SSSAJ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (11)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by McCulley, R. L.
Right arrow Articles by Burke, I. C.
Right arrow Search for Related Content
PubMed
Right arrow Articles by McCulley, R. L.
Right arrow Articles by Burke, I. C.
Agricola
Right arrow Articles by McCulley, R. L.
Right arrow Articles by Burke, I. C.
Related Collections
Right arrow Soil Microbiology
Right arrow Range Soils
Right arrow Biogeochemical Processes
Published in Soil Sci. Soc. Am. J. 68:106-115 (2004).
© 2004 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA

DIVISION S-3—SOIL BIOLOGY & BIOCHEMISTRY

Microbial Community Composition across the Great Plains

Landscape versus Regional Variability

R. L. McCulley*,a and I. C. Burkeb

a Dep. of Biology, Duke Univ., Durham, NC 27708
b Dep. of Forest Science and Graduate Degree Program in Ecology, Colorado State Univ., Ft. Collins, CO 80524

* Corresponding author (mcculley{at}duke.edu).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Rates of nutrient cycling vary across landscape and regional scales. This biogeochemical variability can be partially attributed to patterns in plant community characteristics and abiotic and edaphic conditions across topographic gradients at the landscape-scale or across regional climatic gradients. However, it is also possible that concomitant changes in the microbial communities performing these biogeochemical processes occur across the same spatial scales and may therefore contribute to the observed biogeochemical trends. To assess patterns of microbial community composition across regional and landscape scales, we sampled upland and lowland topographic positions at three grassland communities spanning a 500-mm regional precipitation gradient across the central Great Plains. Soil microbial community composition and biomass were determined using phospholipid fatty acid (PLFA) analysis. Microbial biomass increased across the regional gradient, and different microbial communities were associated with the different grassland community types. The relative abundance of fungi decreased while gram-negative anaerobic bacteria increased from shortgrass steppe to tallgrass prairie. There were no differences in microbial biomass at the landscape-scale, and the only alteration in microbial community composition between upland and lowland landscape positions was a shift toward more nonspecific bacteria in lowlands. The fact that the trends in microbial biomass and community composition at the landscape-scale were less pronounced suggests that variability in microbial community composition is larger regionally across the Great Plains than landscape variability associated with topographical features at any particular site. Alterations in the microbial community may play a role in determining the biogeochemical patterns of grasslands in the Great Plains region.

Abbreviations: ARI, the Fox ranch on the Arickaree River • FAMEs, fatty acid methyl esters • HAYS, the range area managed by Fort Hays State University • KONZA, the Konza Prairie Biological Station • PCA, principal component analysis • PLFA, phospholipid fatty acid • SGS, the Shortgrass Steppe Long-term Ecological Research site • SVR, the Smoky Valley Ranch


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
THE CENTRAL GREAT PLAINS REGION of North America is characterized by strong climatic gradients (in both mean annual precipitation and temperature) and associated trends in plant community composition and structure (Sims et al., 1978; Epstein et al., 1996; Lauenroth et al., 1999; Lane et al., 2000). Floristically distinct plant communities in this region include western semi-arid shortgrass steppe, southern mixed grass prairie, and eastern subhumid tallgrass prairie (Coupland, 1992; Kucera, 1992; Lauenroth and Milchunas, 1992). Alterations in the dominant plant species and functional type composition of these different communities affect production allocation patterns (such as root/shoot ratios) and litter quantity, and quality and timing of inputs (Sims and Singh, 1978; Sims et al., 1978; McCulley, 2002; Murphy et al., 2002). These trends in climate and vegetation interact to produce regional patterns in biogeochemistry. However, at individual sites in the Great Plains, landscape-level variability in plant community composition and production and biogeochemical parameters has also been observed. Previous work has shown that plant production, total soil C and N, potential soil C and N mineralization rates, and soil respiration generally increase from shortgrass steppe to tallgrass prairie and from uplands to lowlands at individual sites within the region (Zak et al., 1994; Briggs and Knapp, 1995; Burke et al., 1999; McCulley, 2002). These biogeochemical trends could be the result of improved abiotic conditions supporting higher microbial activity for longer periods of time during the growing season, increased substrate availability supporting larger microbial biomass pools, and/or shifts in the microbial community composition occurring from west to east across the regional gradient and from uplands to lowlands at the landscape scale.

While there is some support for improved abiotic conditions (less water limitation) and increased substrate availability (net primary production and labile organic matter) across topographic and regional gradients in the Great Plains, little is known about the variability of the microbial community. Several studies have shown microbial biomass increases from west to east across the region (Zak et al., 1994; Vinton and Burke, 1997), but reports on topographic microbial biomass trends vary (Schimel et al., 1985; Zak et al., 1994). Few data exist to compare the composition of the microbial community of these native grasslands. Work from the International Biological Program suggest that fungi dominate grassland microbial communities and that fungal biomass increases from western shortgrass steppe to eastern tallgrass prairie, but results differ depending on the methodology used (Paul et al., 1979; Coleman et al., 1980; Ulehlova, 1992). Soil microbiota perform the biogeochemical transformations that determine ecosystem C and N cycling rates; therefore, quantification of the variability in the microbial community composition is necessary to better understand regional and landscape-level differences in biogeochemical cycling.

The primary objective of this study was to determine whether microbial community composition and biomass differ regionally across three floristically distinct grassland community types (shortgrass steppe, mixed grass, and tallgrass prairie) and topographically across uplands and lowlands at specific sites. Previous work has shown variability in microbial community composition can be caused by alterations in plant species composition (Grayston et al., 1998), management practices (Calderon et al., 2001), soil type (Schutter et al., 2001), and seasonal variability in water, temperature, and substrate availability (Bardgett et al., 1999; Steinberger et al., 1999). Thus, we hypothesized that distinct microbial communities would be associated with the three grassland community types and with the uplands and lowlands at individual sites, reflecting changes in plant community composition and edaphic conditions across the regional and topographic gradients. In an attempt to better understand the major sources of variability in the microbial biomass and community composition and hence, the biogeochemistry of Great Plains grasslands, we also quantified whether spatial variability across the regional scale was larger or smaller than local variability displayed across landscape-level topographic units at individual sites.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Sampling Design
To accomplish these objectives, we established five grassland sites spanning an 800-km transect from eastern Colorado to eastern Kansas (Table 1). All five sites were sampled to determine whether microbial community composition and biomass differed regionally across these grassland community types. To address our objectives associated with landscape-level topographic variability, we sampled four uplands and lowlands at three of the five sites.


View this table:
[in this window]
[in a new window]
  		Table 1. Regional site characteristics with all soil parameters taken from the top 10 cm.{dagger}

 
Study Area
All sites are native grassland managed with moderate levels of cattle grazing and have not been previously cultivated, as evidenced by detailed site histories and well-developed soil profiles that lack a plow layer. The two westernmost sites in eastern Colorado, the Shortgrass Steppe Long-term Ecological Research site (SGS) and The Nature Conservancy owned Fox ranch on the Arickaree River (ARI) are typical shortgrass steppe plant communities, dominated by buffalo grass [Buchloë dactyloides (Nutt.) Columbus] and blue grama grass [Bouteloua gracilis (Kunth) Lag. ex Griffiths, nom. illeg.] (Great Plains Flora Association, 1986; Lauenroth and Milchunas, 1992). Smoky Valley Ranch (SVR), located in western Kansas and also owned by The Nature Conservancy, has a mixture of both shortgrass steppe and mixed grass prairie vegetation. In central Kansas, the range area managed by Fort Hays State University (HAYS) is dominated by plants typically found in southern mixed grass prairie, little bluestem [Schizachyrium scoparium (Michx.) Nash] and sideoats grama [Bouteloua curtipendula (Michx.) Torr.] (Coupland, 1992). The Konza Prairie Biological Station (KONZA) is managed by Kansas State University and is also a Long-Term Ecological Research site. KONZA is a tallgrass prairie with dominant plant species being big bluestem (Andropogon gerardii) and Indian grass [Sorghastrum nutans (L.)] (Silletti and Knapp, 2000). Konza Prairie Biological Station is the only site where typical grassland management includes fire. Our plots at this site were located in a watershed that is burned once every 4 yr. The plots were burned in April 1996 and 2000.

Both mean annual temperature and precipitation increase from west to east across this transect (Table 1). We selected these sites while attempting to minimize textural differences across the gradient. Biogeochemical parameters measured as part of a concurrent study at these sites (McCulley, 2002) generally reflect previously reported trends for this region (e.g., increasing nutrient cycling rates from west to east across the region, Table 1). At each of the five sites, four permanent fenced plots (4 m by 4 m) on a level upland were established in June 1999. These permanent plots represent the "regional dataset" discussed herein.

Field Sampling
To address the questions about regional trends, we ran two 5-m transects from two randomly chosen permanent plots at each site. Along these transects, we collected three soil samples with a soil corer (0- to 10-cm depth, i.d. = 4.8 cm) at 1, 3, and 5 m in October 2000 (n = 6 per site). Previous work from the semi-arid shortgrass steppe has shown that the presence and absence of plants can have significant effects on measured biogeochemical parameters (Hook et al., 1991), and that this effect varies across the regional gradient (Burke et al., 1999). However, the objective of this paper was to detect microbial community variability associated with plant community type across the region and at differing landscape positions; therefore, we sampled soils randomly (both under and between plants) in an attempt to capture the maximum variability representative of the plant community type. One of the 5-m transects sampled in October 2000 was resampled in June 2001 (n = 3 per site) at each of the five sites. June 2001 soil samples were taken from within 50 cm of the October 2000 sample. Both the October 2000 and June 2001 samples are included in the regional dataset (total n = 9 per site).

The landscape-level sampling occurred at SGS, HAYS, and KONZA only, each representing one of the grassland types encountered across this transect (shortgrass steppe, mixed grass, and tallgrass prairie). On the same day we sampled the permanent plots for the regional dataset in October 2000, four uplands and four lowlands not associated with the regional permanent plots were sampled at these sites. At all uplands and lowlands, one soil sample (0–10 cm) was randomly collected (n = 8 per site). These data comprise the landscape dataset.

We removed all soil samples from the soil corer with the smallest amount of disturbance possible, stored them in plastic bags, and put them immediately on dry ice. Samples were kept on dry ice for <24 h before being transferred to a –70°C freezer, where they were stored for up to 6 mo before we performed microbial community composition analysis. While previous work has shown that storage of samples can affect microbial community composition, storage effects were small compared with differences associated with soil type and extraction method (Schutter and Dick, 2000). All of our samples were stored and extracted in a similar manner. Therefore, differences between sites and landscape positions are reflective of actual microbial community composition differences and not simply storage artifacts.

Microbial Community Composition
We assessed microbial community composition using PLFA analysis (Vestal and White, 1989). Phospholipid fatty acid analysis allows quantification of the viable microbial biomass and the taxonomic group composition at the time of sampling based on the total amount of phospholipids extracted and the chemical make-up of the constituent fatty acids (Vestal and White, 1989). Frozen samples were allowed to thaw for 15 to 30 min. Each sample/core was homogenized, and then ~20 g of field moist soil was extracted in a single phase, phosphate buffered CHCl3–CH3OH solution to remove PLFAs (Bligh and Dyer, 1959). We separated PLFAs by silicic acid chromatography and derivitized the PLFAs in an alkaline system to form fatty acid methyl esters (FAMEs) (White et al., 1979). We separated and quantified FAMEs on a Hewlett Packard 5890 Series 2 gas chromatograph (equipped with a 50-m long, 0.33-µm thick, and 0.2-mm i.d. column) linked to a Hewlett Packard 5971A mass spectrometer detector (Hewlett Packard, Palo Alto, CA). Column temperature was held at 70°C for the first 0.5 min of the analysis and was then increased to 110°C (at 20°C min–1) and then to 120°C (at 10°C min–1). The injector and detector were maintained at 240 and 280°C, respectively. Individual FAME spectra were identified using an existing FAME library (D.C. White, personal communication, 1999).

Fatty acid methyl esters are described by standard nomenclature (IUPAC-IUB, 1977) (A:B{omega}C), where "A" indicates the total number of C atoms, "B" the number of unsaturated bonds, and "{omega}C" indicates the number of C atoms between the aliphatic end of the molecule and the first unsaturated bond. Cis and trans isomers are indicated by the suffixes c and t, respectively. Other notations are "Me" for methyl groups, "cy" for cyclopropyl groups, and the prefixes i and a for iso and anteiso methyl branching, respectively.

Soil Characteristics
After soil samples were processed for microbial community composition, a 10-g subsample was dried at 110°C for 3 d to determine gravimetric water content. We air-dried and sieved (2 mm) the remaining soil. From this sieved soil, a subsample was ground in a ball mill, acid-washed to remove inorganic C, and analyzed on a LECO CHN-1000 analyzer ( LECO, St. Joseph, MI) for organic C and total N. Additional subsamples were used for texture analysis using the hydrometer method (Gee and Bauder, 1986). The range area managed by Fort Hays State University soils contained significant quantities of calcium carbonate; therefore, soil organic C was determined via the loss-on-ignition method (Nelson and Sommers, 1996).

Statistical Analyses
We utilized two statistical approaches to determine whether the microbial community composition and biomass data differed regionally across the grassland types and topographically across uplands and lowlands at specific sites. Analysis of variance (ANOVA) was used to determine whether significant differences in the total quantity of PLFA extracted from the soil samples (and the soil parameters) existed across the regional gradient of sites (SGS, ARI, SVR, HAYS, and KONZA) and across the topographic landscape-level gradients (location being either upland or lowland). We transformed data as needed to fit the normality assumptions, and we used least significant difference means separation tests to interpret the significant main effects of ‘site’ and ‘location.’

For the microbial community composition data, we first ran a principal component analysis (PCA) on the mole fraction of 19 FAMEs that were consistently found in all the samples and represent a wide array of microbial functional groups (additional FAMEs identified but excluded from the PCA occurred in low amounts, <1.0% mole in all cases). To maximize the power of the PCA, we incorporated both the landscape and regional datasets into one analysis. We log transformed the mole fraction of each FAME used in the PCA to achieve normal distributions. Once we acquired the first and second principal component axes/weights for each sample, we then ran an ANOVA similar to that used for the total PLFA data to determine whether the principal component axes scores for the samples were significantly different across the regional site gradient and between landscape-level topographic positions (location). To further evaluate the change in microbial community composition, the mole fractions of the FAMEs used in the PCA were separated into their representative taxonomic groups (gram positive bacteria, gram negative bacteria, nonspecific bacteria, actinomycetes, fungi, and protozoa—based on White et al., 1996; Vestal and White, 1989; Myers et al., 2001) and summed so that each sample had one relative abundance value for each taxonomic group. We then ran these taxonomic group data through the aforementioned ANOVA to assess the site and location main effects. The gram-negative and actinomycetes groups were log transformed to achieve normal distributions, and least significant difference means separation tests were used to evaluate significant main effects.

In an attempt to address whether the spatial variability in microbial community composition and biomass across the region was larger or smaller than the local spatial variability across landscape-level topographic units at individual sites, we calculated the proportion of variance accounted for by the main effects of the regional site and landscape-level location from the ANOVAs with the microbial taxonomic group and total PLFA data. This provides a quantitative estimate of each of these main effect sources of variability within their respective analysis of variance models that can then be compared. All ANOVAs and the PCA were performed using SAS, version 8 (SAS Institute, 1989).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Soil Characteristics
Soil C percentage, N percentage, and C/N ratio increased across the region from semi-arid shortgrass steppe to subhumid mixed and tallgrass prairie (Table 1). The coarsest and finest textured soils occurred at ARI and KONZA, respectively. Bulk density mirrored these textural trends, with highest bulk density corresponding to the coarsest soils, and soil pH decreased from west to east across the region (Table 1).

In the landscape dataset, all soil parameters had a significant site main effect (p < 0.05). Trends in C percentage, N percentage, C/N ratio, and clay percentage across the region were similar to those found in the regional dataset (Table 2). Uplands had higher sand content and lower soil moisture values than lowlands (p < 0.005), but other soil characteristics did not differ significantly across landscape positions.


View this table:
[in this window]
[in a new window]
  		Table 2. Landscape soil characteristics (0–10 cm) from uplands and lowlands at three of the regional sites{dagger}.

 
Microbial Biomass
Total PLFA is an estimate of viable microbial biomass at the time of sampling (White et al., 1996). In general, total PLFA increased from semi-arid shortgrass steppe to subhumid mixed grass and tallgrass prairie (Fig. 1 and 2) . Means separation tests on the regional dataset indicated increasing total PLFA with ARI < SVR < SGS, KZ, and HAYS (p < 0.0001). Somewhat different results for the means separation tests on the main effect ‘site’ were found for total PLFA in the landscape dataset (SGS, HAYS < KZ, Fig. 2, p < 0.0001). There were no significant differences between uplands and lowlands at any of the three sites sampled for the landscape dataset (Fig. 2).



View larger version (17K):
[in this window]
[in a new window]
  		Fig. 1. Total phospholipid fatty acids (PLFA) (mean ± 1 S.D.) for the five sites across the regional transect averaged over the October 2000 and June 2001 sampling time periods. Significant differences between sites (p < 0.0001) are represented by different letters on the graph.

 


View larger version (20K):
[in this window]
[in a new window]
  		Fig. 2. Total phospholipid fatty acids (PLFA) (mean ± 1 S.D.) of uplands and lowlands (location) for the landscape dataset. Analysis of variance results indicated significant differences existed between sites (p = 0.0011, indicated by different letters between sites), but there were no significant differences between locations at any of the sites.

 
Microbial Community Composition
The PCA on the combined datasets identified three significant axes explaining a combined total of 58% of the variance in the FAME multivariate data (Fig. 3) . To simplify presentation and interpretation of the data, only the first and second principal component axes are discussed here. Principal Component Axis 1 (PC 1) accounted for 24.5% of the variability and was significantly related to site for both the regional (p < 0.0001) and the landscape dataset (p = 0.015) but was not related to landscape location, suggesting that the sites differed significantly in their microbial community composition. Means separation tests of PC 1 weights were different for the two datasets (regional data set—ARI, KZ > SVR, SGS > HAYS, landscape dataset—SGS > KZ, HAYS). Fatty acid methyl esters indicative of fungi were negatively weighted on PC 1, while actinomycetal and bacterial groups had positive weights (Table 3).



View larger version (27K):
[in this window]
[in a new window]
  		Fig. 3. (a) Ordination of the regional dataset (Principal Component Axes 1 and 2—PC 1 and PC 2; n = 6 for SGS due to three samples from the October 2000 sample date being lost during transport and lab processing; for ARI, two samples occupy almost identical locations on the ordination are therefore represented by one point). Microbial community composition differed among grassland community types, with shortgrass steppe sites (SGS and ARI) located in the first and second quadrants, the tallgrass prairie site (KONZA) located in the third and fourth quadrants, and the mixed grass prairie sites (SVR and HAYS) occurring intermediately in the principal component space between the other two community types. (b) Ordination of the landscape dataset (with the same PC 1 and PC 2 as in Fig. 3a since both datasets were analyzed together in one principal component analysis; n = 3 for uplands at HAYS due to sample loss during transport, and n = 3 for lowlands at KONZA because an outlier data point was excluded from this analysis). Different microbial communities were associated with the different grassland community types and occur in the same general principal component space as in Fig. 3a. Uplands and lowlands also exhibit different microbial communities, with uplands generally higher on the PC 2 axis than lowlands for each site.

 

View this table:
[in this window]
[in a new window]
  		Table 3. Microbial PLFAs receiving large positive (>0.2) or negative weights (<–0.2) on the first and second principal component axes.

 
Principal Component Axis 2 (PC 2) explained 18.9% of the variability in the FAME data, with negative weights on FAMEs indicative of eubacterial anaerobes, gram-negative, gram-positive, and nonspecific bacteria. FAMEs with positive weights on PC 2 included gram-negative and gram-positive bacteria, as well as fungi (Table 3). Similar to PC 1, PC 2 was significantly related to site in both the regional and landscape datasets (p < 0.0001, for both), and means separation tests on PC 2 weights for both datasets indicated shortgrass steppe > mixed grass > tallgrass prairie. Landscape location was also significant (p = 0.004), with uplands having higher PC 2 weights than lowlands particularly at HAYS (Fig. 3b).

Microbial Taxonomic Groups
As noted above, statistical analyses on FAMEs associated with different microbial groups indicate that there were significant differences among the sites in microbial community composition. We separated the mole fraction of the 19 FAMEs into their representative taxonomic groups (gram positive bacteria, gram negative bacteria, nonspecific bacteria, actinomycetes, fungi, and protozoa; Fig. 4) and conducted analysis of variance on the relative abundances of these different groups. In the regional dataset, the relative abundance of all microbial groups was significantly related to site except protozoa (Table 4). The shortgrass steppe site ARI generally had higher relative abundances of FAMEs indicative of gram-positive bacteria, nonspecific bacteria, and actinomycetal taxonomic groups than the other sites (Fig. 4). Konza Prairie Biological Station had a higher relative abundance of FAME cy19:0, an indicator of gram-negative anaerobic bacteria, and lower relative abundances of FAMEs indicative of fungi than the other sites (Fig. 4). Relative abundances of gram-negative bacteria, nonspecific bacteria, and actinomycetal taxonomic groups were also significantly different across sites in the landscape dataset (Table 4), and means separation tests on these groups yielded similar results to those of the regional dataset.



View larger version (40K):
[in this window]
[in a new window]
  		Fig. 4. Mean percentage of mole fraction (across October 2000 and June 2001 sample dates) of the 19 fatty acid methyl esters (FAMEs) included in the principal component analysis for the five sites in the regional dataset and their placement into the microbial taxonomic groups (bacterial, actinomycetal, fungal, and protozoan). Different letters represent significant differences between sites (p < 0.05).

 

View this table:
[in this window]
[in a new window]
  		Table 4. Analysis of variance results (p-values) for the effects of site, location, and their interaction on the microbial functional groups.{dagger}

 
Nonspecific bacteria were the only microbial group to vary across upland and lowland locations (p = 0.012, Table 4). Lowlands had greater relative abundances of these nonspecific bacterial markers than uplands. This result suggests that the FAMEs associated with this microbial group (i14:0, 14:0, 17:0, and 18:0) were responsible for the separation of uplands and lowlands along PC 2 in the PCA.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSIONS
 REFERENCES
 
Microbial Biomass
Microbial biomass and activity are largely governed by soil moisture, temperature, pH, and the quantity and quality of available substrates (Wardle, 1992). The three grassland community types we studied exist across a 500-mm mean annual precipitation and a 4°C mean annual temperature gradient, with net primary production and rates of biogeochemical processes increasing concomitantly from western semi-arid shortgrass steppe to eastern subhumid tallgrass prairie (Table 1). Along this gradient of decreasing climatic limitations and increasing substrate availability, we found larger viable microbial biomass pools in the mixed grass and tallgrass prairie than the shortgrass steppe (although the SGS site is an exception). This result is consistent with previously reported microbial biomass estimates that have shown shortgrass steppe soils have less microbial biomass than tallgrass prairie soils (Zak et al., 1994) and with the microbial biomass trends observed across a regional precipitation gradient in the Judean desert, namely increasing microbial biomass with increasing precipitation (Steinberger et al., 1999). Our microbial biomass values for the tallgrass prairie site, KONZA, are within the range of those reported from restored tallgrass prairie in Illinois (KONZA: 80 to 150 nmol PLFA g soil–1; Illinois: 79 to 123 nmol PLFA g soil–1; Bailey et al., 2002). Overall, our values fall within the range of total PLFA reported from temperate agricultural grasslands in the United Kingdom (Bardgett et al., 1999) and various forest types in Lower Michigan (Myers et al., 2001).

While total PLFA generally increased from shortgrass steppe to tallgrass prairie in both the regional and landscape datasets, means separation tests on the regional dataset indicated one of the shortgrass steppe sites, SGS, had total PLFA values similar to those measured at the mixed grass and tallgrass prairie sites. We believe this result may be related to soil texture. The regional SGS site was intentionally located on a fine-textured soil type to reduce textural differences across the regional dataset. However, the SGS landscape sites as well as the second regional shortgrass steppe site, ARI, were both located on coarser sandy loam soils and had lower total PLFA and C percent values. By modifying soil microclimatic conditions and soil organic C content, soil texture is known to play a role in determining microbial biomass and activity as well as impacting microbial community composition (Schimel et al., 1994; Zak et al., 1994; Scott et al., 1996; Schutter and Dick, 2000). This study was not designed specifically to address the role of soil texture on viable microbial biomass; however, the differences in total PLFA measured between the SGS regional and landscape sites (which were sampled on the same day in October) supports the idea that increased soil clay content leads to more stabilization of soil C and higher microbial biomass (Schimel et al., 1994).

Although previous studies have found increased substrate availability and improved microclimatic conditions in lowlands compared with uplands, and although we measured higher gravimetric soil moisture contents and lower percentages of sand in lowlands than uplands at all sites in the landscape dataset (Table 2), we did not find increased viable microbial biomass in lowlands compared with uplands (Fig. 2). Similarly, Zak et al. (1994) found no differences in microbial biomass across catenary sequences in either the shortgrass steppe or tallgrass prairie. These results may be due to the fact that both studies sampled uplands and lowlands at only one point in time (October in this study; July or August for Zak et al. [1994]). Landscape-level variability in substrate availability, via root exudation, and microclimatic conditions may be accentuated early during the growing season at these sites (Klein, 1977) and might then promote differentiation in total PLFA between uplands and lowlands.

Microbial Community Composition and Taxonomic Groups
The three grassland community types sampled in this study have floristically distinct species assemblages (Coupland, 1992; Kucera, 1992; Lauenroth and Milchunas, 1992). Plant species composition has been shown to influence the size and composition of the microbial pool, most likely via the quantity and types of root exudates made available to the soil microbial community (Klein, 1977; Groffman et al., 1996). Many studies have found distinct microbial communities associated with different management practices, soil type, and/or plant species dominance (Bossio et al., 1998; Grayston et al., 1998; Donnison et al., 2000; Calderon et al., 2001; Schutter et al., 2001). Given these previous results and the known differences in plant species assemblages across our sites, we predicted and found differences in microbial community composition across the different grassland types (Fig. 3).

All the microbial taxonomic groups identified, except protozoa, varied significantly across sites. The main trends in the microbial taxonomic group dataset were that the tallgrass prairie site had higher relative abundances of gram-negative, anaerobic bacteria, and less fungi than the other grassland communities. The tallgrass prairie site in this study had higher percentages of clay and gravimetric soil moisture than all of the other sites; thus, it is not surprising this site had the highest relative abundance of anaerobic bacterial markers, as both factors would result in higher anaerobicity within the soil. Based on the idea that fungi become more abundant as soil pH decreases and the recalcitrance of soil organic matter and litter increase (Alexander, 1977; Grayston et al., 2001), both of which occur across the regional gradient from shortgrass steppe to tallgrass prairie (Table 1; Murphy et al., 2002), we hypothesized that the relative abundance of fungi would increase from west to east across the Great Plains. However, this is not what we found. Previous work has shown that fungal biomass comprises a large portion of the total microbial biomass in tallgrass prairie soils (Miller et al., 1995; Rice et al., 1998), and one comparative study suggested that fungal biomass was eight to nine times greater in tallgrass prairie than shortgrass steppe (Paul et al., 1979). In contrast, our results indicate that fungal biomass does not increase across the gradient and that fungi make up a smaller proportion of the microbial community in tallgrass prairie than shortgrass steppe or mixed grass prairie. These conflicting results could be due to the different techniques employed in these studies or to the fact that the tallgrass prairie in this study was burned the spring before sampling. Burning is known to significantly impact many belowground processes in tallgrass prairie and may consequently have an effect on microbial community composition (Rice et al., 1998).

We identified differences in microbial community composition between uplands and lowlands across the three grassland community types. Lowlands had a higher relative abundance of nonspecific bacterial markers than uplands. This result is contradictory to the findings of Broughton and Gross (2000), who report no significant changes in the microbial community composition along a topographic gradient in a mid-successional grassland on an abandoned field in Michigan. The spatially extensive sampling design employed in this study (n = 3–4 lowlands and uplands per site; compared with the intensive sampling done on only one upland-lowland topographic sequence in Broughton and Gross [2000]) may have increased our probability of encountering shifts in microbial community composition associated with topographic gradients.

Comparison of Regional and Topographic Variability
We found strong evidence that microbial biomass and microbial community composition varies across this regional gradient in the central Great Plains. In addition, we found landscape-level topographic variability in the microbial community composition data but not the biomass dataset. The proportion of variance explained in the ANOVAs by the site and location main effects (the R2 for each main effect) for both the total viable microbial biomass and microbial community composition data (PC 1 and PC 2 weights) indicate that the main effect site accounted for more than ten times of the variation explained by the location main effect. This result suggests that regional variability in microbial biomass and community composition is much larger than landscape-level topographic variability. Most likely this is due to the variability in the controlling factors of microbial biomass and community composition, such as climate, soil organic matter, texture, and plant species composition, being larger across the central Great Plains than the topographic variability in these parameters at any particular site. However, it is important to note that the PCA accounted for only 43.4% of the variability in the microbial community composition dataset; therefore, there was substantial variability in the microbial community that was not accounted for by either site or location.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
REFERENCES
 
Microbial biomass increased from shortgrass steppe to mixed grass and tallgrass prairie and was accompanied by changes in the microbial community composition, such that the different grassland communities appear to have different associated microbial communities. While microbial community composition was slightly different between upland and lowland topographic positions, less variability in both microbial biomass and community composition was explained by the landscape-level topographic effects than the regional site main effect. The total phospholipid values reported in this study compare well with those from restored tallgrass prairie and with the general trends in microbial biomass previously reported for the region using a different technique. However, some interesting discrepancies between the microbial community composition data reported here and that of previous work exist, namely, the low relative abundance of fungi at the tallgrass prairie site and the slight shift in microbial community composition observed here between uplands and lowlands. More temporally and spatially extensive phospholipid analysis work would assist in clarifying the nature of these types of discrepancies and increase our understanding of the controlling variables on microbial community composition.


   ACKNOWLEDGMENTS
 
This work was funded by an NSF Dissertation Improvement award (DEB 0073189), the Colorado Agriculture Experiment Station, and the NSF Long Term Ecological Research Program's Shortgrass Steppe (DEB 9632852) and Konza Prairie Biological Station sites. We thank Dr. Ken Reardon and Shelley Allen for laboratory assistance with the PLFA analysis and Dr. Eugene Kelly for soil characterizations. We also thank Don Zak and two anonymous reviewers for comments that substantially improved the manuscript. Access to the field sites was provided by The Nature Conservancy, USDA-ARS, and Ft. Hays State University.

Received for publication December 16, 2002.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Soil Sci.Home page
V. J. Allison, Z. Yermakov, R. M. Miller, J. D. Jastrow, and R. Matamala
Assessing Soil Microbial Community Composition Across Landscapes: Do Surface Soils Reveal Patterns?
Soil Sci. Soc. Am. J., April 5, 2007; 71(3): 730 - 734.
[Abstract] [Full Text] [PDF]

Home page
 MycologiaHome page
A. Jumpponen and L. C. Johnson
Can rDNA analyses of diverse fungal communities in soil and roots detect effects of environmental manipulations--a case study from tallgrass prairie.
Mycologia, November 1, 2005; 97(6): 1177 - 1194.
[Abstract] [Full Text] [PDF]

Home page
 Soil Sci.Home page
V. J. Allison, R. M. Miller, J. D. Jastrow, R. Matamala, and D. R. Zak
Changes in Soil Microbial Community Structure in a Tallgrass Prairie Chronosequence
Soil Sci. Soc. Am. J., August 4, 2005; 69(5): 1412 - 1421.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (11)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by McCulley, R. L.
Right arrow Articles by Burke, I. C.
Right arrow Search for Related Content
PubMed
Right arrow Articles by McCulley, R. L.
Right arrow Articles by Burke, I. C.
Agricola
Right arrow Articles by McCulley, R. L.
Right arrow Articles by Burke, I. C.
Related Collections
Right arrow Soil Microbiology
Right arrow Range Soils
Right arrow Biogeochemical Processes

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Vadose Zone Journal Journal of Plant Registrations
Journal of Natural Resources
and Life Sciences Education		 Journal of
Environmental Quality