Ecological Applications, 19(6), 2009, pp. 1546–1560 Ó 2009 by the Ecological Society of America Invasive Andropogon gayanus (gamba grass) is an ecosystem transformer of nitrogen relations in Australian savanna N. A. ROSSITER-RACHOR,1,2,3,6 S. A. SETTERFIELD,1,3 M. M. DOUGLAS,1,3,4 L. B. HUTLEY,1,3 G. D. COOK,2,3 5 AND S. SCHMIDT 1 School of Environmental and Life Sciences, Charles Darwin University, Darwin, Northern Territory 0909 Australia 2 CSIRO Tropical Ecosystems Research Centre, Darwin, Northern Territory 0822 Australia 3 Tropical Savannas Management Cooperative Research Centre, Charles Darwin University, Darwin, Northern Territory 0909 Australia 4 School for Environmental Research, Tropical Rivers and Coastal Knowledge (TRaCK) Research Hub, Charles Darwin University, Darwin, Northern Territory 0909 Australia 5 School of Integrative Biology, The University of Queensland, Brisbane, Queensland 4072 Australia Abstract. Invasion by the African grass Andropogon gayanus is drastically altering the understory structure of oligotrophic savannas in tropical Australia. We compared nitrogen (N) relations and phenology of A. gayanus and native grasses to examine the impact of invasion on N cycling and to determine possible reasons for invasiveness of A. gayanus. Andropogon gayanus produced up to 10 and four times more shoot phytomass and root biomass, with up to seven and 2.5 times greater shoot and root N pools than native grass understory. These pronounced differences in phytomass and N pools between A. gayanus and native grasses were associated with an altered N cycle. Most growth occurs in the wet season when, compared with native grasses, dominance of A. gayanus was associated with significantly lower total soil N pools, lower nitrification rates, up to three times lower soil nitrate availability, and up to three times higher soil ammonium availability. Uptake kinetics for different N sources were studied with excised roots of three grass species ex situ. Excised roots of A. gayanus had an over six times higher uptake rate of ammonium than roots of native grasses, while native grass Eriachne triseta had a three times higher uptake rate of nitrate than A. gayanus. We hypothesize that A. gayanus stimulates ammonification but inhibits nitrification, as was shown to occur in its native range in Africa, and that this modification of the soil N cycle is linked to the species’ preference for ammonium as an N source. This mechanism could result in altered soil N relations and could enhance the competitive superiority and persistence of A. gayanus in Australian savannas. Key words: ammonium; exotic grasses; invasive alien species; nitrate; nitrification inhibition; nitrogen cycling; nitrogen uptake. INTRODUCTION It is widely acknowledged that the composition of plant communities is a major determinate of nitrogen (N) cycling, affecting the amount of N stored in various pools in an ecosystem and/or the fluxes of N between these pools (Hooper and Vitousek 1998). In a range of ecosystems, invasive alien plants have altered the composition of resident plant communities, with profound effects on N cycling (see reviews in Ehrenfeld [2003], Levine et al. [2003], Corbin and D’Antonio [2004], D’Antonio and Hobbie [2005], and Liao et al. [2007b]). Invasion by alien plants can affect N cycling by altering the rates of N input (Vitousek and Walker 1989, Witkowski 1991, Stock et al. 1995), quality and quantity of litter (Lindsay and French 2004, Rothstein et al. Manuscript received 7 February 2008; revised 23 October 2008; accepted 27 October 2008; final version received 19 December 2008. Corresponding Editor: J. Gulledge. 6 E-mail: natalie.rossiter@cdu.edu.au 2004), rates of N uptake by plants (Windham and Ehrenfeld 2003), the soil microbial community associated with soil N transformations (Hawkes et al. 2005, Wolfe and Klironomos 2005), the microclimate in which microbially mediated processes occur (Mack and D’Antonio 2003), the rates of N losses via leaching, denitrification (D’Antonio and Hobbie 2005), and volatilization (Rossiter-Rachor et al. 2008). It is essential to determine the mechanisms underlying the impacts of an invader in order to understand the success of the invader (Levine et al. 2006). This information is also fundamental for identifying the likely short- and long-term impacts of invasion, and for providing a basis for management decisions for invaded ecosystems (Corbin and D’Antonio 2004). In the tropical savannas of northern Australia, one of the most significant alien plant invaders is the large African grass Andropogon gayanus (Kunth) (Whitehead and Wilson 2000) (see Plate 1). Andropogon gayanus was introduced into Australia as a pasture species in the 1930s (Oram 1987), but has since invaded beyond 1546 September 2009 GRASS INVASION ALTERS NITROGEN RELATIONS planted areas and into native vegetation (Flores et al. 2005). The rapid spread of A. gayanus across northern Australia has raised widespread concern (Whitehead and Wilson 2000, Russell-Smith et al. 2003, Flores et al. 2005) due to its profound impacts on savanna biodiversity (Brooks et al. 2009), fire regimes (Rossiter et al. 2003), and tree survival (Ferdinands et al. 2006). It has been suggested that the invasive success and persistence of A. gayanus in the savannas of tropical Australia may be partly due to ecophysiological and morphological advantages over native savanna grasses (Rossiter 2001, Clifton 2004). Compared with native grasses, A. gayanus has (1) higher photosynthetic and transpiration rates (Rossiter 2001); (2) higher soil water use (L. B. Hutley, S. A. Setterfield, M. M. Douglas, and N. A. Rossiter-Rachor, unpublished data); (3) longer growth period into the dry season (Rossiter et al. 2003); and (4) earlier onset of growth following early wet season storms (N. A. Rossiter-Rachor, personal observation). In Venezuelan savannas, A. gayanus also has a higher N uptake and nitrogen use efficiency (NUE) than native savanna grasses (Bilbao and Medina 1990). The remarkable drought resistance and growth properties of A. gayanus are well recognized in its native range of Western Africa, and have been attributed to its extensive root system, which accesses water and nutrients at the surface and deeper down the soil profile (Bowden 1964, Groot et al. 1998). In Australian and African savannas soil N levels are generally low, and N availability has been proposed as one of the major constraints on plant growth (Tothill et al. 1985, Solbrig et al. 1996). The apparent paradox of highly productive grasses thriving in low N ecosystems has been investigated in West African savanna (Abbadie and Lata 2006). Andropogon gayanus and several other species in the Andropogonaceae supertribe appear to have evolved a successful mechanism for conserving soil N by inhibiting nitrification (Lata et al. 2000, 2004). Roots of these grasses exude allelopathic compounds that inhibit the activity of nitrifying soil microbes by interrupting metabolic pathways (Subbarao et al. 2007), thus reducing the production of the highly mobile nitrate (NO3) (Lata et al. 2004). Suppression of nitrification and maintaining N as relatively immobile ammonium (NH 4þ) in soil therefore reduces the likelihood of N loss via denitrification and leaching (Subbarao et al. 2007) and may be a key mechanism for increasing the residence time of N within the soil–plant system (Abbadie and Lata 2006). This study compared sites invaded by A. gayanus with non-invaded sites to determine if changes in N relations are associated with invasion. Specifically we aimed to (1) quantify the changes in plant phytomass and N pools in savanna following invasion by A. gayanus, (2) investigate the impact of changes in grass phytomass on soil N relations by quantifying total and available soil N pools, and net ammonification and nitrification rates, and (3) determine how differences in N cycling relate to N 1547 source preferences of A. gayanus and native grasses by examining uptake kinetics of different N forms (NH4þ, NO3, amino acid glycine) in excised roots. METHODS Site description The study site was at Wildman Reserve, Northern Territory, Australia (12843 0 S, 131849 0 E). Air temperature is high throughout the year (mean 278C), while rainfall is highly seasonal (mean 1434 mm) and concentrated in the wet season (October–April). Soils are Kandosols, with a low N concentration (0.01–0.11% N; Day et al. 1979). This savanna is dominated by Eucalyptus miniata (Cunn. Ex Schauer) and E. tetrodonta (F. Muell), with a grass understory dominated by native perennial grasses Alloteropsis semialata (R. Br.) Hitchc. and Eriachne triseta Nees ex Steud. with patches of annual grasses, such as Pseudopogonatherum irritans (Br.) A. Camus. Andropogon gayanus has invaded extensive areas of the reserve over the last 20 years, forming dense, tall, almost monospecific swards up to 4 m high, and replacing the much shorter (;0.5 m) native grass communities (Fig. 1). Andropogon gayanus infestations on the reserve generally range from single plants to patches of 500 m2, within a matrix of uninvaded savanna (Rossiter et al. 2003). The savanna communities of Wildman Reserve are burned frequently (typically annually or biennially) as part of the Reserve’s fire management strategy. During this study, a controlled fuel-reduction burn was carried out by the Park Rangers, at the end of May 2003, after the plant and soil sampling had been carried out for that month (Table 1). This fire frequency is typical for this region, with up to 50% of the northern Australian savannas burnt annually (Russell-Smith et al. 2003). Experimental design Nitrogen relations in native grass savanna were compared with those occurring in A. gayanus invaded savanna using a randomized block design. Five paired plots (blocks) were studied, with each plot pair consisting of an area dominated by native grass (hereafter referred to as ‘‘native grass’’ plots), and a nearby (approximately 50 m distance) A. gayanus-dominated area (hereafter referred to as ‘‘A. gayanus plots’’). Plot pairs were located up to 600 m apart, and each plot was 50 3 50 m in size, with a canopy dominated by E. miniata and E. tetrodonta. In native grass plots, the grass component included several native grass species, while in A. gayanus plots only A. gayanus was present. The study was conducted over two wet-dry season cycles: October 2002–September 2003 (1614 mm) and October 2003– June 2004 (1785 mm) (Table 1). Average gravimetric soil moisture (0–30 cm) ranged from 13.3% (wet season) to 5% (dry season), average soil pH was 5.2 6 0.1 (mean 6 SE), and bulk density ranged from 1.46 g/cm3 at the soil surface, to 1.55 g/cm3 at 30 cm depth. These results were similar to those of Day et al. (1979) and there were no 1548 Ecological Applications Vol. 19, No. 6 N. A. ROSSITER-RACHOR ET AL. FIG. 1. Changes to vegetation structure due to Andropogon gayanus (gamba grass) invasion. Photos of (a) native grass savanna (dominated by Alloteropsis semialata and Eriachne triseta) at Wildman Reserve, in the early dry season (June) and (b) A. gayanusinvaded savanna in the early dry season (June) [photo (b) was taken 100 m to the northwest of the location of photo (a)]. significant differences in these soil characteristics between the two grass plot types. Grass phytomass and nitrogen pools To determine the above ground N pool of A. gayanus and native grasses, the mass and tissue N concentration (%N) of the grass phytomass were quantified. Leaf litter derived from woody savanna plant species was also collected at each sampling time. Samples were collected at key times over two years: (1) dry–wet season transition (October–November); (2) early wet season (December–January); (3) late wet season (February– March); (4) early dry season (May–June); and (5) late dry season (August–September) (Table 1); with the exception of January 2003, when logistics prevented phytomass sampling. Samples were collected from three random replicate quadrats, within each of the five paired native grass and A. gayanus plots (n ¼ total of 15 quadrats per grass plot type per sampling time). Aboveground grass phytomass was quantified by harvesting all aboveground grass material within three randomly placed 2 3 2 m quadrats per plot. This grass phytomass included all dead standing material and also grass material that had fallen to the ground and was disconnected from the plant. Leaf litter from woody plants was also collected from each quadrat. Grass TABLE 1. Summary of plant and soil sampling times throughout this study. 2002 Measurement Nov Plant Aboveground phytomass Belowground biomass x Soil Total N Inorganic N availability Resin bags Soil incubation Fire Fuel reduction burn Dec 2003 Jan Feb Mar x x x x Apr May x Jun Jul Aug Sep Oct Nov x x x x Dec x x x x Notes: The study was conducted over two wet–dry season cycles (October 2002–September 2003; October 2003–June 2004). Samples were collected at five key times of the wet–dry cycle: (1) dry–wet season transition (October–November); (2) early wet season (December–January); (3) late wet season (February–March); (4) early dry season (May–June); and (5) late dry season (August–September). The fuel reduction burn was carried out at the end of May 2003, after the plant and soil sampling had been completed. September 2009 GRASS INVASION ALTERS NITROGEN RELATIONS samples were returned to the lab and sorted into green leaves and stems (live biomass) and dead standing grass (necromass). All grass and leaf litter samples were dried for 48 h at 608C, weighed, ground, and analyzed for percentage of N on a Carlo Erba analyzer (Thermo Electron, Thermo Fisher Scientific, Waltham, Massachusetts, USA). Due to budgetary constraints, only samples from three of the five plot pairs, from the first year of sampling, were analyzed. Nitrogen pools in native grass and A. gayanus plots were calculated on a quadrat basis (product of plant mass and tissue N concentration). For each quadrat, four N pools were calculated: (1) live biomass, (2) necromass, (3) phytomass (sum of live biomass and necromass), and (4) woody leaf litter. Differences between grass and woody litter mass and grass and woody litter N pools between grass types were compared using a three-factor analysis of variance (ANOVA) with factors grass type (fixed), sampling time (fixed), and plot pair (random). Before statistical analyses were undertaken, assumptions of ANOVA were checked using Cochran’s test, and where necessary data were transformed prior to analyses to improve normality and homogeneity. All analyses were done using Statistica 5.5 (StatSoft, Tulsa, Oklahoma, USA). Root nitrogen pools Root biomass was sampled in March 2004 from three of the five plot pairs previously described, using a fixed volume PVC core-sampling device (7.65 3 30 cm; Table 1). In native grass plots, five cores were taken within each of three randomly located quadrats (n ¼ total of 15 cores per native grass plot, and a total of 45 for the study). In A. gayanus plots stratified sampling was undertaken, due to the clumped distribution of A. gayanus. Three cores were collected from the ‘‘tussock region’’ (through the centre of a plant) and three from the ‘‘inter-tussock region’’ (.30 cm from a plant center) TABLE 1. Extended. 2004 Jan x Feb Mar x x x x Apr May Jun x x x x 1549 within each of the three randomly located quadrats (n ¼ total of 18 cores for each A. gayanus plot, and a total of 54 for the study). Basal area of A. gayanus at ground level was measured and the location of each plant in the quadrat was drawn on a location map to determine the area of tussock and inter-tussock regions in each quadrat. The average tussock area was 63% of the 1-m2 quadrat. To determine root mass in each core, soil was air dried in aluminum trays, then sieved to 0.2 mm following Cornelissen et al. (2003). Dead roots, defined as darkened, limp, or deflated (Cornelissen et al. 2003), and roots that clearly belonged to trees or shrubs, were discarded. Root samples were dried at 408C for 48 h, weighed, ground, and analyzed for percentage of N on a Carlo Erba analyzer. In native grass plots, the dry mass of roots per quadrat was calculated using the average mass of roots from the five cores from the quadrat (over the 30 cm corer depth) and expressed as g/m2. In A. gayanus plots, root mass per quadrat was calculated using the average root mass from ‘‘tussock’’ and ‘‘intertussock’’ cores scaled by the proportional area occupied by each region within each quadrat. The root N pool in each quadrat (g N/m2) was calculated as the product of root biomass and root N concentration. Differences between root biomass and N pools between grass types were compared using a two-factor ANOVA with factors grass type (fixed) and plot pair (random). Total soil nitrogen pools Total soil N profile analyses were carried out at four sampling times: (1) wet season, January 2003; (2) dry season, May 2003; (3) wet season, January 2004; and (4) dry season, June 2004 (Table 1). Samples were taken from the top 30 cm of the soil profile, as this is where the majority of soil N is in savanna systems, and the N pool drops off sharply in the lower soil depths (Scholes and Walker 1993). Samples were collected from three random replicate quadrats, within each of the five native grass and A. gayanus plot pairs (n ¼ total of 15 quadrats per grass plot type, per sampling time). A pit was dug (;30 3 30 cm wide 3 40 cm deep) and soil cores were collected from three different depths (0–5 cm, 5–10 cm, 20–30 cm) by inserting a 2 3 10 cm soil corer horizontally into the wall of the pit. Approximately 200 g of soil was taken at each depth (n ¼ 3 soil cores for each depth, per native grass and A. gayanus plot). The sample was placed in a snap-lock polyethylene bag and stored on ice until return to the lab, where the soil was sieved, and approximately 100 g of soil was taken for determination of gravimetric moisture determination (weighed, dried at 1058C for 48, and reweighed; results not presented). The remainder of the soil sample was dried at 408C for 48 h, ground, and then analyzed for total N (% N) on a Carlo Erba analyzer. Total soil N pools for each depth (0–5 cm, 5–10 cm, 20–30 cm) were calculated as the product of soil N concentration and soil bulk density of soil for that depth, and were expressed as g/m2. Differences between total soil N 1550 Ecological Applications Vol. 19, No. 6 N. A. ROSSITER-RACHOR ET AL. PLATE 1. (Top) Dense infestation of Andropogon gayanus (gamba grass) near Adelaide River, Northern Territory, Australia. A. gayanus invasion leads to a near monoculture of the understorey at this site. (Bottom) Savanna tree death due to repeated high-intensity Andropogon gayanus (gamba grass) wildfires (Northern Territory, Australia). A. gayanus fires led to a 53% reduction in tree cover in just 12 years (Ferdinands et al. 2006) at this site. Photo credits: S. A. Setterfield. pools, between grass types were compared for (1) January 2003, (2) May 2003, (3) January 2004, and (4) June 2004, using a four-factor ANOVA with factors grass type (fixed), plot pair (random), depth (fixed), and core (nested in grass type, plot pair, time, and depth). Soil inorganic N availability Two in situ techniques were used to examine the effect of A. gayanus on soil inorganic N (NH4þ and NO3) relations: (1) ion-exchange resin bags and (2) whole-soil incubations, which measure the availability of soluble N ions in the presence and absence of live roots, respectively. Ion exchange resin bags provide an index of plant N availability, as ion accumulation on the resin bags depends on the rates of mineralization, the ion form, water movement in the soil, and plant and microbial uptake; the same factors that determine N availability for plants (Binkley 1984). Alternatively, the whole soil incubation also allows for inorganic N availability and net ammonification and nitrification rates to be quantified in root zone soil (Hart and Firestone 1989). Soil inorganic N availability in the presence of roots was measured using mixed ion-exchange resin bags (Dowex-MR3; Sigma, St. Louis, Missouri, USA) following Schmidt et al. (1998). Measurements were taken over two years at the same five seasons and plot pairs, previously described for the grass phytomass sampling (Table 1). Resin bags were buried at a 458C angle, 5 cm below the soil surface and incubated in situ for 7–14 days. After incubation, the resin was extracted with 1 mol/L KCl, the extract was analyzed NH4þ and NO3 using a Lachat flow injection autoanalyzer (Lachat Instruments, Loveland, Colorado, USA). Inorganic N availability (NH4þ and NO3) over the incubation period (ng Ng1 resind1) was calculated. Differences between inorganic N availability between grass types were compared using a three-factor ANOVA with factors grass type (fixed), sampling time (fixed), and plot pair (random). Post-hoc Tukey’s tests were used to make comparisons between treatment means. Soil inorganic N availability in the absence of roots was determined in the wet season (March 2004) using buried-bag in situ incubations (following Hart and Firestone 1989) (Table 1). This method involved incubating soil inside a gas-permeable, water impermeable polyethylene bag (Hart and Firestone 1989). Measurements were taken within three random replicate quadrats, at three of the five native grass and A. gayanus plot pairs (n ¼ total of 9 quadrats for each grass plot type). At t ¼ 0, two intact 10 3 7.5 cm soil cores were taken from each quadrat. One core was sealed in a polyethylene bag, placed on ice, and returned to the laboratory, where it was sieved to 2 mm to remove large roots. A subsample of the soil was dried to determine gravimetric water content, and the remainder was analyzed for NH4þ and NO3 after extraction with 2 mol/L KCl. The second core taken from each quadrat was carefully removed from the soil with minimal disturbance, tightly sealed in a thick (1.5 mm) polyethylene bag, and immediately returned to the same hole from which the soil core was removed. After a 28-day incubation period in the field, the incubated core was removed and processed as per the initial core. The results were used to calculate the rates of net ammonification and nitrification over the incubation period (calculated as the difference between the pool size in the initial and incubated soil cores). Differences between net ammonification and nitrification rates between grass types were compared using a two-factor ANOVA with factors grass type (fixed) and plot pair (random). Root 15 15 N uptake N uptake by excised roots was measured in two native grasses (A. semialata and E. triseta) and A. gayanus. Fine roots were sampled by carefully excavating three grass tussocks of each grass species with a shovel and transferring tussocks with attached roots and September 2009 GRASS INVASION ALTERS NITROGEN RELATIONS 1551 FIG. 2. Grass (a) live biomass (green leaves and stems), (b) standing necromass (dead leaves and stems), and (c) phytomass (live biomass þ necromass); and (d) woody litter, in native grass and A. gayanus plots at Wildman Reserve (Northern Territory, Australia) from November 2002 to June 2004. Values are means 6 SE. A controlled fuel-reduction burn was carried out at the study site in late May 2003, after that month’s sampling had been completed. soil into plastic bags, ensuring that roots from individual plants were kept separate. Samples were stored on ice, transported to the laboratory within 2 hours, and processed immediately. Fine roots were washed, cut into 2-cm lengths, and transferred into 25 mL of 15Nlabelled (98–99 atom% excess) solution. Roots were incubated in one of three different N sources (NH4þ, NO3, or glycine [GLY]) at one of five N concentrations (1, 10, 100, 300, and 1000 lmol/L N; each with 100 lmol/L CaSO4 added to maintain membrane integrity). Root samples were incubated for 30 minutes at 308C in an agitating (100 rpm) water bath. After the 15N incubation, roots were shaken for 10 minutes in 10 mmol/L KCl to remove 15N from the apoplast, rinsed with deionized water, and dried at 508C for 24 h. Samples were subsequently ground to a fine powder and analyzed for 15N using a continuous flow isotope mass spectrometer (CFIRMS, Micromass Isochrom, Manchester, UK). Differences in N uptake kinetics between the three grass species were compared by determining Vmax (maximum uptake rate over 30 minutes) and Km (substrate concentration at 50% maximum uptake rate). Vmax and Km values were calculated by nonlinear curve fitting of the experimental data to the Michaelis-Menten equation: V ¼ Vmax(S/[Km þ S]), where Vmax is the maximum uptake rate (lmolg130 min1); Km is the Michaelis-Menten affinity constant describing the sub- strate concentration when 50% of maximum uptake occurs (lmol/L); and S is the substrate concentration (lmol/L). RESULTS Aboveground plant material Grass live biomass, necromass, and overall grass phytomass, were significantly higher in A. gayanus plots than in native grass plots (Fig. 2, Table 2a). Mean grass phytomass in A. gayanus plots were more than 10 times greater than that in native grass plots (May 2003; means 502 vs. 44.4 g/m2, respectively, Fig. 2c). Andropogon gayanus and native grasses had significantly different live biomass, necromass and phytomass over time (Fig. 2, Table 2a) and these temporal patterns differed between grass types (Table 2a), particularly for live biomass. Live biomass of A. gayanus and native grasses increased significantly as the wet season progressed. However, native grasses generally reached peak live biomass in the late wet season (March), while A. gayanus reached peak live biomass several months later in the early dry season (May/June) at a time when the native grasses had already senesced, demonstrating that A. gayanus has a much longer growing season than the native grasses (Fig. 2a). The patterns in live biomass production showed small scale spatial variation, as 1552 Ecological Applications Vol. 19, No. 6 N. A. ROSSITER-RACHOR ET AL. TABLE 2. Summary of significant results from a three-factor ANOVA on grass and woody litter mass (g) and N pools (g N/m2), a four-factor ANOVA on total soil N pools (g N/m2), and a three-factor ANOVA on inorganic soil N (NH4þ and NO3) availability (ng N[g resin] 1d1). Factor df Live biomass Necromass Phytomass a) Mass Grass type Time Plot pair Grass type 3 time Grass type 3 plot pair Time 3 plot pair Grass type 3 time 3 plot pair 1, 8, 4, 8, 4, 32, 32, 4 32 180 32 180 180 180 47.06** 30.03*** 38.83*** 17.13*** 42.18** 17.91*** 19.60*** 7.67*** 2.48*** 2.92*** 8.30*** 2.58*** 2.58*** 3.16*** 10.18*** 8.11*** 2.79*** 3.38*** b) N pools Grass type Time Grass type 3 time Grass type 3 plot pair Time 3 plot pair Grass type 3 time 3 plot pair 1, 4, 4, 2, 8, 8, 2 8 8 60 60 60 22.52* 32.21*** 4.94* 5.89** 2.98** 7.41*** c) Total soil N pools January 2003 Grass type June 2003 Grass type Depth January 2004 Grass type Depth Grass type 3 depth June 2004 Depth d) Inorganic soil N availability Grass type Time Grass type 3 time Time 3 plot pair 28.06* 19.89 68.19* 28.02*** Woody litter Total N NO3 39.07*** 3.64** 3.44** 61.13*** 3.27* 3.58** 4.46*** 1, 4 7.87* 1, 4 2, 8 22.46*** 10.81** 1, 4 2, 8 2, 8 45.48** 56.74*** 6.69* 2, 8 33.85*** 1, 7, 7, 28, NH4þ 4 28 28 160 7.15*** 3.15* 1.96** 46.07** 9.91*** 9.37*** 1.66* Note: Values in the table are F values, and significant results are indicated by asterisks. * P , 0.05; ** P , 0.01; *** P , 0.001. indicated by the significant interactions between plot pair and all other factors (Table 2a). However, there was no overall effect of plot pair for any of the grass mass components and the mean squares for their interaction terms were smaller than for all other main effects and the species by time interaction. Trends were the same at all plot pairs, but on a few occasions the difference between native grass and A. gayanus plots were more pronounced at one plot pair than at another. Leaf litter from woody plant species did not vary significantly between grass type but varied significantly over time, with more woody litter in both grass communities at the start of the wet season, decreasing over the wet season, and then increasing again in the dry season (Fig. 2d, Table 2a). In addition to the significant differences between grass type and time, there was significant spatial variation in woody litter at the scale of plot pairs (Table 2a). Aboveground plant N pools The significantly higher grass live biomass, necromass, and total phytomass, in combination with higher tissue N concentrations of A. gayanus (Appendix A), resulted in significantly higher N pools in A. gayanus plots compared with native grass plots (Fig. 3, Table 2b). N pools of A. gayanus phytomass were up to seven times greater than native grass phytomass N pools (March 2003, 1.61 vs. 0.23 g N/m2 for A. gayanus and native grass plots respectively, Fig. 3c). As expected, all grass N pools (live biomass, necromass and phytomass) varied significantly with time (Fig. 3, Table 2b). However the effect of grass type on live biomass N pools differed between sampling times (Table 2b) because the grass N pool for both grass types was zero for several months after fire. Patterns in grass N pools showed small scale spatial variation, as indicated by the significant interactions between plot pair and all other factors and the significant three-way interaction with grass type, plot pair, and time (Table 2b). Trends were the same at all plot pairs, although in some instances the difference between native grass and A. gayanus plots were more pronounced at one plot pair than at another. The N pool of leaf litter from woody species did not vary significantly between plots for the two grass types September 2009 GRASS INVASION ALTERS NITROGEN RELATIONS 1553 FIG. 3. Grass (a) live biomass (green leaves and stems) N pool, (b) standing necromass (dead leaves and stems) N pool, and (c) phytomass (live biomass þ necromass) N pool; and (d) woody litter N pool, in native grass and A. gayanus plots at Wildman Reserve (Northern Territory, Australia) from November 2002 to September 2003. Values are means 6 SE. A controlled fuelreduction burn was carried out at the study site in late May 2003, after that month’s sampling had been completed. (Fig. 3d), but varied significantly over time (Table 2b). There was a higher N pool in both grass types at the start of the wet season, which decreased over the wet season, and increased again in the following dry season (Fig. 3d). Although all woody leaf litter was consumed in the fire (May 2003), the woody litter N pool quickly returned to almost pre-fire levels four months after the fire in September 2003 (Fig. 3d). Belowground root biomass and N pool Root biomass (0–30 cm) was four times greater in A. gayanus plots, compared to the native grass plots (267.3 6 36.0 vs. 64.9 6 9.0 g/m2 [mean 6 SE], respectively; F1,2 ¼ 44.77, P , 0.05). However, root N concentrations (% N) did not differ significantly between native grass and A. gayanus plot pairs (data not presented), ranging from 0.16% to 0.34% N in the native grass and 0.01% to 0.20% N in A. gayanus. Due to biomass differences, the root N pool was significantly higher in A. gayanus plots (0.41 6 0.05 g root N/m2) compared to the native grass plots (0.16 6 0.03 root N/m2), a difference of 2.5 times (F1,2 ¼ 28.37, P , 0.05). Total soil N pools Total soil N pools in A. gayanus plots were significantly lower than those in native grass plots in the wet season (Fig. 4, Table 2c), and this difference was more pronounced with increasing soil depth. Over two wet seasons (January 2003, January 2004) the total soil N pools in A. gayanus plots were 13–28% lower (0–5 cm depth), 18–51% lower (5–10 cm depth), and 25–78% lower (20–30 cm depth) than in the native grass plots (Fig. 4, Table 2c). For example, in January 2004, the mean total soil N at 20–30 cm depth was 93.4 6 4.2 vs. 20.9 6 2.2 g N/m2 in native grass and A. gayanus plots, respectively. However in the dry season, the total soil N pool of A. gayanus plots was either significantly higher (May 2004) or similar (June 2004) to the total N pool of the native grass plots (Fig. 4, Table 2c). Total soil N pools also varied significantly with depth in both grass types (Fig. 4, Table 2c). However, in January 2004 the effect of depth was not the same in both grass types, resulting in a significant interaction between grass type and depth, with a much larger change in total soil N pools at 20–30 cm depth for the native grass plots (Fig. 4, Table 2c). 1554 N. A. ROSSITER-RACHOR ET AL. Ecological Applications Vol. 19, No. 6 Soil inorganic N availability Availability of NH4þ and NO3, as measured with in situ ion exchange resin bags, varied significantly over time in native grass and A. gayanus plots (Fig. 5, Table 2d). Although the general pattern of variation of NH4þ and NO3 availability was similar in both grass types, dissimilarity at some sampling times resulted in a grass type 3 time interaction (Fig. 5, Table 2d). Post-hoc tests revealed that NH4þ availability was significantly (up to three times) higher in A. gayanus plots at several times in the wet season (January, March, and November 2003; March 2004; Fig. 5b). For example, in March 2003, the mean NH4þ availability was 1643 6 444 vs. 500 6 109 ng NH4þ[g resin]1d1 in A. gayanus and native grass plots respectively. In contrast, NO3 availability was significantly (up to three times) lower in A. gayanus plots compared to native grass plots at several times in the wet season (November 2002, January and March 2003, November 2003; Fig. 5c). For example, in November 2002, the mean NO3 availability was 226 6 69 vs. 727 6 155 ng NO3[g resin]1d1, in A. gayanus and native grass plots, respectively. Although all plot pairs exhibited the same general trend of NH4þ and NO3 availability, some had a greater magnitude of difference than others at some times, resulting in a plot pair 3 time interaction (Table 2d). Net ammonification rates as measured with in situ soil incubations did not vary significantly with grass type or plot pair (means 0.67 vs. 0.73 mg NH4þ[kg soil]1[28 d]1 in A. gayanus and native grass, respectively). In contrast, net nitrification rates were significantly (53%) lower in A. gayanus plots, compared to those in native grass plots (means 0.23 vs. 0.50 mg NO3[kg soil]1[28 d]1, respectively; F1,2 ¼ 50.70, P , 0.05. N uptake by excised roots FIG. 4. Mean total soil N in native grass and A. gayanus plots at Wildman Reserve (Northern Territory, Australia) over two wet–dry season cycles (January 2003 and May 2003; January 2004 and June 2004). Data are means 6 SE. A controlled fuel-reduction burn was carried out at the study site in late May 2003, after that month’s sampling had been completed. At concentrations ranging from 1 to 1000 lmol/L N, excised roots of A. gayanus and native grass A. semialata exhibited a similar order of preference for uptake of the three N sources with NH4þ . GLY . NO3, whereas native grass E. triseta incorporated N sources in the order of preference NH4þ . NO3 . GLY (Fig. 6). Compared with the native grasses, A. gayanus had greater incorporation rates of each of the three N sources, but this was most pronounced with NH4þ (Fig. 6). The calculated maximum uptake rate (Vmax) for NH4þ of A. gayanus was six and 7.5 times higher than Vmax of E. triseta and A. semialata, respectively (Fig. 6). In contrast to NH4þ, E. triseta had the highest Vmax for NO3, with approximately three times higher Vmax than A. gayanus and A. semialata (Fig. 6). Uptake of GLY by A. gayanus did not follow Michaelis-Menten kinetics, but was linear in the experimental conditions, while the native grasses did exhibit Michaelis-Menten kinetics for GLY uptake (Fig. 6). Substrate affinity (Km value, substrate concentration at which 50% of the maximum uptake rate is reached) for NH4þ was three and ten times September 2009 GRASS INVASION ALTERS NITROGEN RELATIONS 1555 FIG. 5. (a) Monthly rainfall (mm) over the study period, and availability of (b) NH4þ and (c) NO3 (as determined with in situ ion exchange resin bags) in native grass and A. gayanus plots, at Wildman Reserve (Northern Territory, Australia) from November 2002 to March 2004. Values are means 6 SE. Asterisks indicate significant differences among grass types (P , 0.05). See Table 2d for statistics. A controlled fuel-reduction burn was carried out at the study site in late May 2003, after that month’s sampling had been completed. greater in A. gayanus than in A. semialata and E. triseta, respectively (Fig. 6). Alloteropsis semialata had a greater affinity for NO3, with over five times higher Km values than A. gayanus and E. triseta, and both the native grasses had a similar affinity for GLY. DISCUSSION This study demonstrates that invasion of low-nutrient Australian savanna by Andropogon gayanus alters the composition of the community understory and the above- and belowground N pools (Fig. 7). Associated 1556 N. A. ROSSITER-RACHOR ET AL. FIG. 6. Uptake of different N forms by excised roots of A. gayanus and native grass species E. triseta and A. semialata. Roots were incubated for 30 minutes in 1, 10, 100, 300, and 1000 lmol/L 15N-labeled (98–99 atom% enriched) ammonium, glycine, and nitrate. Data are averages of three replicates (6SD). Vmax (maximum uptake rate over 30 minutes) and Km (substrate concentration at 50% maximum uptake rate) were calculated from uptake curves. Uptake of glycine by A. gayanus did not follow Michaelis-Menten kinetics but was instead linear in the studied concentration range. Ecological Applications Vol. 19, No. 6 with the large N pools in the phytomass of A. gayanusdominated understory were marked changes in soil N relations. In a growing season, A. gayanus produced an order of magnitude more phytomass than native grasses, resulting in seven and 2.5 times greater shoot and root N pools. The higher growth rate and live biomass accumulation of A. gayanus is supported by up to four times greater root biomass per volume of soil, greater rooting depth, and more efficient uptake of N from soil compared with native grasses. Roots of A. gayanus had a 52 times higher maximum uptake rate (Vmax) for ammonium than for nitrate, and six to 7.5 times greater Vmax for ammonium than native grasses. The kinetics suggest that A. gayanus possesses a low affinity/high capacity uptake system for ammonium which would allow efficient uptake from mmol/L concentrations, although a high affinity/low capacity uptake system for lmol/L concentrations may also be present, since plants generally have high and low affinity transport systems (Loqué and von Wirén 2004). Of the studied grasses, A. gayanus had the most pronounced difference in ammonium and nitrate uptake, with only high affinity/low capacity uptake of nitrate. Andropogon gayanus also had lower nitrification rates in its root zone compared to native grasses. Taken together, these findings indicate that A. gayanus has a superior ability to acquire ammonium from high concentrations, a low preference for nitrate, and may inhibit nitrification in its root zone. Nitrification inhibition has been demonstrated for A. gayanus and other African grasses (Subbarao et al. 2007). Savanna trees have a similar order of preference for N sources as A. gayanus (NH4þ . glycine . NO3) (Schmidt and Stewart 1999) suggesting that ammonium is the main N source for savanna plants, followed by organic N (amino acid-N). Ammonium was taken up preferentially by the native grasses, but nitrate was preferred over organic N (glycine) by E. triseta, while nitrate uptake was lower than organic N in A. semialata but higher than in A. gayanus. The greater capacity for ammonium uptake of A. gayanus could affect native savanna plant species via direct competition for ammonium and/or reduced availability of nitrate. Wedin and Tilman (1990) showed that individual grass species can influence soil N availability within just a few years, and they suggested that competition for N may lead to positive feedbacks between the processes controlling species composition. The significantly greater ammonium availability in A. gayanus plots indicates that savanna N relations were altered by the presence of the invasive grass. In the presence of grass roots, soil ammonium availability was three times higher in soil associated with A. gayanus invasion compared to soil associated with native grasses. In support of the notion that the presence of A. gayanus changes soil N relations, similar ammonium availability was detected in soil of all grasses in the absence of roots. One possible explanation for the discrepancy between September 2009 GRASS INVASION ALTERS NITROGEN RELATIONS 1557 FIG. 7. Model of the effect of A. gayanus invasion on ecosystem nitrogen cycling in a savanna woodland (after Liao et al. 2007a). Symbols denote an decrease in response to A. gayanus invasion (#), an increase in response to A. gayanus invasion ("), or no change in response to A. gayanus invasion (¼).Question marks (‘‘?’’) denote processes that were not directly measured here. Data are from Rossiter (2001). soil ammonium relations measured in the presence or absence of roots is differences in microbial activity. Microbial activity may be stimulated in the presence of A. gayanus roots and result in increased rates of mineralization and ammonification. Further research has to substantiate this hypothesis, as little is known about soil microbial communities in Australian savanna soils, or the potential impact of A. gayanus invasion on microbial communities. However, both techniques, in situ resin and whole soil incubations, detected lower nitrification rates in soil associated with A. gayanus when compared to rates from native grass soils. The lower nitrification rates in soil associated with A. gayanus, in combination with the greater availability of ammonium, supports the notion that the presence of A. gayanus inhibits nitrification in soil. This concurs with observations from Columbian savanna (SylvesterBradley et al. 1988) where after six weeks incubation soil nitrate levels had increased 6.5 times in control plots but only 1.2-times in A. gayanus plots. Furthermore, recent research by J. C. Lata, P. Jouquet, X. Raynaud, R. Lensi, L. Abbadie, and S. Barot (unpublished manuscript) has documented that in its native range in Western Africa A. gayanus reduces the nitrification potential of soil by 78%. This is also consistent with previous studies (Lensi et al. 1992, Lata et al. 2004) demonstrating that the African grass Hyparrhenia diplandra decreased the nitrification potential of African savanna soils. A transplant experiment by Lata et al. (2000) showed that the presence of H. diplandra altered nitrification by inhibiting the activity of nitrifying bacteria in the soil, and we propose a similar mechanism for A. gayanus in Australian savannas. Allelochemicals inhibiting nitrification have been studied intensively (reviewed by Subbarao et al. 2006, 2007) and numerous compounds inhibit growth of nitrifying bacteria, although few studies have investigated these compounds in situ and associated with plant roots. It was recently shown that African grass Brachiaria humidicola exudes as yet unidentified compounds which inhibit function of nitrifying microbe Nitrosomonas europea by blocking the enzymes converting ammonium to nitrite (Subbarao et al. 2007). While it has long been hypothesized that grasses have mechanisms for inhibiting nitrification (Theron 1951, Munro 1966, Meiklejohn 1968), it was only recently demonstrated that grasses, including Melinis minutiflora, Megathyrus maximus (formerly Pannicum maximum), Brachiaria decumbens and A. gayanus exude compounds that inhibit biological nitrification (Subbarao et al. 2007). The effect of these so-called ‘‘biological nitrification inhibition (BNI) compounds’’ released from A. gayanus roots was considerable when compared with artificial nitrification inhibitor allylthiourea (AT), since A. gayanus roots produced 7.7 AT units/g root dry mass (Subbarao et al. 2007). 1558 Ecological Applications Vol. 19, No. 6 N. A. ROSSITER-RACHOR ET AL. In the current study, the potential advantage for A. gayanus provided by suppressed nitrification is that N loss from an invaded site could be markedly reduced. Soil moisture dynamics at our study site suggest that given high intensity rainfall associated with the north Australian monsoon, drainage from the root zone in the top 1 m of soil is significant and in the order of 200–300 mm per annum (L. B. Hutley, S. A. Setterfield, M. M. Douglas, and N. A. Rossiter-Rachor, unpublished data). Given that nitrate availability in these savannas is up to 16-fold greater than ammonium (Schmidt et al. 1998), it is likely that nitrate leaches from soil in the wet season. Consequently, inhibition of nitrification and rapid uptake of ammonium would provide an important competitive advantage to A. gayanus in this ecosystem, complementing previously documented traits relating to higher photosynthetic and transpiration rates, higher soil water use, and longer growing period, by maximizing N retention and minimizing N loss in this low N system. In addition to the pronounced changes in available N pools, A. gayanus invasion led to significant decreases in total soil N pools in the wet season (the growing season). Decreases in total soil N have been reported for a range of alien grass invasions (Johnson and Wedin 1997, Christian and Wilson 1999, Ibarra-Flores et al. 1999, Kourtev et al. 1999, Reed et al. 2005, Sperry et al. 2006, Drenovsky and Batten 2007). For example, total soil N pools were 40% lower in Bromus tectorum invaded communities in southeastern Utah, USA (Sperry et al. 2006). However in some studies the reduction in total soil N due to alien grass invasion varied spatially (Ibarra-Flores et al. 1999, Kourtev et al. 1999), seasonally (Mack and D’Antonio 2003, this study), or following disturbance events such as fire (Reed et al. 2005). The reduction in total soil N following alien plant invasion can be due to their effects on N uptake and N transformation rates, or alterations in disturbance regimes (Corbin and D’Antonio 2004). In the current study the large decreases in total soil N in the wet season are most likely to be caused by increases in N uptake due to A. gayanus’ higher root and shoot production, deeper rooting depth and longer growing season. Burning of A. gayanus phytomass in the frequent savanna fires could also lead to increases in N uptake by A. gayanus to support its rapid post-fire growth and reestablishment. In this study, the high levels of A. gayanus live biomass had almost completely regrown just one year after complete combustion, and this was accompanied by large reductions in total soil N throughout the entire soil profile. Regular burning of grass phytomass would also result in significant increases in N losses from the ecosystem via volatilization of the large pool of N stored in the aboveground biomass (Rossiter-Rachor et al. 2008), with possible implications for total soil N pools. In the longer term, A. gayanus fires could have further effects on the total soil N pool by decreasing tree density, due to increases in fire intensity (Rossiter et al. 2003). Although it was originally thought that it may take several decades for significant decreases in savanna tree cover to occur, a recent study showed that repeated high intensity A. gayanus wildfires led to a 53% reduction in tree cover in just 12 years (Ferdinands et al. 2006). High-intensity A. gayanus fires and the subsequent reduction in tree cover could lead to further reductions in total soil N, due to the loss of ‘‘islands of fertility’’ associated with trees (reviewed by Schmidt and Lamble 2002). One study, in the Australian savannas, estimated that a 20% reduction in tree density could result in a loss of up to 21 000 kg of N (21 metric tons) from a 10-km2 area (Ludwig et al. 2000). This could be an important additional effect of A. gayanus on savanna N cycling, and should be examined in future studies. In summary, we have demonstrated that A. gayanus significantly alters plant and soil N relations in the mesic savannas of Northern Australia, providing a clear example of how a single species may alter ecosystem N dynamics. We hypothesize that A. gayanus inhibits nitrification, as it does in its native range in Africa. This mechanism could play a role in the invasive success and persistence of A. gayanus in the low N savannas, in addition to the previously described ecophysiological and morphological traits, as it could allow A. gayanus to regulate soil N relations in the invaded system and outcompete native species for N. These changes in N relations, combined with the 8-times higher fire intensities following A. gayanus invasion (Rossiter et al. 2003) and the resultant loss of woody cover (Ferdinands et al. 2006), clearly make A. gayanus an ecosystem transformer of the mesic savannas of northern Australia. ACKNOWLEDGMENTS We thank the Parks and Wildlife Commission of the Northern Territory for access to Wildman Reserve, and the rangers at Wildman Reserve. K. McGuinness (Charles Darwin University) and G. Quinn (Deakin University) provided much appreciated advice on design and analysis. We thank J. Barratt and numerous volunteers for field and lab assistance. We thank J.-C. Lata and G. V. Subbarao for helpful discussions on the mechanism of nitrification inhibition by African grasses. Earlier versions of the manuscript were improved through valuable comments from C. D’Antonio, D. Richardson, T. Grice, and two anonymous reviewers. This research was supported by Doctoral Research Scholarships from Charles Darwin University, CSIRO Sustainable Ecosystems, and the Tropical Savannas Management CRC. LITERATURE CITED Abbadie, L., and J. C. Lata. 2006. Nitrogen dynamics in the soil–plant system. Pages 277–298 in L. Abbadie, J. Gignoux, X. Le Roux, and M. Lepage, editors. Lamto: structure, functioning and dynamics of a savanna ecosystem. Springer, New York, New York, USA. Bilbao, B., and E. Medina. 1990. Nitrogen use efficiency for growth in a cultivated African grass and a native South American pasture grass. Journal of Biogeography 17:421–425. Binkley, D. 1984. Ion exchange resin bags: factors affecting estimates of N availability. Soil Science Society of America Journal 48:1181–1184. Bowden, B. N. 1964. Studies of Andropogon gayanus Kunth. III: an outline of its biology. Journal of Ecology 52:255–271. September 2009 GRASS INVASION ALTERS NITROGEN RELATIONS Brooks, K., S. A. Setterfield, and M. M. Douglas. 2009. Exotic grass invasions: applying a conceptual framework to the dynamics of degradation and restoration in Australia’s tropical savannas. Restoration Ecology, in press. Christian, J. M., and S. D. Wilson. 1999. Long-term impacts of introduced grass in the Northern Great Plains. Ecology 80: 2397–2407. Clifton, P. 2004. Effects of Andropogon gayanus on microclimate and woody seedling recruitment. Thesis. Charles Darwin University, Darwin, Australia. Corbin, J. D., and C. M. D’Antonio. 2004. Effects of exotic species on soil nitrogen cycling: implications for restoration. Weed Technology 18:1464–1467. Cornelissen, J. H. C., S. Lavorel, E. Garnier, S. Diaz, N. Buchmann, D. E. Gurvich, P. B. Reich, H. ter Steege, H. D. Morgan, M. G. A. van der Heijden, J. G. Pausas, and H. Poorter. 2003. A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Australian Journal of Botany 51:335–380. D’Antonio, C. M., and S. E. Hobbie. 2005. Plant species effects on ecosystem processes: insights from invasive species. Pages 65–84 in D. F. Sax and J. J. Stachowicz, editors. Species invasions. Freeman, New York, New York, USA. Day, K. J., C. J. Harrison, and H. R. M. van Cuylenburg. 1979. Land resources of Wildman River Station, N.T. Northern Territory Parks and Wildlife Commission, Darwin, Australia. Drenovsky, R. E., and K. M. Batten. 2007. Invasion by Aegilops tricuncialis (barb goatgrass) slows carbon and nutrient cycling in a serpentine grassland. Biological Invasions 9:107–116. Ehrenfeld, J. G. 2003. Effects of exotic plant invasions on soil nutrient cycling processes. Ecosystems 6:503–523. Ferdinands, K., S. Setterfield, M. Douglas, and J. Barratt. 2006. Africanising the tropical woodlands: canopy loss and tree death following gamba grass (Andropogon gayanus) invasion. Page 95 in C. Preston, J. H. Watts, and N. D. Crossman, Proceedings of the 15th Australian Weeds Conference. Weed Management Society of South Australia, Adelaide, Australia. Flores, T. A., S. A. Setterfield, and M. M. Douglas. 2005. Seedling recruitment of the exotic grass Andropogon gayanus in northern Australia. Australian Journal of Botany 53:1–7. Groot, J. J. R., M. Traoré, and D. Koné. 1998. Description du système racinaire de trois espèces. Fourragères en zone soudano-sahélienne Biotechnology, Agronomy, Society and Environment 2:106–119. Hart, S. C., and M. K. Firestone. 1989. Evaluation of three in situ soil nitrogen availability assays. Canadian Journal of Forest Research 19:185–191. Hawkes, C. V., I. F. Wren, D. J. Herman, and M. K. Firestone. 2005. Plant invasion alters nitrogen cycling by modifying the soil nitrifying community. Ecology Letters 8:976–985. Hooper, D. U., and P. M. Vitousek. 1998. Effects of plant composition and diversity on nutrient cycling. Ecological Monographs 68:121–149. Ibarra-Flores, F., J. R. Cox, M. Martin-Riveria, T. A. Crowl, B. E. Norton, R. E. Banner, and R. W. Miller. 1999. Soil physicochemical changes following Buffelgrass establishment in Mexico. Arid Soil Research and Rehabilitation 13:39–52. Johnson, N. C., and D. A. Wedin. 1997. Soil carbon, nutrients, and mycorrhizae during conversion of dry tropical forest to grassland. Ecological Applications 7:171–182. Kourtev, P. S., W. Z. Huang, and J. G. Ehrenfeld. 1999. Differences in earthworm densities and nitrogen dynamics in soils under exotic and native plant species. Biological Invasions 1:237–245. Lata, J. C., V. Degrange, X. Raynaud, P. A. Maron, R. Lensi, and L. Abbadie. 2004. Grass populations control nitrification in savanna soils. Functional Ecology 18:605–611. 1559 Lata, J. C., K. Guillaume, V. Degrange, L. Abbadie, and R. Lensi. 2000. Relationships between root density of the African grass Hyparrhenia diplandra and nitrification at the decimetric scale: an inhibition–stimulation balance hypothesis. Proceedings of the Royal Society B 267:595–600. Lensi, R., A. M. Domenach, and L. Abbadie. 1992. Field study of nitrification and dentrification in a wet savanna of west Africa (Lamto, Côte d’Ivoire). Plant and Soil 147:107–113. Levine, J. A., M. Villa, C. M. D’Antonio, J. S. Dukes, K. Grigulis, and S. Lavorel. 2003. Mechanisms underlying the impacts of exotic plant invasions. Proceedings of the Royal Society of London 270:775–781. Levine, J. M., E. Pachepsky, B. E. Kendall, S. G. Yelenik, and J. H. R. Lambers. 2006. Plant–soil feedbacks and invasive spread. Ecology Letters 9:1005–1014. Liao, C., Y. Luo, L. Jiang, X. Zhou, X. Wu, C. Fang, J. Chen, and B. Li. 2007a. Invasion of Spartina alterniflora enhanced ecosystem carbon and nitrogen stocks in the Yangtze Estuary, China. Ecosystems 10:1351–1361. Liao, C., R. Peng, Y. Luo, X. Zhou, X. Wu, C. Fang, J. Chen, and B. Li. 2007b. Altered ecosystem carbon and nitrogen cycles by plant invasion: a meta-analysis. New Phytologist 177:706–714. Lindsay, E. A., and K. French. 2004. Chrysanthemoides monilifera ssp. rotundata invasion alters decomposition rates in coastal areas of south-eastern Australia. Forest Ecology and Management 198:387–399. Loqué, D., and N. von Wirén. 2004. Regulatory levels for the transport of ammonium in plant roots. Journal of Experimental Botany 55:1293–1305. Ludwig, J. A., J. A. Wiens, and D. J. Tongway. 2000. A scaling rule for landscape patches and how it applies to conserving soil resources in savannas. Ecosystems 3:84–97. Mack, M. C., and C. M. D’Antonio. 2003. Exotic grasses alter controls over soil nitrogen dynamics in a Hawaiian woodland. Ecological Applications 13:154–166. Meiklejohn, J. 1968. Numbers of nitrifying bacteria in some Rhodesian soils under natural grass and improved pastures. Journal of Applied Ecology 5:291–300. Munro, P. E. 1966. Inhibition of nitrite-oxidizers by roots of grass. Journal of Applied Ecology 3:227–229. Oram, R. N. 1987. Register of Australian herbage plant cultivars: Andropogon gayanus Kunth (gamba grass). Journal of Australian Institute of Agricultural Science 53:123–124. Reed, H. E., T. R. Seastedt, and J. M. Blair. 2005. Ecological consequences of C4 grass invasion of a C4 grassland: a dilemma for management. Ecological Applications 15: 1560–1569. Rossiter, N. A. 2001. Comparative ecophysiology and fire ecology of native and exotic savanna grasses. Thesis. Northern Territory University, Darwin, Australia. Rossiter, N. A., S. A. Setterfield, M. M. Douglas, and L. B. Hutley. 2003. Testing the grass-fire cycle: alien grass invasion in the tropical savannas of northern Australia. Diversity and Distributions 9:169–176. Rossiter-Rachor, N. A., S. A. Setterfield, M. M. Douglas, L. B. Hutley, and G. D. Cook. 2008. Andropogon gayanus (gamba grass) invasion increases fire-mediated nitrogen losses in the tropical savannas of northern Australia. Ecosystems 11: 77–88. Rothstein, D., P. M. Vitousek, and L. Simmons. 2004. An exotic tree accelerates decomposition and nutrient turnover in a Hawaiian montane forest. Ecosystems 7:805–814. Russell-Smith, J., C. Yates, A. Edwards, G. E. Allan, G. D. Cook, P. Cooke, B. Heath, and R. Smith. 2003. Contemporary fire regimes of northern Australia, 1997–2001: change since Aboriginal occupancy, challenges for sustainable management. International Journal of Wildland Fire 12: 282–297. 1560 N. A. ROSSITER-RACHOR ET AL. Schmidt, S., and R. E. Lamble. 2002. Does clearing of woody vegetation result in sustainable pastures? Considerations for nutrient dynamics. Rangeland Journal 24:96–111. Schmidt, S., and G. R. Stewart. 1999. Glycine metabolism by plant roots and its occurrence in Australian plant communities. Australian Journal of Plant Physiology 26:253–264. Schmidt, S., G. R. Stewart, M. H. Turnbull, P. D. Erskine, and M. H. Ashwath. 1998. Nitrogen relations of natural and disturbed plant communities in tropical Australia. Oecologia 117:95–104. Scholes, R. J., and B. H. Walker. 1993. An African savanna: synthesis of the Nylsvley study. Cambridge University Press, Cambridge, UK. Solbrig, O. T., E. Medina, and J. F. Silva. 1996. Determinants of tropical savannas. Pages 31–39 in O. T. Solbrig, E. Medina, and J. F. Silva, editors. Biodiversity and savanna ecosystem processes: a global perspective. Springer-Verlag, Berlin, Germany. Sperry, L. J., J. Belnap, and R. D. Evans. 2006. Bromus tectorum invasion alters nitrogen dynamics in an undisturbed arid grassland ecosystem. Ecology 87:603–615. Stock, W. D., K. T. Wienand, and A. C. Baker. 1995. Impacts of invading N2 fixing Acacia species on patterns of nutrient cycling in two Cape ecosystems: evidence from soil incubations studies and 15N abundance values. Oecologia 101:375– 382. Subbarao, G. V., M. Rondon, O. Ito, T. Ishikawa, I. M. Rao, K. Nakahara, C. Lascano, and W. L. Berry. 2007. Biological nitrification inhibition (BNI): is it a widespread phenomenon? Plant and Soil 294:5–18. Subbarao, G. V., K. L. Sahrawat, W. L. Berry, K. Nakahara, T. Ishikawa, K. Suenaga, M. Rondon, and I. M. Rao. 2006. Scope and strategies for regulation of nitrification in Ecological Applications Vol. 19, No. 6 agricultural systems: challenges and opportunities. Critical Reviews in Plant Sciences 25:303–335. Sylvester-Bradley, R., D. Mosquera, and J. E. Mendez. 1988. Inhibition of nitrate accumulation in tropical grassland soil: effect of nitrogen fertilization and soil disturbance. Journal of Soil Science 39:407–416. Theron, J. J. 1951. The influence of plants on the mineralization of nitrogen and the maintenance of organic matter in the soil. Journal of Agricultural Science 41:289–296. Tothill, J. C., H. A. Nix, J. P. Stanton, and M. J. Russell. 1985. Land use and productive potential of Australian savanna lands. Pages 125–141 in J. C. Tothill and J. J. Mott, editors. Ecology and management of the World’s savannas. Australian Academy of Science, Canberra, Australia. Vitousek, P. M., and L. R. Walker. 1989. Biological invasion by Myrica faya in Hawaii: plant demography, nitrogen fixation, ecosystems effects. Ecological Monographs 59:247–265. Wedin, D., and D. Tilman. 1990. Species effects on nitrogen cycling: a test with perennial grasses. Oecologia 84:433–441. Whitehead, P., and C. Wilson. 2000. Exotic grasses in northern Australia: species that should be sent home. Pages 1–8 in Proceedings of the Northern Grassy Landscapes Conference, Katherine, Northern Territory, 29–31 August 2000. Tropical Savannas CRC, Darwin, Australia. Windham, L., and J. G. Ehrenfeld. 2003. Net impact of a plant invasion on nitrogen-cycling processes within a brackish tidal marsh. Ecological Applications 13:883–896. Witowski, E. T. F. 1991. Effects of invasive alien acacias on nutrient cycling in the coastal lowlands of the Cape fynbos. Journal of Applied Ecology 28:1–5. Wolfe, B. E., and J. N. Klironomos. 2005. Breaking new ground: soil communities and exotic plant invasion. BioScience 55:477–487. APPENDIX Grass and woody litter tissue nitrogen concentrations in native grass and Andropogon gayanus plots (Ecological Archives A019061-A1).