Relative performance of native and exotic grass species in response to amendment of drastically disturbed serpentine substrates

Authors

  • RYAN E. O'DELL,

    1. Department of Land, Air and Water Resources (LAWR), Plant and Environmental Sciences Building (PES), University of California, One Shields Avenue, Davis, CA 95616–8627, USA
    Search for more papers by this author
  • VICTOR P. CLAASSEN

    1. Department of Land, Air and Water Resources (LAWR), Plant and Environmental Sciences Building (PES), University of California, One Shields Avenue, Davis, CA 95616–8627, USA
    Search for more papers by this author

Ryan E. O'Dell, Department of Land, Air and Water Resources (LAWR), Plant and Environmental Sciences Building (PES), University of California, One Shields Avenue, Davis, CA 95616–8627, USA (fax 1 530 752 1552; e-mail reodell@ucdavis.edu).

Summary

  • 1Limited information exists on approaches to effectively revegetate severely disturbed, barren, subgrade (unconsolidated parent material) serpentine substrates. Additionally, little is known about the invasion potential of exotic grasses from highly invaded non-serpentine environments into adjacent, relatively uninvaded serpentine environments following restoration efforts that use substrate nutrient enrichment to accelerate revegetation. This study investigated approaches to the revegetation of subgrade serpentine substrates with native, serpentine-tolerant grass species while limiting the establishment and reproduction of invasive annual grass species.
  • 2Biomass production was measured for two native serpentine perennial grass species and both biomass and seed production (fitness) were measured for one native annual and two invasive annual grass species. All species were grown on subgrade serpentine substrate amended with garden (yard) waste compost, slow-release nitrogen (N), phosphorus (P) and potassium (K) fertilizers and CaSO42H2O (gypsum) in both glasshouse and field environments. The primary goals of this study were to: (i) determine how substrate amendment with compost changed with time and (ii) identify the substrate amended combination that maximized native grass species biomass and seed production while limiting the productivity of invasive annual grass species.
  • 3Compost amendment of the subgrade serpentine substrate greatly increased plant-available N, P, K and calcium (Ca) levels, decreased plant-available heavy metals, and increased the cation exchange capacity. Substrate inline image concentration remained low for up to 60 days following compost amendment but then increased more than six-fold in the 120 days thereafter. No other substrate properties changed substantially over a period of 180 days.
  • 4Amendment of the serpentine substrate with 30% volume compost per volume substrate, and 220 mg each of N, P and K per kg substrate, maximized biomass and seed production of the native grass species. This amendment level, however, resulted in an undesirable increase in biomass and seed production (fitness) of invasive annual grass species, exceeding that of the native species.
  • 5 Synthesis and applications. Revegetation of drastically disturbed, barren, subgrade serpentine substrates with native, serpentine-tolerant plant species can be successfully achieved with a combination of yard waste compost (organic) and slow-release (inorganic) fertilizer amendment. Enrichment of low-fertility serpentine substrates to promote native plant growth, however, may encourage invasion of undesirable plant species into the native plant community, leading to habitat degradation. Therefore, aggressive control methods may be required to prevent invasion of exotic species into the revegetation community.

Introduction

Serpentine is common along the west coast of the USA, including the northern California coast ranges, Sierra Nevada foothills and south-western Oregon. Serpentine also occurs in Cuba, UK, Spain, Italy, Japan, South Africa, New Caledonia, New Zealand and south-west Australia. In California, serpentine occupies approximately 1·5% (6000 km2/406 260 km2) of the land area (Harrison, Inouye & Safford 2002). Maintaining vegetative cover is important for serpentine substrates because they are highly erosive and contain high concentrations of toxic heavy metals and asbestos (Kruckeberg 1984; Brooks 1987; Baker, Proctor & Reeves 1992). Barren serpentine sites are susceptible to wind erosion and represent a major source of naturally occurring asbestos pollution. Inhalation of air-borne asbestos can lead to asbestosis (scarring of lung tissue) and lung cancer in humans (Preger 1978).

Serpentine substrates are stressful environments for plant growth because of their adverse nutrient conditions, which limit plant productivity and reduce ground cover, even in some undisturbed sites. Serpentine substrates are frequently deficient in the essential macronutrients nitrogen (N), phosphorus (P) and potassium (K), have low calcium : magnesium (Ca:Mg) molar ratios and potentially high levels of phytotoxic heavy metals, including nickel (Ni), chromium (Cr) and cobalt (Co). The exceptionally low Ca:Mg molar ratio of serpentine soils is often cited as the most significant feature that differentiates it from soils derived from non-serpentine parent materials (Kruckeberg 1984; Roberts & Proctor 1992; Burt et al. 2001). Serpentine soils typically have Ca:Mg molar ratios lower than 1, while non-serpentine soils often have ratios substantially greater than 1 (Kruckeberg 1984; Burt et al. 2001). Macronutrient deficiency, low Ca:Mg molar ratio and high heavy metal concentrations appear to be common features of serpentine landscapes world-wide.

Floras rich in endemics have evolved in response to the chemically adverse features of serpentine soils. Serpentine plant communities in the northern California coast ranges collectively include an estimated 120 endemic taxa (5·5% of the total endemic taxa in California; Safford, Viers & Harrison 2005). Despite having a rich diversity of endemic taxa, serpentine soils are generally regarded as low-productivity environments in which the accompanying flora often exhibits typical morphological stress tolerance features, such as xeromorphic foliage, reduction in stature and density, and high root : shoot biomass ratios, as well as physiological features, including tolerance of N, P and Ca deficiency and Mg and heavy metal toxicity (Kruckeberg 1984; Brooks 1987). The stressful conditions imposed by serpentine environments act to exclude intolerant species, including many of the invasive annual species that now dominate non-serpentine plant communities throughout California (Huenneke et al. 1990; Harrison 1999; Weiss 1999; Safford & Harrison 2001).

When serpentine landscapes are disturbed by construction, erosion or mining, the relatively thin topsoil layer is easily lost, resulting in a loss of water-holding capacity, cation exchange capacity (CEC), organic matter (OM), plant essential nutrients, plant seed and mycorrhizal propagules. The inability of such sites to become revegetated by natural processes leaves them susceptible to erosion and mass wasting, resulting in further degradation of the environment through sedimentation of local watersheds. One such site in the northern Californian coast ranges is a 0·1-km2 road-cut at Colusa HWY 20, milepost 1.5 (henceforth referred to as the Colusa 1·5 site), that resulted from large-scale slope reconstruction following a landslide more than a decade previous to this study. The site is a steep 2:1 (horizontal : vertical, 27 degree) west-facing slope, with an overall slope length of approximately 220 m. Previous attempts to revegetate the site using conventional amendments and non-serpentine plant species were not successful and the slope has remained virtually barren since slope construction. Previous treatments included application of soluble N, P and K fertilizers, elemental sulphur (S), CaSO42H2O (gypsum) and a typical California roadside revegetation seed mix including commercial and native accessions of non-serpentine origins, including the invasive exotic annual grass red brome Bromus madritensis L. (Poaceae).

Native vegetation on undisturbed soils around the Colusa 1.5 site consists largely of serpentine chaparral dominated by the serpentine endemic, sclerophyllous, evergreen shrubs sticky whiteleaf manzanita Arctostaphylos viscida C. Parry ssp. pulchella (Howell) (Ericaceae), leather oak Quercus durata var. durata Jepson (Fagaceae) and white-flowered muskbrush Ceanothus jepsonii E. Greene var. albiflorus J. Howell (Rhamnaceae), corresponding to the leather oak vegetation series described by Sawyer & Keeler-Wolf (1995) and serpentine chaparral community described in detail by Kruckeberg (1984). Native grass species at the site include the perennials Bromus laevipes (chinook brome; Poaceae) Shear (Poaceae) and squirreltail Elymus elymoides (Raf.) Swezey (Poaceae) and the annual small fescue Vulpia microstachys (Nutt.) Munro (Poaceae). Many invasive exotic annual grass species, including Bromus madritensis and barbed goatgrass Aegilops triuncialis L. (Poaceae), have been observed growing in disturbed areas at and surrounding the site as well as in adjacent non-serpentine grassland communities, which are dominated by these invasive species.

A combination of glasshouse and field studies was initiated to investigate the effectiveness of using yard waste compost (an organic soil amendment) in combination with inorganic slow-release N, P, and K fertilizers and CaSO42H2O (gypsum) to promote the establishment and growth of serpentine-tolerant native grasses on subgrade (unconsolidated parent material) serpentine substrate from the site. Of all plant growth forms, grasses offer the greatest protection from surface substrate erosion because of their fibrous root systems and low-growing, thatchy canopy. Elymus elymoides, Bromus laevipes and Vulpia microstachys were included in the study in order to examine the response of serpentine native species to improved substrate nutrient and physical conditions on these subgrade substrates. Additionally, two invasive annual species occurring at the site, Bromus madritensis and Aegilops triuncialis, were included to study the potential for increased non-native species invasion as a result of improved substrate nutrient conditions. The primary goals of this study were to: (i) determine how substrate amendment with compost changed with time, and (ii) identify the substrate amendment combination that maximized native grass species biomass and seed production while limiting the productivity of invasive annual grass species.

Materials and methods

seed and substrate collection; substrate analysis

Seeds of all grass species used in this study were collected at the Colusa 1·5 site. Serpentine substrate was collected from the large, non-vegetated, subgrade, cut-slope face at the Colusa 1·5 site. Parent material at the Colusa 1.5 site was a serpentinitic foliated breccia containing an average of 46% (v/v) rock fragments (> 2 mm) embedded in a bluish-grey (Munsell colour 10 GY 6/1) sandy clay loam matrix (D. McGahan, unpublished data). The serpentinitic minerals of the parent material had been verified to include chrysotile asbestos amounting to 1–5% by mass, as analysed by dispersion staining and polarized light microscopy (Asbestech, Carmichael, CA). Serpentine topsoil 0–30 cm depth (Henneke soil series; clayey-skeletal, magnesic, thermic, lithic argixeroll) was collected from an adjacent vegetated, undisturbed slope at Colusa 1·5 and included as a functional soil comparison to the amended, subgrade serpentine treatments. Yard waste compost (US EPA Title 40, Part 503, Composting regulations; thermophilic process followed by a 90-day aerobic curing period) was obtained from the city of Redding municipal compost facility, Redding, California, USA. The yard waste compost consisted of decomposed yard clippings including grass, leaves and chipped stems. Each substrate was screened to 2-mm particle size to form a consistent matrix for glasshouse pot trials.

Unamended, subgrade serpentine substrate was analysed for total major and trace elements (Acme Analytical Laboratories Ltd, Vancouver, Canada; analysis suites 4A and 4B). In order to evaluate the nutrient content of the different substrate compost amendment treatments used in the study, screened compost was thoroughly mixed with various proportions of screened barren serpentine substrate (0%, 10%, 20%, 30%, 40%, 50% and 100% volume compost per total pot volume), with three replicates per treatment. Additional unplanted replicates of the 30% treatment were constructed, placed in the glasshouse, and subject to the same conditions as the planted treatments. These additional samples were analysed after various lengths of incubation to evaluate changes in substrate physical and nutrient conditions as compost decomposition and curing progressed. Triplicate samples of the glasshouse pot study unamended serpentine substrate, amended serpentine substrates and serpentine topsoil were collected at the study onset. Additionally, samples were randomly collected at the start of the study from the unamended field study and 30% compost-amended plots established at the Colusa 1·5 site. All samples were submitted for analysis of initial physical properties and available nutrients (A & L Western Agricultural Laboratories Inc., Modesto, CA; analysis suite S3C). Triplicate samples of the 30% compost samples were also submitted after 30, 60, 90 and 180 days incubation. The following methods were used to analyse the various treatments for plant nutrients: saturated paste pH (Rhoades & Miyamoto 1990), bicarbonate extractable inline image (Olsen & Sommers 1982), ammonium acetate (1 m, neutral pH) extractable Ca2+ and Mg2+ (Thomas 1982), KCl (2 m) extractable inline image (Keeney & Nelson 1982) and DTPA (0·1 m) extract for Ni2+, Cr3+ and Co2+ (Lindsay & Norvell 1978). Total C and total N were analysed by micro-Dumas dry combustion using a Carlo Erba NA 1500 NC elemental analyser (Fisions Instruments, Milan, Italy) (Dumas 1831 cited in Bremner & Mulvaney 1982).

glasshouse pot study establishment and harvest

In winter 2003, barren serpentine substrate from the Colusa 1·5 barren cut-slope face was amended with various combinations of yard waste compost, powdered agricultural-grade CaSO4 · 2H2O (gypsum), powdered, reagent-grade MgHPO4 · 3H2O and slow-release, polymer-coated ProTurf 38-0-0 and ProTurf 0-0-45 fertilizers (Scotts Company, Marysville, OH) to create the treatments summarized in Table 1. As no plants had ever been successfully grown on the subgrade serpentine substrate and as some of the fertilizers were contained in slow-release formulations that restricted release rates, high total nutrient rates of amendment were applied in order to guarantee sufficiency of these nutrients. Plastic pots with 900 cm3 volume (100 cm2 substrate surface area) were filled with the various treatments and seeded to obtain five seedlings per pot for Bromus laevipes (500 seeds m−2), three for Elymus elymoides (300 seeds m−2), 15 for Bromus madritensis (1500 seeds m−2), 15 for Vulpia microstachys (1500 seeds m−2) and four for Aegilops triuncialis (400 seeds m−2). These plant densities approximate those observed in undisturbed vegetation for each species near the Colusa 1·5 site. The seeded pots were completely randomized in a glasshouse at 21 °C and maintained at field moisture capacity with distilled water through the duration of the study. During the course of the study, glasshouse temperature, humidity and illumination were similar to ambient conditions for Davis, California.

Table 1.  Quantities of amendment applied to the subgrade serpentine substrate. –, amendment not present in treatment.
TreatmentCompost (% v/v)N (mg kg−1)P (mg kg−1)K (mg kg−1)Ca (mg kg−1)
0%  0
10% 10
20% 20
30% 30
30% + NPK 30220220220
30% + NPK + Ca 30220220220550
NPK + Ca220220220550
40% 40
50% 50
100%100
Serp. topsoil

All plants were harvested at 165 days post-seeding. At harvest, the perennial species Bromus laevipes and Elymus elymoides were still growing, whereas the annual species Vulpia microstachys, Bromus madritensis and Aegilops triuncialis had senesced following flowering and seed set. Above-ground biomass was clipped at the substrate surface and separated into shoot and seed components. The substrate was washed away to recover the root biomass. The separate biomass components were then oven dried at 60 °C and weighed.

field study establishment and harvest

Twenty 0·7-m2 unamended serpentine (control) and 200·7-m2 30% (v/v) yard waste compost-amended plots were established as a group on the subgrade, cut-slope face of the Colusa 1·5 site in winter 2004. The 30% compost-amended plots were amended by applying compost to the surface of the substrate and tilling it into the serpentine substrate with a tractor-mounted backhoe, to a depth of 30 cm. The control plots were tilled to a depth of 30 cm, but were not amended. Each plot was randomly selected and broadcast seeded with a single species at the same rate as that of the pot study, producing three unamended control and three compost-amended plots for each of the five grass species. Plots were subject to the prevailing environmental conditions of the site, which are very similar to that for Davis, California, for the duration of the study.

All plots were harvested at 165 days post-seeding. At harvest, the perennial species Bromus laevipes and Elymus elymoides were still growing, whereas the annual species Vulpia microstachys, Bromus madritensis and Aegilops triuncialis had senesced following flowering and seed set. Above-ground biomass was clipped at the substrate surface and separated into shoot and seed components. Root biomass was not sampled. The separate shoot biomass components were oven dried at 60 °C and weighed.

data analysis

The reproductive fitness index (henceforth referred to as the fitness index) for the annual species examined in this study was calculated as the ratio of ‘seeds out : seeds in’. This ratio is also known as the intrinsic growth rate of the population (Begon, Harper & Townsend 1986). A fitness index > 1 suggests population expansion, whereas fitness < 1 suggests population decline.

Annual grass species have discrete generations and are semelparous, meaning that individuals in the population germinate during the wet season (winter in California), with flowering and seed set occurring during the growing season (spring and summer) (Begon, Harper & Townsend 1986; Gulmon 1992). Factors affecting fitness include seed mortality, seed and plant predation and density-dependent mortality. The glasshouse study eliminated the effects of most of these factors by utilizing pre-germinated seedlings grown in a predator-free enclosure. In the field environment, such factors were probably present and may have had a minor impact on fitness. Calculation of the fitness index in this study assumed that 100% of the annual grass seeds produced the previous growing season would germinate and establish. The germination rate of individual seeds used in this study were very high, with a 94% and 97% germination rate for both Vulpia microstachys and Bromus madritensis, respectively. A study of Vulpia microstachys and soft chess Bromus hordeaceus (a close invasive annual relative of Bromus madritensis) seed germination in a serpentine grassland found similarly high germination rates (97% and 98%, respectively) by the beginning of winter for seeds produced the previous growing season (Gulmon 1992).

Seeds of Aegilops triuncialis represent a special case for the evaluation of seed germination. When Aegilops triuncialis is naturally dispersed as dimorphic seed pairs, the smaller seeds of these sibling pairs are chemically inhibited by larger sibling seeds and exhibit prolonged dormancy (more than one wet season–dry season cycle; Dyer 2004). Chemically suppressed seeds contribute to the seed bank and may germinate in subsequent years once the chemical inhibitor is leached away. For this study, seed pairs were disarticulated, after which germination of individual seeds was approximately 95%.

All treatment results were evaluated by one-way analysis of variance (anova) with mean separation from the unamended control (0% compost) established by Fisher's LSD. The significance level for all statistical tests was set at α= 0·05. All statistical analyses were conducted using Statistica 6·1 (Statsoft Inc., Tulsa, OK).

Results

substrate nutrients

The greatest portion of the serpentine substrate by mass was composed of SiO2, Fe2O3, Al2O3 and MgO, while CaO, K2O, P2O5, Na2O and MnO were relatively minor constituents (Table 2). Ni and Cr were contained at much higher concentrations in the serpentine substrate than Co. Initial nutrient content of the amended barren serpentine substrates increased in proportion to increased compost addition (Table 3). Although amendment with 10% (v/v) compost increased most of the measured nutrient levels of the subgrade serpentine substrate to that of the serpentine topsoil, compost amendment at or above 30% was required to generate biomass production equal to or greater than serpentine topsoil for all five grass species. Details of plant response data are given below.

Table 2.  Elemental composition of unamended subgrade serpentine substrate. Concentrations are expressed on an oxide or elemental basis
SiO2 (%)Al2O3 (%)Fe2O3 (%)MgO (%)CaO (%)Na2O (%)K2O (%)P2O5 (%)MnO (%)Ni (mg kg−1)Cr (mg kg−1)Co (mg kg−1)
38·961·556·6836·640·480·020·040·040·1021161067102
Table 3.  Glasshouse compost-amended subgrade serpentine substrates and serpentine topsoil treatment initial nutrient concentrations. Means; n = 3; –, no data
Glasshouse treatmentspHCEC (molc kg−1)Total C (mg kg−1)Total N (mg kg−1)N (inline image) (mg kg−1)P (inline image) (mg kg−1)Ca (mg kg−1)Mg (mg kg−1)Ca:Mg molNi (mg kg−1)Cr (mg kg−1)Co (mg kg−1)
0%8·50·09  3365  13410 4 498 7950·44·310·100·14
5%8·50·11  9765  5521111 659 8050·5
10%8·80·13 15057  948 820 851 8780·6
20%8·90·15 28301 189114281096 8710·8
30%9·00·17 47795 321818321332 8540·91·03< 0·10< 0·10
40%8·90·20 62162 449420421643 8861·1
50%8·70·23 82237 575923521913 8751·3
100%8·60·302707412006836652663 8541·9< 0·10< 0·10< 0·10
Serp topsoil7·10·29 12552  89018 4 77030170·218·710·100·16

Amendment of the subgrade serpentine substrate with 30% (v/v) compost increased sum of cations (CEC) two-fold, total C eight-fold, total N 14-fold, N (inline image) two-fold, P (inline image) eight-fold, K+ 55-fold and Ca2+ three-fold. Mg2+ concentration and pH remained generally unchanged. Substrate nutrient conditions for the unamended and 30% compost-amended field plots at the Colusa 1·5 site were very similar to the 0% and 30% compost treatments for the glasshouse study (Table 4). When the serpentine substrate was amended with 30% compost, all of the initial physical and nutrient parameters were equivalent to, or greatly exceeded that of, the serpentine topsoil. CEC and Mg2+, however, were much higher in the serpentine topsoil than in the 30% compost-amended serpentine substrate. Ni2+, Cr3+ and Co2+ levels of both the unamended subgrade serpentine substrates and serpentine topsoil were generally low and in all cases below 20 mg kg−1.

Table 4.  Field plot unamended (0%) and compost-amended (30%) treatment initial nutrient concentrations. Means, n= 3 randomly selected points within the plot. –, no data
Field plot treatmentspHCEC (molc kg−1)Total C (mg kg−1)Total N (mg kg−1)N (inline image) (mg kg−1)P (inline image) (mg kg−1)Ca (mg kg−1)Mg (mg kg−1)Ca:Mg molNi (mg kg−1)Cr (mg kg−1)Co (mg kg−1)
0%8·60·07 4210 213 4 7 507 4970·6
30%8·70·176443448021437141314131·3

Several substrate properties of the 30% compost treatment changed substantially during the 180-day incubation period (Table 5). CEC increased approximately 1·5-fold, and N (inline image) increased sixfold during this period. Levels of N (inline image) remained low for the first 30 days of the study, but increased markedly in the next 30 days. P (inline image) concentration, in contrast, fluctuated greatly over the incubation period. Both Ca2+ and Mg2+ concentrations increased slightly, resulting in a relatively stable substrate Ca:Mg molar ratio over the incubation period.

Table 5.  30% compost treatment nutrient concentrations after 0-, 30-, 60-, 90- and 180-day incubation periods in the glasshouse environment. Means; n= 3; –, no data
DayspHCEC (molc kg−1)Total C (mg kg−1)Total N (mg kg−1)N (inline image) (mg kg−1)P (inline image) (mg kg−1)Ca (mg kg−1)Mg (mg kg−1)Ca:Mg molNi (mg kg−1)Cr (mg kg−1)Co (mg kg−1)
09·00·17477953218 18321332 8530·91·03< 0·10< 0·10
308·70·17 1824113610130·7
608·50·22465293483 7115140315470·6
908·60·22498173668 9190152914300·7
1808·50·2347514369611546188212900·9

biomass production

Biomass production of all five grass species increased with compost amendment alone and compost amendment in combination with NPK fertilizer in the glasshouse environment. Each species varied, however, with respect to substrate tolerance, degree of amendment response and shoot vs. root biomass allocation trends. Seedlings of Bromus laevipes, Vulpia microstachys and Bromus madritensis displayed signs of root tip necrosis and inhibited root elongation after germination and early growth on the unamended serpentine substrate. Both Elymus elymoides and Aegilops triuncialis, however, had healthy white roots that extended deep into the unamended substrate, indicating their inherent tolerance of subgrade serpentine substrates.

Above-ground and below-ground biomass of the two native perennial species, Bromus laevipes and Elymus elymoides, increased approximately two-fold for each 10% increase in compost amendment up to 30% (Fig. 1). Addition of 30% or more compost to the subgrade serpentine substrate resulted in significant increases of both shoot and root biomass in Bromus laevipes compared with the unamended (0%) control (Fisher's LSD, n = 3, P < 0·05). Elymus elymoides root biomass increased significantly at the 30% compost amendment level relative to the unamended (0%) control (Fisher's LSD, n = 3, P < 0·01), but shoot biomass did not significantly increase at compost amendment levels less than 50% (Fisher's LSD, n= 3, P= 0·05). The 30% + NPK treatment resulted in a four-fold biomass production increase for both Bromus laevipes and Elymus elymoides compared with the 30% compost treatment. The 30% + NPK + Ca treatment did not result in a significantly greater increase in biomass production compared with the 30% + NPK fertilizer treatment for either Bromus laevipes or Elymus elymoides (Fisher's LSD, n = 3, P = 0·05). In both species, addition of 30% compost alone yielded approximately the same shoot and root biomass as the serpentine topsoil.

Figure 1.

Shoot and root biomass production for Bromus laevipes (Brla), Elymus elymoides (Elel), Vulpia microstachys (Vumi), Bromus madritensis (Brma) and Aegilops triuncialis (Aetr) subject to study treatments in a glasshouse environment. The statistical significance of study treatments compared with zero control (0%) by anova, Fisher's LSD is as follows: NS, not significant; * = P < 0·05 (significant); **P < 0·01 (highly significant); ***P < 0·001 (very highly significant). Bars denote means ± SE, n= 3. Grey bar, seed biomass; black bar, shoot or root biomass. Significance corresponds to total shoot (seed and shoot combined) biomass.

Annual grass growth response to amendment was similar to that of the perennial species, although biomass allocation patterns differed (Fig. 1). All three annual grass species had relatively greater biomass allocation to shoots than roots, resulting in lower root : shoot ratios than the two native perennial species (Table 6). Both shoot and root biomass were significantly greater for the 30% + NPK treatment than the 0% control or 10%, 20% and 30% compost treatments for all three species (Fisher's LSD, n = 3, P < 0·05). Aegilops triuncialis demonstrated greater tolerance of the subgrade serpentine substrate by producing more shoot and root biomass at amendment levels at and lower than 30% compost than the native annual Vulpia microstachys or invasive annual Bromus madritensis (Fig. 1). Shoot and root biomass production did not significantly differ between Vulpia microstachys and Bromus madritensis at any of the treatments (Fisher's LSD, n = 3, P = 0·05). All three annual grass species had approximately the same shoot and root biomass production at the 30% compost level as the serpentine topsoil.

Table 6.  Root : shoot ratios for the study species subject to the study treatments. Means; n= 3; –, no ratio because of zero value shoot or root biomass
SpeciesTreatmentSerp. topsoil
0%10%20%30%30% + NPK30% + NPK + Ca40%50%100%
Bromus laevipes 0·40·91·30·80·81·10·70·61·6
Elymus elymoides 2·11·41·51·60·80·61·40·90·61·8
Vulpia microstachys 0·20·50·70·50·40·80·60·30·8
Bromus madritensis 0·10·71·00·70·50·71·00·61·1
Aegilops triuncialis 1·21·61·41·10·60·41·41·10·61·7

Similar to the glasshouse study, all five grass species in the field study produced significantly greater biomass for the 30% compost treatment than the unamended (0%) control (Fisher's LSD, n = 3, P < 0·05; Fig. 1 vs. Fig. 2). Although above-ground biomass was similar between the glasshouse and field environments for the two perennial grass species, biomass production for all three annual grass species was much greater in the field environment than the glasshouse environment, perhaps due to local differences in growing conditions between the two environments.

Figure 2.

Shoot biomass production for Bromus laevipes (Brla), Elymus elymoides (Elel), Vulpia microstachys (Vumi), Bromus madritensis (Brma) and Aegilops triuncialis (Aetr) subject to unamended control (0%) and compost treatment (30%) in the field environment. Statistical significance of 30% compared with 0% treatment by anova, Fisher's LSD is as follows: NS, not significant; *P < 0·05 (significant); **P < 0·01 (highly significant); ***P < 0·001 (very highly significant). Bars denote means ± SE, n= 3. Grey bar, seed biomass; black bar, shoot biomass. Significance corresponds to total shoot (seed and shoot combined) biomass.

annual grass seed production

In general, compost, NPK and Ca amendment of the subgrade serpentine substrate increased seed biomass production for Vulpia microstachys and Bromus madritensis in the glasshouse environment (Fig. 1). Although 30% compost significantly increased above-ground biomass production of Vulpia microstachys and Bromus madritensis compared with the unamended control (Fisher's LSD, n = 3, P < 0·05), seed production did not significantly increase for either species with the same treatment (Fisher's LSD, n = 3, P = 0·05). Bromus madritensis seed biomass production was significantly greater (six-fold) for the 30% + NPK than the 30% compost treatment (Fisher's LSD, n = 3, P < 0·05). In contrast, Vulpia microstachys seed production was not significantly greater for the 30% + NPK treatment than the 30% compost treatment (Fisher's LSD, n = 3, P = 0·05). Vulpia microstachys seed production was significantly greater, however, for the 30% + NPK + Ca treatment than the 30% + NPK treatment (Fisher's LSD, n= 3, P < 0·05), indicating a positive response to Ca addition. In contrast to Vulpia microstachys and Bromus madritensis, there was no significant increase in seed biomass production with amendment at any level for Aegilops triuncialis (Fisher's LSD, n = 3, P = 0·05). Unlike the glasshouse environment, Vulpia microstachys and Bromus madritensis seed biomass production was significantly greater for the compost treatment (30%) than the unamended (0%) control in the field environment (Fisher's LSD, n = 3, P < 0·05; Fig. 2).

Fitness indices for all three annual grass species in the glasshouse environment exceeded 1 at compost amendment levels at and greater than 20% and were less than 1 at compost levels less than 20% (Fig. 3). Fitness indices for Vulpia microstachys and Bromus madritensis tended to increase with increasing levels of amendment, but only became significant with very high levels of compost amendment (> 30%) and fertilizer (30% + NPK) compared with the unamended control (Fisher's LSD, n = 3, P < 0·05). Fitness indices for Aegilops triuncialis, however, tended to remain near 1 even with high levels of amendment. In the field environment, fitness indices for all three species were significantly greater for the compost-amended (30%) plots than the unamended (0%) control plots (Fisher's LSD, n = 3, P < 0·05; Fig. 4). Fitness indices for all three species were much greater than 1 in the compost-amended plots and below 1 for the unamended control plots, and were more than three-fold, 1·5-fold, and sevenfold greater, respectively, in the field than the glasshouse environment (Fig. 4 vs. Fig. 3).

Figure 3.

Fitness indices for Vulpia microstachys (Vumi; top panel), Bromus madritensis (Brma; middle panel) and Aegilops triuncialis (Aetr; bottom panel) subject to study treatments in a glasshouse environment. Statistical significance of study treatments compared with zero control (0%) by anova, Fisher's LSD is as follows: NS, not significant; *P < 0·05 (significant); **P < 0·01 (highly significant); ***P < 0·001 (very highly significant). Bars denote means ± SE, n = 3.

Figure 4.

Fitness indices for Vulpia microstachys (Vumi), Bromus madritensis (Brma) and Aegilops triuncialis (Aetr) subject to unamended control (0%) and compost treatment (30%) in the field environment. Statistical significance of 30% compared with 0% treatment by anova, Fisher's LSD is as follows: NS, not significant; *P < 0·05 (significant); **P < 0·01 (highly significant); ***P < 0·001 (very highly significant). Bars denote means ± SE, n= 3.

Discussion

Amendment of the subgrade serpentine substrate with 30% (v/v) compost and 220 mg kg substrate−1 each of N, P and K as slow-release fertilizer greatly increased shoot and root biomass of the native perennial grass species Bromus laevipes and Elymus elymoides, as well as shoot biomass, root biomass and seed production of the native annual grass species Vulpia microstachys. The same amendments, however, also resulted in an undesirable increase in biomass and seed production of the invasive annual grass species Bromus madritensis and Aegilops triuncialis. Increases in invasive grass biomass productivity on infertile serpentine soils in response to NPKCa fertilization have been documented in several studies. In those studies, fertilization of California coast range serpentine soils with NPK resulted in significant (up to four-fold) increases in above-ground biomass production in California native serpentine grass species including Elymus elymoides and Vulpia microstachys (Turitzin 1982; Smith & Kay 1986; Koide et al. 1988). The same treatment, however, also resulted in significant (up to 10-fold) increases in above-ground biomass production for invasive annual grass species including Bromus madritensis (Turitzin 1982; Smith & Kay 1986; Huenneke et al. 1990). The findings of this study and previous studies suggest that deficiency of N, P and/or K is highly limiting to plant growth on serpentine substrates in California. Therefore, amendment of the serpentine substrate with N, P and K in the form of organic (compost) and/or inorganic (fertilizer) compounds is crucial for sustainable plant productivity and cover of the revegetation community.

While sufficient substrate levels of N, P and K appear to be important for plant biomass production, substrate Ca availability appears to be important for root structural integrity and elongation. Elymus elymoides and Aegilops triuncialis seedlings demonstrated a high tolerance of low substrate Ca:Mg molar ratios by rooting deeply into the unamended serpentine substrate. Seedlings of Bromus laevipes, Vulpia microstachys, and Bromus madritensis, however, had signs of root tip necrosis and greatly inhibited root elongation when grown on the unamended serpentine substrate. Amendment of the low Ca subgrade serpentine substrate (0·4 Ca:Mg molar) with minimal amounts (20% v/v) of the Ca-rich yard waste compost amendment (1·9 Ca:Mg molar) doubled the substrate Ca:Mg molar ratio to 0·8. The increase in substrate Ca:Mg molar ratio greatly improved the ability of Bromus laevipes, Vulpia microstachys and Bromus madritensis to root into the substrate. Although some native, serpentine-tolerant plant species including Elymus elymoides are exceedingly tolerant of low substrate Ca:Mg molar ratios, Ca amendment of subgrade serpentine substrate is essential for the establishment and healthy root growth of less-tolerant, but desirable, native serpentine plant species.

Organic amendments containing Ca appear to be able to ameliorate substrate Ca deficiencies in a way that inorganic Ca amendments such as gypsum do not. Ca addition to compost-amended serpentine substrates (this study) and NPK-amended serpentine soils (Turitzin 1982; Carter, Proctor & Slingsby 1988; Huenneke et al. 1990) resulted in little additional increase in biomass production of serpentine native annual, serpentine native perennial or non-serpentine invasive annual grass species. The lack of, and in some cases negative, response to Ca amendment may be the result of the ability of Ca2+ to displace Mg2+ from the cation exchange complex, thereby increasing substrate solution Mg2+ concentration to levels that may inhibit plant growth (Meyer 1980). A high Ca content, but low-release rate granular organic, mineral or synthetic resin polymer cation exchange complex may therefore prove to be a more favourable amendment to increase low substrate Ca:Mg molar ratios than more soluble sources of Ca such as gypsum. High Ca exchange complex amendments, such as compost, have the ability to provide a bioavailable source of Ca without the potentially adverse effects of Mg displacement from the substrate exchange complex into the substrate solution.

High substrate concentrations of Mg (low substrate Ca:Mg molar ratios) are antagonistic to plant Ca uptake (Marschner 2002). Physiological mechanisms conveying tolerance to serpentine substrates, including active Ca2+ uptake (calcium-proton antiporter CAX1 in Arabidopsis thaliana (mouseear cress; Brassicaceae), Mg2+ exclusion and Mg2+ toxicity tolerance, have been identified in numerous serpentine-tolerant plant species globally and may be present in Elymus elymoides and Aegilops triuncialis as well (Lyon et al. 1971; Shewry & Peterson 1975; Wallace, Jones & Alexander 1982; Lee & Reeves 1989; Bradshaw 2005; O'Dell & Claassen 2006; O'Dell, James & Richards 2006). Elymus elymoides has been shown to have a shoot Ca:Mg molar ratio two-fold that of the serpentine soil supporting it, suggesting the presence of such tolerance mechanisms (Smith & Kay 1986).

Aegilops triuncialis has demonstrated its tolerance of serpentine substrates by invading native serpentine grasslands throughout California (Harrison 1999). Evidence is currently lacking as to whether distinct serpentine-tolerant ecotypes of this species have evolved since the species’ introduction to California in the early 20th century (J. McKay & K. Rice, unpublished data). The native origin of Aegilops triuncialis from stressful substrates in western and central Asia, the short period of time since it invaded California grasslands in the coast ranges approximately 40 years ago, and its current widespread distribution in serpentine grasslands today, suggest that this species may have contained the tolerance mechanisms required for survival on serpentine substrates prior to its introduction. Selection of appropriate revegetation plant species and seed source is critical for successful vegetation establishment and productivity. Site-collected seed of native serpentine-tolerant plant species is superior for the revegetation of serpentine substrate (O'Dell & Claassen 2006). Native species growing on serpentine substrates have the stress tolerance features required for establishment and survival on serpentine, as well as adaptations to the local climate. Although readily available, most seed procured from commercial sources is of non-serpentine origins and, thus, poorly adapted for establishment and survival on serpentine.

In addition to increasing substrate fertility, organic matter in compost has also been demonstrated to chelate toxic heavy metals such as Ni and Cr in serpentine soils, thereby reducing their negative effects on root growth (Halstead 1968; Fernández, Seoane, & Merino 1999). Although substrate Ni concentration was elevated in the subgrade serpentine substrate and serpentine topsoil compared with typical non-serpentine soils, the concentrations were still lower than that shown to impact root growth in sensitive crop species (Baccouch, Chaoui & El Ferjani 1998). Organic matter amendments such as compost further improve root growth and overall health of plants on serpentine through increased porosity and infiltration, enhancement of microbial activity resulting in the formation of stable soil aggregates, and by providing a long-lasting nutrient supply. The multiple benefits provided by compost, including a long-lasting nutrient supply, a rich source of bioavailable Ca and improved substrate physical properties, make it an effective amendment for the revegetation of subgrade serpentine substrates.

Compost amendment dramatically increased initial total N, N (inline image) and P (inline image) in the subgrade serpentine substrate. Relatively low substrate inline image availability in the first 30 days following compost amendment required the addition of a fertilizer containing N in order to achieve maximum plant productivity. In one study, NPK fertilizer applied in combination with an organic amendment optimized above-ground biomass production for Avena sativa L. (oats; Poaceae) on a serpentine soil (Halstead 1968). NPK fertilization alone, however, resulted in only moderate biomass increases. In another study, only the application of NPK fertilizer in combination with manure resulted in stable vegetation cover for more than two growing seasons on subgrade serpentine asbestos mine wastes (Moore & Zimmermann 1977). Amendment with either fertilizer or manure alone resulted in poor vegetative cover that rapidly deteriorated after the first growing season. The results of these two studies support our finding that the highest plant productivity on serpentine substrates was achieved when organic and NPK fertilizer amendments were applied together. The elevated rate of N release from fertilizers can compensate for slow initial N release rates from organic amendments (Claassen & Hogan 1998).

In addition to greatly increasing biomass and seed production of native grass species, NPK fertilizer and organic amendments have also resulted in undesirable increases in invasive grass biomass and seed production on serpentine substrates (Turitzin 1982; Smith & Kay 1986; Koide et al. 1988; Huenneke et al. 1990). In California north coast range serpentine grasslands, NPKCa fertilization resulted in significant increases of invasive annual forb biomass along with simultaneous significant declines in native annual forb abundance, producing communities with greatly reduced species richness that were dominated by only a few invasive annual grass species (Koide et al. 1988; Huenneke et al. 1990). The adverse chemical nature of serpentine substrates, including deficiency of N, P, K and Ca, is perhaps the single most important feature that maintains relatively pristine, native serpentine plant communities predominately free of invasive species.

The necessary amendment of drastically disturbed, subgrade serpentine substrates in order to promote native plant establishment may diminish the unique serpentine chemical features that function to exclude invasive species (Huenneke et al. 1990; Harrison 1999; Safford & Harrison 2001). Additionally, the close proximity of serpentine road-cuts to roads makes such sites extremely vulnerable to invasion (Williamson & Harrison 2002; Gelbard & Harrison 2003). As a result, aggressive invasive annual species control may be required to prevent encroachment from adjacent non-serpentine communities into serpentine revegetation communities. Effective control methods for invasive annual species on serpentine substrates include timely mowing (early spring, before seed-set) and burning (summer/early autumn, after seed-set), followed by reseeding (prior to the first autumn rains) with native serpentine revegetation species in order to maintain a dense stand of native vegetation (Harrison, Inouye & Safford 2002; Seabloom et al. 2003).

Acknowledgements

This study was funded by California Department of Transportation RTA #65A0098 through the Caltrans Research and Innovation program and #43A0073 through the Caltrans Stormwater program. We extend our thanks to J. K. McKay and S. P. Harrison for manuscript review and constructive comments.

Ancillary