Microbial community composition on native and drastically disturbed serpentine soils

Microbial community composition on native and drastically disturbed serpentine soils

Soil Biology & Biochemistry 37 (2005) 1427–1435 www.elsevier.com/locate/soilbio Microbial community composition on native and drastically disturbed s...

315KB Sizes 1 Downloads 15 Views

Soil Biology & Biochemistry 37 (2005) 1427–1435 www.elsevier.com/locate/soilbio

Microbial community composition on native and drastically disturbed serpentine soils Shira H. DeGrood*, Victor P. Claassen, Kate M. Scow Land, Air, and Water Resources Department, Plant and Environmental Sciences, University of California, Building, 1 Shields Avenue, Davis, CA 95616, USA Received 8 July 2004; received in revised form 9 December 2004; accepted 20 December 2004

Abstract During construction of roads, entire hillsides can be cut away, dramatically disturbing the ecosystem. Microbial communities play important, but poorly understood roles in revegetating roadcuts because of the many functions they perform in nutrient cycling, plant symbioses, decomposition, and other ecosystem processes. Our objective was to determine relationships among microbial community composition, soil chemistry, and disturbance on a serpentine soil disturbed by a roadcut and then partially revegetated. We hypothesized that the adjacent undisturbed serpentine soil would have a different microbial community composition from barren and revegetated sections of the roadcut and that undisturbed soils would have the greatest microbial biomass and diversity. We measured phospholipid fatty acids (PLFA) and soil nutrient concentrations on barren and revegetated sections of the roadcut and on adjacent undisturbed serpentine and nonserpentine soils. Most roadcut samples had soil chemistry similar to the serpentine reference soil. The microbial biomass and diversity of barren sites was lower than that of revegetated or the serpentine reference site. The nonserpentine reference site had significantly (P%0.05) greater microbial biomass than serpentine or disturbed sites but significantly lower relative proportions of actinomycetes, and slow growth biomarkers. The Barren site had the lowest microbial biomass and a significantly (P%0.05) greater proportion of that biomass was fungi. Barren, revegetated, and serpentine sites all had dissimilar microbial community composition. The composition of the revegetated communities, however, was intermediate between the serpentine reference and barren soils, suggesting that community composition of revegetated soils is approaching that of an undisturbed site with similar soil chemistry. q 2005 Elsevier Ltd. All rights reserved. Keywords: Phospholipid fatty acid analysis; Disturbance; Serpentine; Revegetation; Microbial community composition; Soil characteristics; Heavy metals; Restoration; Biomarkers

1. Introduction Drastic disturbance is defined as removal of all the topsoil and biological materials on a site (Box, 1978), as occurs with deep excavation, landslide or mining activity. Disturbance of the soil or substrate causes dramatic changes in the taxonomic and functional diversity of soil microbial communities (Buckley and Schmidt, 2001; Chow et al., 2003; Laiho et al., 2003; Ponder and Tadros, 2002) and the effects of disturbance can be long lasting (Mummey et al., 2002b). Degraded sites may be restored to the original level of biotic integrity or rehabilitated to a modified sustainable * Corresponding author. Tel.: C1 530 752 0416; fax: C1 530 752 1552. E-mail address: [email protected] (S.H. DeGrood). 0038-0717/$ - see front matter q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.soilbio.2004.12.013

ecosystem. In either case, many essential functions for a sustainable community are mediated by soil microorganisms (SER, 2002), including nutrient cycling (Ponder and Tadros, 2002), soil structure (Axelrood et al., 2002; Mummey et al., 2002a), and biological interactions (vanBruggen and Termorshuizen, 2003; Weller, 2002). Therefore, successful revegetation projects depend heavily on regeneration of microbial diversity. Bradshaw (1984a,b) summarized a conceptual model of ecosystem development in which the goal of restoration is to shift a skeletal or disturbed soil material from a state of low biomass and diversity to greater biomass and diversity. Under this model, restoration of original microbial diversity, biomass, and, more accurately, complete microbial community composition, should facilitate sustainable restoration on disturbed sites.


S.H. DeGrood et al. / Soil Biology & Biochemistry 37 (2005) 1427–1435

Microbial populations, however, have received less study on restoration or rehabilitation projects than aboveground communities, despite the essential functions they provide. In areas, where soil conditions are marginal, the functions of soil microbes are critical for supporting plant growth and revegetation success. Ecosystem recovery on serpentine derived soils, for instance, is may be retarded by harsh edaphic conditions, including high concentrations of heavy metals such as chromium, nickel, cobalt, and manganese; a low calcium:magnesium ratio (Ca:Mg); and deficiencies in nitrogen, phosphorus, and potassium (Prasad and De Oliveira Freitas, 1999; Proctor, 1999; Robinson et al., 1997; Turitzen, 1982). The geological parent material, including the ultramafic mineral serpentinite, generates the low Ca:Mg and high heavy metal content (Proctor, 1999). The unique properties of serpentine soils are reported to reduce plant productivity and to support unique plant communities (Baker et al., 1991; Brooks, 1987; Harrison, 1999; Roberts and Proctor, 1992). The limited available data suggest that microbial and fungal biomass and community structure also differs between serpentinite derived soils and adjacent nonserpentine soils (Amir and Pineau, 1998a,b; Lipman, 1926; Maas and Stuntz, 1969). Our study site is located on serpentine substrates in central California. After several landslide events closed the highway, a 49,000 m2 roadcut was constructed in 1992 and seeded with native and exotic erosion control plant species. However, these revegetation attempts have only been successful on sections of the roadcut, where the original topsoil was respread. The rest of the slope remains almost completely barren over a decade later. Lack of plant establishment may be due to physical disturbance of the soil (Hart et al., 1999), disruption of the soil microbial community structure (Klironomos, 2002), and/or shifts in chemical and physical properties of the soil (Chiarucci et al., 1998). The objective of this study was to investigate changes in microbial communities, soil chemistry, and revegetation of a roadcut site on serpentine substrates. We hypothesized that the microbial community of an undisturbed reference serpentine soil would have greater microbial biomass, diversity, and different community composition than barren or revegetated substrates on the roadcut. We also hypothesized that the community composition of the reference site would be more similar to the revegetated sites than to the barren site. We used phospholipid fatty acid (PLFA) analysis to measure microbial community composition (Baath et al., 1998; Bossio et al., 1998; Song et al., 1999; Zelles, 1999), both to generate a fingerprint of dominant members of the microbial community and to estimate microbial biomass. The total number of PLFAs in a sample was used as a proxy indicator of microbial diversity to test Bradshaw’s conceptual model of ecosystem development (Bradshaw, 1984a,b), which investigates the relationship between total microbial biomass and microbial diversity.

2. Methods 2.1. Study site The study site is located in the Coast Range of central California, 55 km due north of San Francisco (Colusa County State Route 20mile 1.5). This steep (308), westfacing roadcut is about 91 m tall, 213 m wide, and extends back from the highway about 228 m, for a total area of about 49,000 m2. After slope stabilization in 1992, the California Department of Transportation emplaced five horizontal cement interceptor trenches to prevent water runoff across the slope. The surface was seeded with native and erosion-control grasses and shrubs. Plant biomass was reduced to zero within 4 years except on two areas, where topsoil from the site was reapplied. The roadcut substrates had serpentine mineralogy as indicated by the detection of the mineral groups serpentine and chlorite in the clay fraction of the site soils and substrates by X-ray diffraction (D.G. McGahan, personal communication). Serpentinization, or hydrothermal alteration of ultramafic parent materials, is often an incomplete process, resulting in a wide variety of ‘serpentinitized’ soils and substrates in the field (Burt et al., 2001). For simplicity, however, the substrates and soils in this study are referred to simply as ‘serpentine’, despite the heterogeneous nature of this mineralogical group. Two revegetated areas were sampled. The first (Reveg1) received a thin (2–3 cm) topsoil overlay and had moderate revegetation (20% plant cover). The second (Reveg2) received a 1.5 m deep backfill of serpentine topsoil and had dense vegetation (80% plant cover). Samples were collected from under groundcover, which consisted mainly of Vulpia microstachys and Bromus madritensis, and were near shrub species, including Arctostaphylos viscida and Ceanothus jepsonii (R. E. O’Dell, personal communication). Shrubs were less than 0.5 m tall. Samples were also collected from an adjacent non-vegetated area (Barren), where no revegetation treatment had been applied (0% plant cover). Four replicate samples were taken at random from each reference site. Each sample was collected to a depth of 10 cm by a trowel, which was rinsed with water after each sampling. Reference samples were taken from native serpentine and nonserpentine soils to represent undisturbed soil conditions. The serpentine reference site was on the opposite side of the hill from the roadcut, about 30 m away from the top of the roadcut. The nonserpentine reference site was on the slope facing the roadcut, about 200 m away. Both reference sites had similar percent slope as the roadcut. The serpentine reference site was mapped as a Henneke sandy loam (Clayey-skeletal, magnesic, thermic Lithic Argixeroll) (N.R.C.S., 2001). Samples were taken from under herbaceous cover, which included Lomatium marginatum and Agoseris

S.H. DeGrood et al. / Soil Biology & Biochemistry 37 (2005) 1427–1435

heterophylla (40% plant cover), surrounded by shrubs, including Quercus durata spp durata, Heteromeles arbutitolia, and Ceanothus jepsonii (R.E. O’Dell, personal communication). Shrubs reached approximately 2 m tall. The nonserpentine reference site was located on the slope of a hill opposite from the roadcut in an area mapped as a Skyhigh gravelly clay loam (Fine, smectitic, thermic Mollic Haploxeralf). Samples were taken from under grassland areas, mainly Vulpia microstachys, Avena fatua, and Triteleia laxa, near scattered black oak (Quercus kelloggii) (75% plant cover). Grasses reached 0.25 m and trees were approximately 20 m tall. Four replicate samples were taken at random from each reference site. Each sample was collected to a depth of 10 cm by a trowel, which was rinsed with water after each sampling. 2.2. Soil chemistry analysis Prior to analysis, soils were air-dried and sieved (!2 mm). Soils were analyzed at AandL Western Agricultural Laboratories, Inc. (Modesto, CA, analysis suite S3C) for organic matter (OM; loss on ignition, 350 8C, 2 h), electrical conductivity (EC; saturated paste), pH (dilute CaCl2), nutrient cations (neutral ammonium acetate), cation exchange capacity (sum of cations plus H), extractable P (Bray and Olsen), NOK 3 , and DTPA extractable Zn, Mn, Fe, Cu, Co, Ni, and Mo and BaCl2- extractable B as measured by ICP (Gavlak et al., 2003). Total N was analyzed by dry combustion on a Carlo Erba NA 1500 (Fisons Instruments, Beverly MA). 2.3. Phospholipid fatty acid analysis PLFA analysis was carried out as described in Bossio et al. (1998) with the following modification: 5 g dry weight of soil was extracted rather than 8 g soil. In brief, fatty acids were directly extracted from soil samples using chloroform: methanol: phosphate buffer. PLFAs were separated from neutral and glycolipid fatty acids on a solid phase extraction column (0.58 Si; Supelco Inc., Bellafonte, PA). After mild alkaline methanolysis, samples were analyzed using a Hewlett Packard 6890 Gas Chromatograph with 25 m Ultra 2 (5% phenyl)-methylpolysiloxane column (J & W Scientific, Folsom, CA). Fatty acids were quantified by comparison of the peak areas with those of an internal standard 19:0 peak. The peaks were named using bacterial standards and identification software from the Microbial Identification System (Microbial ID, Inc., Newark, DE). Prior to first use, peak identification was verified by comparing mass spectrometry EI spectra to spectra from standards and molecular weights were confirmed with chemical ionization spectra, using a Varian 3400 gas chromatograph interfaced with a Finigan ITD 806 mass spectrometer.


In fatty acid nomenclature, the basic form is ‘A:BuC’, where A is the total number of carbons, B is the number of double bonds, and C is the position of the double bond from the methyl end of the molecule. The suffixes ‘c’ and ‘t’ stand for cis and trans, the prefixes ‘i’, ‘ai’, and ‘me’ refer to iso, anteiso, and mid-chain methyl branching, and the prefix ‘cy’ refers to cyclopropyl rings (Navarrete et al., 2000). 2.4. Statistical analyses Multivariate analyses were performed using CANOCO version 4.0 (Microcomputer Power, Inc., Ithaca, NY). Correspondence Analysis (CA) was used to determine the similarity of soil chemistry across sampling locations. Samples with similar chemistry will have similar CA scores and thus will group closer together when plotted. Samples and soil nutrients on the same side of an axis are positively correlated, while variables on opposite sides of an axis are negatively correlated. For this study, CA was performed using samples from all five treatments and included: NOK 3 , Olsen-P, Bray-P, K, Ca, Mg, S, B, Cu, Fe, H, Mo, Na, Zn, Co, Cr, Mn, Ni, pH, cation exchange capacity (CEC), electrical conductivity (EC), and organic matter (OM). Canonical Correspondence Analysis (CCA) was used to examine the relationship among PLFA fingerprints, soil chemistry, and sample location. Like CA, similar CCA scores indicate samples with similar PLFA profiles. Soil chemistry variables were tested to determine which, if any, had a significant (P%0.05) effect on microbial community composition using the Monte Carlo permutation test. Vectors of greater magnitude and forming smaller angles with an axis are more strongly correlated with that axis. CCA was performed using all five treatments and using treatments with similar soil chemistry. CA and CCA can be unduly influenced by rare fatty acids. Fatty acids that only appear in a few samples are usually unreliably represented because they have values near the detection limit. Therefore, we omitted fatty acids that were present in less than 25% of the samples. To meet the CCA requirement that there must be fewer environmental variables than samples, the following environmental variables were analyzed in the CCA: Bray-P, K, Ca, Mg, S, Cu, Fe, Mo, Co, Mn, Ni, pH, CEC, and organic matter (OM). Eight microbial indices were calculated using biomarkers described in Bossio et al. (1998) (Table 1) and total microbial diversity was estimated by counting the total number of PLFAs present in each sample. ANOVA was performed on each microbial index after testing for homogeneity of variance (SAS version 8.02, 1999) followed by a Tukey mean comparison. The relationship between total microbial biomass and microbial diversity was investigated by plotting total microbial biomass against the number of PLFAs in each sample, to test the conceptual model proposed by Bradshaw (1984a,b).


S.H. DeGrood et al. / Soil Biology & Biochemistry 37 (2005) 1427–1435

Table 1 Biomarker indices investigated for significant differences Designation


Total Biomass % Gram (C) % Gram (K) % Fungi % Actinomycetes % Slow growth Fungal:Bacterial biomass GramK:GramC

Sum of all fatty acids, in nmole/g soil (DW) Branched fatty acids (i, a, Me)/Total Biomass Monounsaturated fatty acids (A:1uC)/Total Biomass 18:3 w6c (6,9,12) C18:2 w6, 9c/Total Biomass 10Me18:0C10 Me16:0/Total Biomass Cyclic fatty acids (cy)/Total Biomass 18:2 w6, 9c/i15:0C a15:0C15:0C i16:0C16:1u5c C i17:0C a17:0C17:0cy C17:0C18:1u7c C19:0cy Monounsaturated fatty acids (A:1uC) /Branched fatty acids (i, a, Me)

3. Results 3.1. Soil nutrients Revegetated roadcut and serpentine soils had similar extractable concentrations of OM, P, pH, Ca, Mg, Ca:Mg ratio, CEC, and micronutrients (Table 2). The Barren roadcut soils were similar but had lower concentrations of plantavailable Ca and Mg than either Reveg site, although the molar Ca:Mg ratios were similar. Nonserpentine soils had greater concentrations of several important macro and micronutrients than the serpentine and roadcut samples, including NOK 3 , P, K, Ca, Mo, and Zn. Nonserpentine samples had lower Mg and Ni concentration and lower Ca:Mg ratio. On a Correspondence Analysis biplot, Axis 1 explained 96.8% of the variation, while Axis 2 only explained 1.8% of

the variation in the data (Fig. 1). Serpentine and nonserpentine reference sites had dissimilar soil chemistry profiles. Samples with serpentine soil chemistry were enriched in Mg, Ni, Co, and Fe and deficient in P, K, Ca, B, Cu, and Zn compared to nonserpentine samples. The chemistry of most roadcut samples was similar to the serpentine reference site and grouped to the left of the origin. One replicate of each of the Barren and Reveg2 soil, however, had soil chemistry that was intermediate between the serpentine and nonserpentine samples. 3.2. PLFA multivariate analysis Canonical correspondence analysis (CCA) Axis 1 explained 48.5% of the variation and separated treatments based on soil chemistry (Fig. 2). The Reveg2 sample that

Table 2 Means and standard error of measured soil variables

OMa TotalNa BrayPb OlsenPb Kb Sb Mob Cab Mgb Ca:Mgb Nib Mnb Cob Crb Bb Cub Feb Hc Nab Znb PH CECc ECd a b c d e

In percent (w/w). mg kgK1. meq 100 gK1. mmhos cmK1. Below detection limit.






1.35 (0.17) 0.03 (0.01) 2.8 (0.43) 2.88 (0.51) 36.65 (4.83) 3.45 (1.98) O0.1e 480.13 (122.86) 1467.50 (188.37) 0.37 (0.15) 3.52 (0.61) 2.08 (0.27) 0.11 (0.01) O0.1e 0.08 (0.03) 0.18 (0.03) 2.03 (0.25) 0 (0) 9.35 (0.22) O0.1e 7.63 (0.08) 14.60 (1.25) O0.1e

2.08 (0.17) 0.07 (0.01) 2.83 (0.57) 4.35 (0.10) 59.68 (11.76) 1.30 (0.79) O0.1e 514.18 (45.74) 1798.00 (59.59) 0.29 (0.03) 5.87 (0.39) 4.83 (1.22) 0.13 (0.03) O0.1e 0.13 (0.03) 0.23 (0.03) 10.48 (5.64) 0 (0) 9.55 (0.34) 0.23 (0.05) 7.13 (0.05) 17.55 (0.35) 0.15 (0.03)

2.95 (0.22) 0.11 (0.00) 5.58 (1.09) 3.73 (0.85) 166.68 (22.89) 1.23 (0.77) O0.1e 819.55 (228.77) 2358.50 (299.31) 0.36 (0.10) 17.04 (2.64) 5.55 (0.47) 0.15 (0.02) O0.1e 0.25 (0.03) 0.38 (0.05) 11.03 (4.16) 0 (0) 10.20 (0.84) 0.40 (0.00) 7.18 (0.05) 23.93 (2.78) 0.15 (0.03)

3.35 (0.38) 0.13 (0.1) 3.08 (0.60) 4.83 (0.51) 91.03 (6.81) 0.83 (0.44) O0.1e 713.78 (37.39) 2889.50 (137.11) 0.25 (0.02) 18.71 (1.40) 7.05 (0.50) 0.16 (0.02) O0.1e 0.25 (0.03) 0.63 (0.09) 15.33 (0.79) 0.33 (0.21) 9.10 (0.34) 0.35 (0.06) 7.05 (0.13) 27.93 (1.27) 0.20 (0)

3.68 (0.18) 0.20 (0.01) 17.68 (3.44) 13.28 (2.35) 313.75 (14.58) 2.18 (0.86) 0.12 (0.02) 2992.25 (208.59) 707.03 (33.73) 4.28 (0.45) 2.14 (0.24) 6.35 (0.98) O0.1 O0.1 0.88 (0.06) 6.63 (0.53) 8.50 (1.66) 0.25 (0.17) 11.63 (0.89) 2.05 (0.18) 7.08 (0.14) 21.83 (0.94) 0.23 (0.03)

S.H. DeGrood et al. / Soil Biology & Biochemistry 37 (2005) 1427–1435


Fig. 1. CA ordination plot of soil chemistry for samples from a partially revegetated roadcut on serpentine soil and adjacent serpentine and nonserpentine soils in California, USA. Axis 1 and 2 explain 96.8 and 1.8% of the total variation, respectively. Axes are scaled to represent proportional significance of Axis 1 and 2. Plus symbols represent CA scores for environmental variables and are labeled to indicate which variable it represents. Circles indicate major groupings of samples with similar nutrient profiles. Reveg indicates revegetated roadcut samples. Barren indicates nonvegetated roadcut samples. One Reveg2 and one Barren sample were outside of the groupings. Serp indicates serpentine reference samples and Nonserp indicates nonserpentine references samples.

had soil chemistry somewhat similar to the nonserpentine reference also had a PLFA profile more similar to the nonserpentine reference than to the other Reveg samples. Axis 2 explained 19.8% of the variation and separated serpentine from Barren samples. Reveg1 and nonserpentine samples grouped around the origin, while CCA scores for Reveg2 samples had a wide range of positive values. Monte Carlo permutation tests indicated that K, OM, and pH significantly (P%0.05) explained variation in PLFA profiles of all five treatments. A second CCA combining only treatments with serpentine soil chemistry (serpentine reference, Reveg1, Reveg2, Barren) indicated that PLFA profiles of Reveg samples were intermediate between undisturbed serpentine and Barren samples and undisturbed serpentine and Barren PLFA profiles were dissimilar from each other (Fig. 3). Axis 1, which explained 40.6% of the variation, separated Barren from Reveg and serpentine samples. Axis 2, which explained 25.9% of the variation, separated Reveg and Barren from serpentine samples. Of the environmental variables included in the analysis, OM, pH, and K significantly explained variation in PLFA profiles of samples with serpentine soil chemistry.

Fig. 2. CCA ordination plot of PLFA profiles for samples from a partially revegetated roadcut on serpentine soil and adjacent serpentine and nonserpentine soils in California, USA. Axis 1 and 2 explain 48.5 and 19.8% of the total variation, respectively. Arrows signify vectors of nutrients that are significant by the Monte Carlo test (P%0.05). Circles indicate distribution of CA scores for each treatment. Reveg indicates revegetated roadcut samples. Barren indicates nonvegetated roadcut samples. Serp indicates serpentine reference samples and Nonserp indicates nonserpentine references samples.

3.3. PLFA biomarkers Tukey tests of biomarker indices showed that biomarkers for slow growth (Fdfn4,dfd15Z3.74, PZ0.0266; Fig. 4a) was significantly lower in the nonserpentine reference soil than the serpentine reference soil, as was the relative proportion of actinomycetes (Fdfn4,dfd15Z8.32, PZ0.001; Fig. 4b). Microbial biomass ranged from 16.3 to 100.4 nanomoles PLFA gK1 soil, with the highest values in the nonserpentine reference soil and lowest in the Barren samples (Fdfn4,dfd15Z 11.7, PZ0.0002; Fig. 4c). The relative proportion of fungi was significantly greater in the Barren than any vegetated site (Fdfn4,dfd15Z16.30, P!0.0001; Fig. 4d). The relative proportion of GramC was not significantly (PR0.05) different by a Tukey test (Fig. 4e), although ANOVA did show significant differences (Fdfn4,dfd15Z3.12, PZ0.0471 in the relative proportion of Gram- bacterial biomarkers (Fdfn4,dfd15Z1.85, PZ0.1722; Fig. 4f) nor in ratios of GramK:GramC (Fdfn4,dfd15Z1.93, PZ0.1585; Fig. 4g) and fungi:bacteria (Fdfn4,dfd15Z2.65, PZ0.0748; Fig. 4h).

Fig. 3. CCA ordination plot of PLFA profiles for samples from a partially revegetated roadcut on serpentine soil and adjacent serpentine soils in California, USA. Axis 1 and 2 explain 40.6 and 25.9% of the total variation, respectively. Arrows signify vectors of chemicals that are significant by the Monte Carlo test (P%0.05). Circles indicate distribution of CA scores for each treatment. Reveg indicates revegetated roadcut samples. Barren indicates nonvegetated roadcut samples. Serp indicates serpentine reference.


S.H. DeGrood et al. / Soil Biology & Biochemistry 37 (2005) 1427–1435

Fig. 4. Means of Various PLFA Biomarkers from a partially revegetated roadcut on serpentine soil and adjacent serpentine and nonserpentine soils in California, USA. ANOVA and Tukey tests (P%0.05) were used to compare treatments within each index. Bars represent means for Barren (B), revegetated 1 (R1), revegetated 2 (R2), serpentine reference (S), and nonserpentine reference (N). Letters above the bars indicate significant differences between each treatment. Indices are: Relative Proportion Slow Growth Biomarkers (a), Relative Proportion Actinomycetes (b), Total Biomass (c), Relative Proportion Fungi (d), Relative Proportion GramC (e), Relative Proportion GramK (f), GramK/GramC (g), and Fungi/Bacteria (h).

3.4. Ecosystem development

4. Discussion

A strong relationship was observed between microbial biomass and the total number of PLFAs, an indicator of microbial diversity (Fig. 5). Barren samples, which had the lowest biomass, also had the lowest microbial diversity, whereas the opposite was true for nonserpentine samples. The biomass and diversity of Reveg samples ranged widely within their grouping, though together they were intermediate in value between the Barren and nonserpentine soils. Most sites showed the same relationship of biomass and diversity between samples. The Barren site, however, had one sample with greater diversity but without an appreciative increase in biomass.

We evaluated whether microbial community composition followed Bradshaw’s (1984a,b) conceptual model of ecosystem development. This model examines the relationship between biomass and diversity of plant communities as important variables when comparing disturbed, recovering, and untouched ecosystems. Revegetation of an amended ecosystem may occur without achieving the original biomass and diversity. Ecosystems that are at ‘replacement’, have original biomass concentrations but diversity is still low. We found that revegetated sites had increased microbial diversity and biomass above barren sites and that the proportion of diversity to biomass was similar to

S.H. DeGrood et al. / Soil Biology & Biochemistry 37 (2005) 1427–1435

Fig. 5. Ecosystem Recovery Plot of total number of PLFAs plotted against total microbial biomass (in nanomoles PLFA gK1 soil) for a partially revegetated roadcut on serpentine soil and adjacent serpentine and nonserpentine soils in California, USA. Reveg indicates revegetated roadcut samples. Barren indicates nonvegetated roadcut samples. Serp indicates serpentine reference samples and Nonserp indicates nonserpentine references samples. Arrow indicates distribution of Reveg1 and Reveg2 samples. Circles indicate distribution of all other treatments.

that of the serpentine reference site. According to Bradshaw’s (1984a,b) model, the soil microbial community composition in the revegetated sites is undergoing true restoration, rather than simply replacement. Further analyses of the PLFA data indicated that microbial community composition has not been completely restored on the revegetated sections of the slope. Though Barren and the adjacent serpentine reference had similar soil chemistry, their microbial communities were different. Revegetated soils also had soil chemistry similar to the Barren and serpentine reference site but microbial community composition had similarities to both. These results support our original hypothesis that the composition of revegetated soils would converge on that of serpentine reference soils. A similar trend was observed at a mine overburden reclamation site in Wyoming, where fatty acid fingerprints from a 20-year-old revegetation site were more similar to the reference site than a 6-year-old revegetation site (Mummey et al., 2002a). Despite that, the microbial biomass and concentration of biomarkers for bacteria (15:0, cy17:0, cy19:0, i15:1, i17:0, ai17:0) and fungi (18:2 u6c) were significantly different between the reference and 20-year-old reclamation site (Mummey et al., 2002b). Selection during restoration of microorganisms capable of tolerating high heavy metal and Mg concentrations may explain why microbial community composition on Reveg sites was approaching that of the undisturbed site with serpentine, rather than nonserpentine, soil chemistry. Heavy metal additions to nonserpentine soils caused changes in the microbial community composition (Baath et al., 1998; Pennanen, 2001). However, heavy metal additions on serpentine soils only caused a minor shift in microbial community compared with nonserpentine soils


(He´ry et al., 2003), suggesting that the serpentine microbial community has greater adaptations to higher heavy metal conditions than nonserpentine microbial community. Topsoil addition was effective to revegetate the exposed substrates on the roadcut. Revegetation was unsuccessful on sections of the roadcut that had been left exposed. The detrimental effect on microbial and plant biomass from topsoil loss has been documented in other locations. In New Zealand, the top 31 cm of soil was removed from a pastureland through topsoil mining (Hart et al., 1999). Mining decreased total soil carbon from 8.2 to 3.9% and soil N from 0.51 to 0.24%. Even 3 years after revegetation, microbial C and mineralizable N was 35 and 38% lower on the mined site than the control plot. Aboveground plant biomass was 30% lower on the mined than control plot. Microbial community responses were also evaluated on irrigated farmland in Arkansas, from which the topsoil was scraped off for land leveling (Brye et al., 2003). Fungal biomass and bacterial biomass, estimated using direct-count microscopy, were significantly lower a month after leveling than a few months before leveling. However, the overall ratio of fungi:bacteria remained the same, suggesting that the microbial community composition did not change despite the overall reduction in microbial biomass. The finding of our study and others suggests that loss of topsoil severely effects microbial biomass and community composition. The microbial community will not recover without topsoil development. However, there will be little topsoil development without above- and belowground biomass development. Without active remediation efforts, slope sections without topsoil will likely remain a degraded environment for decades to come (Bradshaw, 1984a,b). The serpentine and nonserpentine reference sites differed in microbial community composition and extractable soil chemistry. CCA vectors indicated that nonserpentine microbial communities were associated with higher organic matter concentration and higher pH and lower concentrations of K than serpentine communities. We originally hypothesized that common serpentine characteristics, such as low Ca:Mg and high Ni would be the source of differences in microbial community composition between serpentine and nonserpentine soils, because they are considered important predictors of plant community composition (Harrison, 1999; Johnson and Proctor, 1981; Proctor et al., 1981). Studies of vegetation on Californian serpentine soils have indicated that low Ca concentrations play an important role in excluding certain plant species from serpentine grasslands (Brooks, 1987; Harrison, 1999; Kruckeberg, 1985). Alexander et al. (1989) concluded that increasing timber yield on Northern Californian serpentine forests was related to increasing Ca:Mg. A study by Robinson et al. (1997) also suggested that Mg and Ni are important in defining plant community composition on New Zealand serpentinite. Ca:Mg levels at our site were slightly higher than those of eight serpentine sites throughout Europe and New Caledonia but were still dramatically lower


S.H. DeGrood et al. / Soil Biology & Biochemistry 37 (2005) 1427–1435

than common Ca:Mg levels for nonserpentine soils in Europe (Brooks, 1987). We were surprised in our study to find that neither Ca, Mg nor Ca:Mg (data not shown) significantly explained variation in microbial community patterns. The significance of higher OM and K (a nutrient cation commonly adsorbed on organic matter) suggests that the microbes are responding to increased soil organic matter levels, probably because of the associated increase in nutrient availability and water holding capacity. In California’s Mediterranean climate, most microbial activity stops in the summer as soils dry out. Increased water holding capacity would increase springtime microbial activity and influence microbial community composition and biomass. Serpentine and nonserpentine reference sites also differed significantly in several microbial biomarkers. Total microbial biomass was significantly lower in the serpentine than nonserpentine soil. Other studies have noted similar trends in microbial biomass on serpentine soils (Amir and Pineau, 1998a; Lipman, 1926). This study also indicated that the relative proportion of actinomycetes was significantly greater in serpentine than in nonserpentine samples. Unusually high relative proportions of actinomycetes on serpentine soils have also been seen on New Caledonia serpentine derived soils. When bacteria from serpentinite-derived soils were cultured on soil agar plates and identified, Amir and Pineau (1998a) found that over 80% of the culturable bacteria were actinomycetes. In another study, serpentine and nonserpentine soils in New Caledonian were spiked with Ni and ITS-PCR was used to examine genetic diversity (He´ry et al., 2003). Visual inspection of banding patterns showed a difference in microbial community composition between serpentine and nonserpentine soils. After 14 days of incubation, both soils showed a new band that, when cloned, consisted mainly of Actinomycetes species. The noted ability of Actiomycetes to adapt to high concentration of heavy metals is likely related to the high concentration and quantity of Actinomycetes species. Because they fill numerous environmental niches, it is reasonable to expect that some of these species have developed resistance to heavy metals. Barren samples had the greatest relative proportion of fungi biomarkers, although the fungi:bacteria was not significantly different. Although fungi biomass decreases after disturbance (Mummey et al., 2002a; Klein, 2000), fungal biomass is greater than bacterial four (Mummey et al., 2002a) and 6 years (Klein, 2000) after disturbance. As noted by Reichardt et al. (2001), fungi are more tolerant to drought conditions than bacteria. Our site experiences drought conditions due to slope engineering and California’s Mediterranean climate, which may select against bacteria to a greater degree than fungi. In conclusion, we found evidence for a trajectory towards restoration, rather than replacement, by a similar microbial community as barren soils became revegetated. In contrast, non-topsoil amended soils have remained barren, with low

microbial biomass and diversity compared to the undisturbed serpentine reference soils. These data indicate that revegetation did not occur without appropriate management, such as topsoil amendment, and that by such treatment, the microbial community composition is comparable on disturbed and undisturbed serpentinite derived substrates. Conservation of microbial community composition may be due, in part, to similarities in plant community composition or to the continued generation of soil organic matter as much as to the unique soil chemistry of serpentinite derived soils. Future studies should evaluate if microbial functions important for restoration are recovered along with microbial community composition and if critical soil functions can be restored using soil organic amendments that are more readily available than harvested topsoil.

Acknowledgements We would like to use this opportunity to thank Dr Rebecca Drenovsky for her assistance with statistical interpretation of the data, Donald McGahan for his help with soil taxonomy and mineralogy, and Ryan O’Dell for his expertise on plant taxonomy. We would also like to thank the California Department of Transportation (RTA 65A0098) and the Consortium for Research at McLaughlin for making this study possible. Additional support came from the NIEHS Superfund Basic Research Program (2P42 ES04699) and the EPA Center for Ecological Health Research.

References Alexander, E., Adamson, C., Zinke, P., Graham, R., 1989. Soils and conifer forest productivity on serpentinized peridotite of the Trinity Ophiolite, California. Soil Science 148, 412–423. Amir, H., Pineau, R., 1998a. Influence of plants and cropping on microbiological characteristics of some New Caledonian ultramafic soils. Australian Journal of Soil Research 36, 457–471. Amir, H., Pineau, R., 1998b. Effects of metals on the germination and growth of fungal isolates from New Caledonian ultramafic soils. Soil Biology & Biochemistry 30, 2043–2054. Axelrood, P.E., Chow, M.L., Arnold, C.S., Lu, K., McDermott, J.M., Davies, J., 2002. Cultivation-dependent characterization of bacterial diversity from British Columbia forest soils subjected to disturbance. Canadian Journal of Microbiology 48, 643–654. Baath, E., Diaz-Ravina, M., Frostegard, A., Campbell, C.D., 1998. Effect of metal-rich sludge amendments on the soil microbial community. Applied and Environmental Microbiology 64, 238–245. Baker, A.J.M., Proctor, J., Reeves, R.D. (Eds.), 1991. The Vegetation of Ultramafic (Serpentine) Soils: Proceedings of the First International Conference on Serpentine Ecology. Davis, California, University of California, Davis. Bossio, D.A., Scow, K.M., Gunapala, N., Graham, K.J., 1998. Determinants of soil microbial communities: Effects of agricultural management, season, and soil type on phospholipid fatty acid profiles. Microbial Ecology 36, 1–12.

S.H. DeGrood et al. / Soil Biology & Biochemistry 37 (2005) 1427–1435 Box, T.W., 1978. The significance and responsibility of rehabilitating drastically disturbed land, in: Schaller, F.W., Sutton, P. (Eds.), Reclamation of Drastically Disturbed Lands. American Society of Agronomy, Madison, WI, pp. 1–10. Bradshaw, A.D., 1984a. Land Restoration—now and in the future. Proceedings of the Royal Society of London Series B-Biological Sciences 223, 1–23. Bradshaw, A.D., 1984b. Ecological principles and land reclamation practice. Landscape Planning 11, 35–48. Brooks, R.R., 1987. Serpentine and Its Vegetation: a Multidisciplinary Approach. Dioscorides Press, Portland, Oregon. Brye, K.R., Slaton, N.A., Savin, M.C., Norman, R.J., Miller, D.M., 2003. Short-term effects of land leveling on soil physical properties and microbial biomass. Soil Science Society of American Journal 67, 1405– 1417. Buckley, D.H., Schmidt, T.M., 2001. Environmental factors influencing the distribution of rRNA from Verrucomicrobia in soil. FEMS Microbiology Ecology 35, 105–112. Burt, R., Fillmore, M., Wilson, M.A., Gross, E.R., Langridge, R.W., Lammers, D.A., 2001. Soil properties of selected pedons on ultramafic rocks in Klamath Mountains, Oregon. Communications in Soil Science and Plant Analysis 32, 2145–2175. Chiarucci, A., Robinson, B.H., Bonini, I., Petit, D., Brooks, R.R., De Dominicis, V., 1998. Vegetation of Tuscan ultramafic soils in relation to edaphic and physical factors. Folia Geobotanica 33, 113–131. Chow, J.C., Watson, J.G., Ashbaugh, L.L., Magliano, K.L., 2003. Similarities and differences in PM10 chemical source profiles for geological dust from the San Joaquin Valley, California. Atmospheric Environment 37, 1317–1340. Gavlak, R., Horneck, D., Miller, R., Kotuby-Amacher, J., 2003. Soil, Plant and Water Reference Methods for the Western Region. Utah State University, Logan, UT. Harrison, S., 1999. Local and regional diversity in a patchy landscape: native, alien and endemic herbs on serpentine soils. Ecology 80, 70–80. Hart, P.B.S., West, A.W., Kings, J.A., Watts, H.M., Howe, J.C., 1999. Land restoration management after topsoil mining and implications for restoration policy guidelines in New Zealand. Land Degradation and Development 10, 435–453. He´ry, M., Nazaret, S., Jaffre´, T., Normand, P., Navarro, E., 2003. Adaptation to nickel spiking of bacterial communities in neocaledonian soils. Environmental Microbiology 5, 3–12. Johnston, W.R., Proctor, J., 1981. Growth of Serpentine and NonSerpentine Races of Festuca-Rubra in Solutions Simulating the Chemical Conditions in a Toxic Serpentine Soil. Journal of Ecology 69, 855–869. Klein, D.A., Paschke, M.W., 2000. A soil microbial community structuralfunctional index: the microscopy-based total/active/active fungal/bacterial (TA/AFB) biovolumes ratio. Applied Soil Ecology 14, 257–268. Klironomos, J.N., 2002. Feedback with soil biota contributes to plant rarity and invasiveness in communities. Nature 417, 67–70. Kruckeberg, A.R., 1985. Vegetation, Geology, Soils and Management Problems. University of California Press. Laiho, R., Sanchez, F., Tiarks, A., Dougherty, P.M., Trettin, C.C., 2003. Impacts of intensive forestry on early rotation trends in site carbon pools in the southeastern US. Forest Ecology and Management 174, 177–189.


Lipman, C.B., 1926. The bacterial flora of serpentine soils. Journal of Bacteriology 12, 315–318. Maas, J.L., Stuntz, D.E., 1969. Mycoecology on serpentine soil. Mycologia 61, 1106–1116. Mummey, D.L., Stahl, P.D., Buyer, J.S., 2002a. Microbial biomarkers as an indicator of ecosystem recovery following surface mine reclamation. Applied Soil Ecology 21, 251–259. Mummey, D.L., Stahl, P.D., Buyer, J.S., 2002b. Soil microbiological properties 20 years after surface mine reclamation: spatial analysis of reclaimed and undisturbed sites. Soil Biology & Biochemistry 34, 1717–1725. N.R.C.S., 2001. Soil Survey of Colusa County, California. [Online] (http://www.ca.nrcs.usda.gov/mlra02/colusa.html). Verified June 25, 2004. Navarrete, A., Peacock, A., Macnaughton, S.J., Urmeneta, J., MasCastella, J., White, D.C., Guerrero, R., 2000. Physiological status and community composition of microbial mats of the Ebro Delta, Spain, by signature lipid biomarkers. Microbial Ecology 39, 92–99. Pennanen, T., 2001. Microbial communities in boreal coniferous forest humus exposed to heavy metals and changes in soil pH—a summary of the use of phospholipid fatty acids. Biolog (R) and H-3-thymidine incorporation methods in field studies. Geoderma 100, 91–126. Ponder, F., Tadros, M., 2002. Phospholipid fatty acids in forest soil four years after organic matter removal and soil compaction. Applied Soil Ecology 19, 173–182. Prasad, M., De Oliveira Freitas, H., 1999. Feasible biotechnological and bioremediation strategies forserpentine soils and mine spoils. Electronic Journal of Biotechnology 2. Proctor, J., 1999. Toxins, nutrient shortages and droughts: the serpentine challenge. Trends in Ecology and Evolution 14, 334–335. Proctor, J., Johnston, W.R., Cottam, D.A., Wilson, A.B., 1981. Fieldcapacity water extracts from serpentine soils. Nature 294, 245–246. Reichardt, W., Briones, A., de Jesus, R., Padre, B., 2001. Microbial population shifts in experimental rice systems. Applied Soil Ecology 17, 151–163. Roberts, B.A., Proctor, J. (Eds.), 1992. The Ecology of Areas with Serpentinized Rocks: a World View. Kulwer, Boston, MA. Robinson, B., Brooks, R., Kirkman, J., Gregg, P., Alvarez, H., 1997. Edaphic influences on a New Zealand ultramafic (‘serpentine’) flora: a statistical approach. Plant and Soil 188, 11–20. SER, 2002. The SER primer on ecological restoration. Society for Ecological Restoration International Science and Policy Working Group. Song, X.H., Hopke, P.K., Bruns, M.A., Graham, K., Scow, K., 1999. Pattern recognition of soil samples based on the microbial fatty acid contents. Environmental Science and Technology 33, 3524–3530. Turitzen, S.N., 1982. Nutrient limitation to plant growth in a California serpentine grassland. American Midland Naturalist 107, 95–99. vanBruggen, A., Termorshuizen, A., 2003. Integrated approaches to root disease management in organic farming systems. Australasian Plant Pathology 32, 141–156. Weller, D.M., 2002. Microbial populations responsible for specific soil suppressiveness to plant pathogens. Annual Review of Phytopathology 40, 309–348. Zelles, L., 1999. Fatty acid patterns of phospholipids and lipopolysaccharides in the characterization of microbial communities in soil: a review. Biology and Fertility of Soils 29, 111–129.