). Whether N2 is of some use is questionable, since the N2 fixation in pastures with clover (Trifolium ssp.) is estimated to be 5 to 30 kg ha-1 (Killham, 1994). Growth stimulation in nutrient-poor ecosystems is likely to diminish, unless an increase in temperature, which will probably accompany the increasing atmospheric CO2 concentration, provides the necessary nutrients by stimulating decomposition of native SOM.
Earlier studies have also shown that increased plant growth under elevated CO2 does not necessarily imply an increased N uptake. Although Hocking and Meyer (1991) found an increased N uptake in wheat (Triticum aestivum L.), this was not observed in cockleburr (Xanthium occidentale Bertol.) (Hocking and Meyer, 1985), corn (Zea mays L.) (Hocking and Meyer, 1991), L. perenne (Van Ginkel et al., 1997), and in a mixed culture of tropical plant species (Körner and Arnone, 1992). Hungate et al. (1996) concluded that the effect of elevated CO2 on N uptake by plants is species dependent, but probably similar in "functional groups". This study confirms the results for L. perenne and shows that slower-growing grasses such as A. capillaris and F. ovina respond in a similar way.
The shift in N distribution from leaves to roots under elevated CO2 (Hocking and Meyer, 1985, 1991; Soussana et al., 1996; Van Ginkel et al., 1997) was consistently found for all three grass species in both N treatments, probably due to an increased nutrient use efficiency (Wong, 1979; Goudriaan and de Ruiter, 1983), even under the severe N stress in the low N treatment. As a result, the N percentage in the leaves was substantially reduced and the C/N ratio increased. Changes in N percentage and C/N ratio were much smaller in the roots because of the higher N allocation to the roots combined with a stronger increase in root production under elevated CO2 than in the leaves. From a classical point of view, this would imply that decomposition rates of leaves would be more affected by elevated CO2 than those of roots. However, Magid et al. (1996) found no differences in decomposition rate of L. perenne leaves despite differences in C/N ratio, whereas Van Ginkel et al. (1996) observed differences in decomposition rate of grass roots that had similar C/N ratios. The decreased decomposition rates of grass roots (Gorissen et al., 1995; Jongen et al., 1995; Van Ginkel et al., 1996) will reduce the active mineral N pool because of the longer turnover times of this plant material. This will further worsen conditions in N-limited ecosystems for a lasting growth stimulation under elevated CO2, but will probably affect species composition due to differences in the relative nutrient requirements of species (Berendse et al., 1987, 1989). According to Wedin and Tilman (1990), competition for N may be driven to a high degree through litter feedback of species; species with low N supply rates from litter are the best N competitors. Experiments with leaf and root material are now underway to obtain more information about possible differences in decomposition rates.
|
ACKNOWLEDGMENTS |
|---|
Received for publication July 24, 1998.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsResultsDiscussionREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
