Abstract

Key Words

N availability, soil models, SNAP, land-use change, woodlands, semi-arid.

Introduction

The ‘Mulga Lands’ bioregion (Figure 1) represents a large, fragile part of arid and semi-arid (200-500 mm mean annual rainfall) Australia where mulga (Acacia aneura F. Muell. ex Benth.) and other Acacia spp. are the dominant vegetation (Wilson 1999). Australia is reported to have one of the highest rates of vegetation clearance in the world, with only four other countries exceeding the estimated rate of clearance of native vegetation in Australia in 1999 (Australian State of the Environment Committee 2001). In 2001, it was only next to Brazil in total area of land cleared during the 1999-2001 period (DNR&M 2003). Between 1999 and 2001, 0.8% of the Mulga Lands bioregion in Queensland was cleared of woody vegetation (DNR&M 2003). This amounted to 27% of the total woody vegetation cleared in Queensland during this period, leaving only 51% of the Mulga Lands bioregion covered by woody vegetation (DNR&M 2003). Dalal et al. (2004) have reported that soil carbon (C) and nitrogen (N) stocks decline in the Mulga Lands after clearing of the mulga for introduced buffel grass pasture. However, little information is available on N mineralisation in the Mulga Lands or on the impacts of land-use change from mulga to buffel grass pasture on N availability. Yet, prediction of N availability is important for plant N demand and for input to forest growth models, such as CenW, to assess the impacts of climate variability and change (Kirschbaum 1999).

SNAP (Soil Nitrogen Availability Predictor), is an empirical model that can predict the basal rate of N mineralisation during short-term aerobic laboratory incubations of either undisturbed or bulked and mixed soil (Paul et al. 2002). Empirical models are simple mathematical equations that can simulate experimental results. SNAP combines a simple laboratory measurement of the basal rate (k) of N mineralisation with the modifying effects of daily temperature and water content to predict seasonal and annual rates of N mineralisation in soils (Paul et al. 2002). They aimed to develop a predictive model that was simple and accurate and that could be used by both researchers and site mangers to predict N mineralisation across a range of sites. Their study indicated that k could most accurately be calculated from the amount of N mineralised after an incubation period of between 20-60 days duration, with a number of sampling times during this period in order to ensure that there is increase in N mineralisation. Although Paul et al. (2002) were able to verify the predicted rates of N mineralisation using data from 9 native forests, 12 radiata pine plantations and 12 eucalypt plantations, none of these were situated in sub-tropical Queensland. Paul et al. (2002) were also able to conclude that SNAP was an easily implemented and cost-effective method to estimate daily, seasonal and annual field rates of mineralisation across a range of soils using a simple, laboratory-based measurement of N mineralisation.

Figure 1. Map showing the study site and the distribution of mulga (Acacia aneura) woodland and open-forest in Queensland and Australia. (Sources: Geoscience Australia 2003, Queensland Herbarium 2003, Harms and Dalal 2004)

SNAP uses a laboratory incubation of N mineralisation amongst other parameters to predict field N mineralisation rates. Aerobic laboratory incubation under controlled environmental conditions is one of the most frequently used methods to assess N availability. González-Prieto et al. (1995) recommended a medium-term (6-week) incubation with intermediate measurements of N mineralisation over the short-term (2 weeks) rather than a single measurement using either one of these two incubation methods. A medium-term incubation of 6 weeks (42 days) lies within the 20-60 day incubation period as recommended by Paul et al. (2002), who also recommended a number of sampling times during the incubation period, particularly for disturbed soils, to ascertain N mineralisation increase with time of incubation during this period.

The objectives of the current study were (a) to predict field N mineralisation rates for sub-tropical mulga and an adjacent 20-year-old buffel grass pasture using SNAP and (b) to compare actual field N mineralisation rates between the adjacent 20-year-old buffel grass pasture, mulga woodlands and cropping area, recently sown to buffel grass.

Methods

Site and Soil Description

The study site (Figure 1) is located on the ‘Mulga View’ property near St George (27°59’S, 148°33’E), southern Queensland, Australia and has been described in detail by Dalal et al. (2004). The soil type is a red Kandosol (Isbell 2002) or Rhodic Paleustalf (Soil Survey Staff 2003), with a clay content of 12% and soil pH of 6.0 in the 0-10 cm layer, increasing with depth to 22% clay and pH (1:5 soil:water) 6.2 in the 60-100 cm soil layer (Dalal et al. 2004). Mean annual temperature at St George is 21°C and mean annual rainfall and pan evaporation are 516 mm and 1954 mm, respectively.

Prior to clearing in 1980, the whole site was under mulga-dominated woodland or open-forest. A portion of the cleared area was ploughed and sown to buffel grass (pasture area) and an adjoining portion (cropped area) was ploughed and sown to wheat (Triticum aestivum L.). Cattle have grazed the pasture area (old buffel), with varying intensity, while the cropped area has grown mostly wheat using conventional cultivation, but also a few crops of sorghum, and was sown to buffel grass in 2001 (new buffel). The cereal crops have usually received 2.5 kg N/ha as monoammonium phosphate, but no other fertilisers, while pastures have received no fertilisation. The average wheat and sorghum yields have been about 0.8 t/ha (Dalal et al. 2004), removing approximately 16 kg N/ha. The uncleared mulga area carries a high-density, open mulga forest, with an estimated aboveground biomass of 80 t C/ha (Harms and Dalal 2004). The same sites were used for both laboratory and field incubations described below.

Laboratory incubation

Four replicated soil samples from the 0-10 and 10-30 cm depths under mulga and 20-year-old buffel grass pasture were collected in October 2003 and aerobically incubated at 25°C and 60% water-holding capacity for six weeks. Every two weeks destructive sampling was performed to measure mineral N during the incubation period. Daily N mineralisation rates were calculated in mg/kg and used as the input into SNAP.

Field N mineralisation

In situ N mineralisation tubes were inserted to a depth of 1 m in each land use area using the technique developed by Raison et al. (1987). Six replications of three stainless steel tubes (50 mm internal diameter) were inserted in each land use area in January 2004, excluding roots. At each random sampling point, one of the tubes was immediately removed to provide baseline mineral N for that sampling cycle, one was capped to prevent N loss through leaching and the remaining tube was not capped to allow for leaching. Both these tubes (capped and uncapped) were left in situ for approximately six weeks (dependent upon rainfall) after which they were removed and three more tubes inserted to repeat the cycle. Tubes were separated into 0-10, 10-60 and 60-90 cm soil depths.

Plant N uptake was also measured using the Raison in situ N mineralisation method described above. Assuming the rates of net ammonification and nitrification were the same for soils inside and outside the tubes, the amount of mineral N taken up (plant N uptake, N_u) by the vegetation during an exposure period was estimated by:

where N_{e(t + 1)o} is the mineral N content of the uncapped tube at the end of the exposure period, and N_{b(t + 1)} is the mineral N content of bulk soil at the end of the exposure period (Raison et al. 1987).

Chemical and statistical analysis

Ammonium-N and NO₃^--N were extracted from soils by shaking field-moist soil for 1 h in 2 m KCl (1:5 soil: solution). Ammonium-N (Crooke and Simpson 1971) and NO₃^--N (Best 1976) in the extracts were determined colorimetrically on an autoanalyser (Technicon 1977). The difference in soil mineral N between the capped tube and the baseline tube was used to measure the N dynamics for the given sampling period. The dynamics for a given N form are cumulative and, for nitrate N, negative values were assumed to show losses through immobilisation or denitrification (Blumfield and Xu 2003). Differences between the capped and uncapped cores were assumed to demonstrate N losses through leaching. Plant N uptake was also calculated from the uncapped tubes as described by Raison et al. (1987). All statistical analysis was performed using the STATISTICA software package (StatSoft 2001).

Results and Discussion

Laboratory Incubation - SNAP

The N mineralisation rates for both mulga and the 20-year-old buffel grass pasture (old buffel) derived from the laboratory incubation are displayed in Figure 2 for both soil depths. Mulga soil had greater mineral N production over the six weeks in both soil depths compared to that of the old buffel, although mineral N in the old buffel soil did increase during this period. Table 1 gives the net mineral N derived from the laboratory incubation, the daily rate inserted into SNAP, the resultant annual rate as predicted by SNAP and the actual field N mineralisation rate measured in situ between January and March 2004. The N mineralisation rate predicted by SNAP was greater in the 0-10 cm soil depth than the 10-30 cm depth for both mulga and old buffel, while the annual net N mineralisation rates were greater for mulga in the 0-10 cm soil depth and greater for old buffel in the 10-30 cm depth. In the entire 0-30 cm soil depth SNAP predicted that the uncleared mulga would mineralise 92.3 kg N/ha.year and the adjacent old buffel would mineralise 58.3 kg N/ha.year.

Table 1. SNAP model inputs and outputs after a 6-week aerobic laboratory incubation of mulga and 20-year-old buffel grass pasture from Mulga View, St George, Queensland^†.

Site	Mineral N produced over 6 weeks (mg N/kg)	Input into SNAP (mg N/kg.day)	Predicted N mineralisation rate output from SNAP (kg N/ha.yr)	Field N mineralisation rate, January – March 2004 (kg N/ha.yr)
0-10 cm
Mulga	10.30 (3.6)	0.25 (0.09)	69.00	73.33
Old buffel	7.64 (0.9)	0.18 (0.02)	42.24	56.60
10-30 cm
Mulga	1.46 (0.2)	0.04 (0.01)	23.32	30.22
Old buffel	1.38 (0.3)	0.03 (0.01)	16.07	55.96

^† Data in a column are means (n = 6); data in parenthesis are standard errors of the means.

Predicted field N mineralisation - Old buffel v. mulga

Field N mineralisation rates were calculated over two sampling cycles - January to March and March to May 2004. These results were used for comparison with the results predicted by SNAP. During the first sampling cycle mulga mineralised 11.3 kg N/ha in the 0-10 cm depth and 4.6 kg N/ha in the 10-30 cm depth. This is equivalent to 73.3 kg N/ha.yr and 30.2 kg N/ha.yr for the 0-10 and 10-30 cm depths, respectively. SNAP predicts N mineralisation rates of 69.0 and 23.3 kg N/ha.yr, respectively for these two depths. This prediction was reasonably accurate for the 0-10 cm depth, but the SNAP-predicted rate of N mineralisation was 30% lower than the measured rate in the 10-30 cm depth. The field N mineralisation rate under mulga in the first sampling period for the top 0-30 cm profile is 103.3 kg N/ha.yr. This is slightly greater than, but compares favourably with the SNAP-predicted rate of 92.3 kg N/ha.year.

During the same period (Jan – March 2004), the old buffel mineralised 8.7 kg N/ha and 8.6 kg N/ha in the 0-10 and 10-30 cm depths, respectively. This is equivalent to 56.6 kg N/ha.yr and 56.0 kg N/ha.yr, respectively, totalling 112.6 kg N/ha.yr for the top 30 cm soil profile. SNAP predicts field N mineralisation rates for old buffel of 42.2 kg N/ha.yr and 16.1 kg N/ha.yr for the 0-10 cm and 10-30 cm depths respectively, with a total of 58.3 kg N/ha.yr. Therefore, SNAP slightly underestimated field N mineralisation in the 0-10 cm depth, but grossly underestimated it in the 10-30 cm depth for old buffel. For the entire 30 cm profile for this sampling period, SNAP predicted approximately half the actual N mineralisation measured in the field.

SNAP was not as successful in predicting field N mineralisation rates for the buffel grass pasture compared to mulga, but this is presumably because SNAP was intended for use in forests, particularly plantation forests, rather than pastures. However, the accuracy of SNAP predictions was much greater in the 0-10 cm soil depth than the 10-30 cm depth for both land-use areas and could be modified further to accurately predict field N mineralisation rates for mulga woodlands.

Between March and May 2004, field N mineralisation rates for mulga were 2.2 kg N/ha and -7.9 kg N/ha for the 0-10 and 10-30 cm depths, respectively. These values equate to 14.5 kg N/ha.yr and -51.3 kg N/ha.yr, respectively, and the total field N mineralisation rate for the top 30 cm profile becomes -36.8 kg N/ha.yr. Similarly for old buffel, field N mineralisation rates were -3.0 kg N/ha in the 0-10 cm soil depth and -1.7 kg N/ha in the 10-30 cm soil depth. These values become -19.9 kg N/ha.yr and -11.3 kg N/ha.yr, respectively and total -31.2kg N/ha.yr. This indicates that during the March-May sampling cycle net N immobilisation is occurring in the top 30 cm. Immobilisation (the assimilation of mineral N into microbial biomass) takes up mineral N (NH₄⁺-N + NO₃^--N) and converts it into organic form. Ammonium N is preferentially immobilised compared to NO₃^--N, but NO₃^--N immobilisation may dominate when NH₄⁺-N is limited (Recous et al. 1990), which is often the case for arable soils.

Actual field N mineralisation rates are shown in Table 2. SNAP did not allow for negative N mineralisation rates, even during long periods with little rainfall and cooling period when N immobilisation could be the dominant process, converting inorganic N to organic N for microbial assimilation. Therefore, SNAP predictions can either underestimate or overestimate the field N mineralisation rate at any given point in time, but will generally overestimate for periods where N immobilisation is occurring. Cumulative N mineralisation rates over the entire period studied could also be used for comparison, particularly as SNAP calculates field N mineralisation rates on a seasonal and annual basis. Yet, cumulative field N mineralisation rates (Table 2) over the two sampling cycles (January – May 2004) indicate large differences between measured field N mineralisation rates and those predicted by SNAP (Table 1).

Table 2. Field mineral N (January – May 2004, 112 days) under mulga, 20-year-old buffel grass pasture (old buffel), and cropped (new buffel) at Mulga View, St George, southern Queensland, Australia^†.

Site	NH4+-N (kg/ha)	NO3--N (kg/ha)	Mineral N- 112 days (kg/ha)	Yearly Mineral N (kg N/ha.year)
0-10 cm
Mulga	0.50 (2.0) a	11.96 (1.9) ab	12.46 (3.0) ab	40.61 (9.9) ab
New buffel	5.07 (2.4) a	15.16 (2.8) a	20.23 (4.6) a	65.92 (15.1) a
Old buffel	-0.84 (2.9) a	6.47 (1.8) b	5.64 (4.1) b	18.36 (13.4) b
10-60 cm
Mulga	-3.59 (3.7) a	7.31 (7.2) a	3.72 (9.4) a	12.11 (30.5) a
New buffel	5.86 (1.4) b	27.72 (7.1) b	33.58 (7.5) b	109.44 (24.5) b
Old buffel	-2.65 (2.0) a	19.77 (5.1) ab	17.11 (6.0) ab	55.77 (19.7) ab
60-90 cm
Mulga	0.13 (1.4) a	-3.10 (2.1) a	-2.97 (1.7) a	-9.67 (5.6) a
New buffel	0.46 (0.9) a	6.29 (2.1) b	6.76 (2.3) b	22.01 (7.6) b
Old buffel	-1.20 (0.8) a	4.68 (1.3) b	3.48 (1.7) b	11.34 (5.6) b
0-90 cm
Mulga	-2.96 (5.6) a	16.17 (7.3) a	13.21 (10.1) a	43.06 (32.9) a
New buffel	11.39 (2.8) b	49.18 (7.2) b	60.57 (8.3) b	197.38 (27.0) b
Old buffel	-4.69 (3.9) a	30.92 (5.6) ab	20.23 (8.1) a	85.48 (26.5) a

^† Data in a column are means (n = 6); data in parenthesis are standard errors of the means. Means within a column followed by the same letter are not different at the 5% level of significance using least significant difference (l.s.d.).

Actual field N mineralisation - Old buffel v. new buffel v. mulga

Preliminary results (Table 2) indicate that for the period from January to May 2004, cumulative field N mineralisation in the 0-10 cm depth was 12.5 kg N/ha in the mulga, 20.2 kg N/ha in the new buffel and 5.6 kg N/ha in the old buffel. Nitrate-N was always the dominant inorganic N form in these three areas with 12.0, 15.2 and 6.5 kg N/ha, respectively, derived from nitrification. The abundance of NO₃^--N is an important component for plant nutrition (Bubb et al. 1998), and the drying and re-wetting cycles would have promoted N mineralisation, particularly nitrification (Attiwill and Adams 1993; Bubb et al. 1998). In the 10-60 cm depth, cumulative field N mineralisation values were 3.7 kg N/ha, 33.6 kg N/ha and 17.1 kg N/ha, respectively for mulga, new buffel and old buffel. Again nitrification was the dominant process. Mulga mineralises less N at this depth than either of the buffel grass areas, similarly for the 60-90 cm soil depth, indicating that mulga mineralises most of its N in the surface 10 cm, while buffel can make use of N at depth. Mulga may also be capable of fixing atmospheric N₂ in biological surface crusts (Schmidt and Lamble 2002; Pate et al. 1998), which would also increase mineral N in the surface soil.

Numerous authors have consistently reported lower rates of N mineralisation in older pastures compared to the forests from which they were created (Neill et al. 1995, 1999). The old buffel has much lower mineral N than the mulga indicating the same is happening in the Mulga Lands, however, the new buffel has much greater mineral N than either the old buffel or the mulga. Neill et al. (1999) have stated that soil N turnover is maintained for a decade or so, but eventually slows in older pastures, indicating that the high rates of N mineralisation in the new buffel may be due to the addition of fertilisers to the previous crops, no stock grazing and new roots adding N to the system.

Figure 2. Mineral N during aerobic incubation of the 0-10 and 10-30 cm soil depths from adjacent mulga and 20-year-old buffel grass pasture at Mulga View, St George, Queensland.

Figure 3. Plant N uptake for each sampling cycle in the 0-90 cm soil profile at Mulga View, St George, Qld, Australia.

Plant N uptake

Plant N uptake can also be estimated from the in situ coring technique (Raison et al. 1987). Our results indicate that N uptake was dominant in the new buffel grass pasture for each sampling period (Figure 3), while Figure 4 displays the cumulative plant N uptake for the entire 0-90 cm soil profile. Plant N uptake (Figure 3) decreased in all three land-uses in the second sampling period between March and May after the growing season. Bubb et al. (1998) reported that plant N uptake was generally prominent during periods when net mineralisation was positive, however, they also reported instances where significant levels of plant N uptake coincided with negative net mineralisation (immobilisation).

Cumulative plant N uptake was significantly different between mulga and the new buffel (p<0.05), but mulga was not different to the old buffel, neither was the old buffel different to the new buffel. This higher cumulative plant N uptake in the new buffel may be the result of unchecked growth (i.e. no grazing or harvesting) and the previous use of fertiliser. This is in contrast to the old buffel, which is intermittently grazed by cattle with no added fertiliser, and the uncleared mulga that had lower levels of mineral N throughout the profile than the new buffel. Cumulative mineral N in the 0-10 cm depth was greater than cumulative plant N uptake for all land-use areas (Figure 5) possibly due to lower root N uptake from the drier top layer. Plant N uptake in the new buffel increased as soil mineral N increased, decreased in the old buffel as soil mineral N decreased, yet decreased in the mulga with an increase in mineral N. Mineral N in the mulga far exceeded plant N uptake, indicating that mulga must obtain some mineral N from biological N₂-fixation of surface crusts, termites or root nodules, as mulga mineralises very little N at depth (Table 2).

Figure 4. Cumulative plant N uptake for the 0-90 cm soil profile at Mulga View, St George, Qld, Australia.

Figure 5. Cumulative net N mineralisation and plant uptake in the 0-10 cm soil depth during January to May 2004. Data points represent the means of 5 replicates.

Conclusion

Clearing the native mulga for the introduction of exotic buffel grass pastures has been detrimental to soil organic matter quality and quantity in the mulga soils of southern Queensaland. New pastures can alleviate the degradation in the short-term, but long-term grazing for beef production, the removal of nutrients in the animal produce and no fertilisation can exacerbate the problem. This will lead to large bare soil patches and a decline in plant productivity that will reduce the available feed for grazing livestock. Consequently, the removal of mulga forest over a 20-year period in south-western Queensland for pasture and cultivation may have not only contributed to an enhanced greenhouse effect, but also threatened the sustainability of the cleared Mulga Lands.

SNAP proved capable of predicting field N mineralisation rates in the uncleared mulga, but was less successful for the 20-year-old buffel grass pasture, yet further simulations are required to obtain accurate seasonal predictions of field N mineralisation rates in the mulga. However, SNAP may be used in forests and rangelands to predict field N mineralisation rates when extensive fieldwork is not a feasible option.

Acknowledgements

The authors would like to thank Mr Ian Hill for access to ‘Mulga View’, Mr Ben Harms and Ms Christine McCallum of the Queensland Department of Natural Resources Mines and Energy for field and laboratory assistance and Dr Keryn Paul of CSIRO Forestry and Forest Products for kindly authorising the use of SNAP.

References

Attiwill PM, Adams MA (1993) Tansley Review No. 50: Nutrient cycling in forests. New Phytologist 124, 561-582.

Australian State of the Environment Committee (2001) State of the Environment. (CSIRO Publishing: Collingwood) http://www.deh.gov.au/soe/2001/pubs/soe2001.pdf (last accessed 18/06/04).

Best EK (1976) An automated method for the determination of nitrate-nitrogen in soil extracts. Queensland Journal of Agricultural and Animal Sciences 33, 161-166.

Blumfield TJ, Xu ZH (2003) Impact of harvest residues on soil mineral nitrogen dynamics following clearfall harvesting of a hoop pine plantation in subtropical Australia. Forest Ecology and Management 179, 55-67.

Bubb KA, Xu ZH, Simpson JA, Saffigna PG (1998) In situ measurements of soil mineral-nitrogen fluxes in hoop pine plantations of subtropical Australia. New Zealand Journal of Forestry Science 28, 152-164.

Crooke WM, Simpson WE (1971) Determination of ammonium in Kjeldahl digests of crops by an automated procedure. Journal of the Science of Food and Agriculture 22, 9-10.

Dalal RC, Harms B, Krull E, Wang W (2004) Total soil carbon and nitrogen and their pools following Mulga (Acacia aneura) clearing for pasture development and cropping 1. Total and labile carbon. Australian Journal of Soil Research (submitted).

DNR&M - Department of Natural Resources and Mines, (2003) Land cover change in Queensland 1999-2001. A State-wide Landcover and Trees Study (SLATS) Report. http://www.dnr.qld.gov.au/slats/ 77p. The State of Queensland: Queensland Department of Natural Resources and Mines, Brisbane.

Geoscience Australia (2003) Vegetation – Pre-European Settlement, 1788. Department of Industry, Tourism and Resources: Commonwealth of Australia http://www.ga.gov.au/nmd/products/thematic/vegmap.jsp

González-Prieto SJ, Carballas M, Villar MC, Carballas T (1995) Organic nitrogen mineralisation in temperate humid-zone soils after two and six weeks of aerobic incubation. Biology and Fertility of Soils 20, 237-242.

Harms BJ, Dalal RC (2004) The sustainability of Mulga Lands after clearing – what do changes in soil carbon and nitrogen show? In Conserving Soil and Water for Society: Sharing Solutions. Proc. 13th International Soil Conservation Organisation Conference. Eds. SR Raine, AJW Biggs, NW Menzies, DM Freebairn, PE Tolmie 4-9th July, Brisbane. ASSSI/IECA. Paper 1015. 6p.

Isbell RF (2002) The Australian Soil Classification. 2^nd ed. CSIRO Publishing: Melbourne.

Kirschbaum MUF (1999) CenW, a forest growth model with linked carbon, energy, nutrient and water cycles. Ecological Modelling 118, 17-59.

Neill C, Piccolo MC, Steudler PA, Melillo JM, Feigl BJ, Cerri CC (1999) Nitrogen dynamics in soils of forests and active pastures in the western Brazilian Amazon Basin. Soil Biology and Biochemistry 27, 1167-1175.

Neill C, Piccolo MC, Melillo JM, Steudler PA, Cerri CC (1999) Nitrogen dynamics in Amazon forest and pasture soils measured by ¹⁵N pool dilution. Soil Biology and Biochemistry 31, 567-572.

Pate JS, Unkovich MJ, Erskine PD, Stewart GR (1998) Australian mulga ecosystems- ¹³C and ¹⁵N natural abundances of biota components and their ecophysiological significance. Plant, Cell and Environment 21, 1231-1242.

Paul KI, Polglase PJ, O’Connell AM, Carlyle JC, Smethurst PJ and Khanna PK (2002) Soil nitrogen availability predictor (SNAP): a simple model for predicting mineralisation of nitrogen in forest soils. Australian Journal of Soil Research 40, 1011-1026.

Queensland Herbarium (2003) Regional Ecosystem Mapping, Version 4, September 2003. Environmental Protection Agency: Queensland Government http://www.epa.qld.gov.au/nature_conservation/biodiversity/regional_ecosystems/

Raison RJ, Connell MJ, Khanna PK (1987) Methodology for studying fluxes of soil mineral-N in situ. Soil Biology and Biochemistry 29, 1557-1563.

Recous S, Mary B, Faurie G (1990) Microbial immobilisation of ammonium and nitrate in cultivated soils. Soil Biology and Biochemistry 22, 913-922.

Schmidt S, Lamble RE (2002) Nutrient dynamics in Queensland savannas: implications for the sustainability of land clearing for pasture production. The Rangeland Journal 24, 96-111.

Soil Survey Staff (2003) Keys to Soil Taxonomy. 9^th ed. USDA Natural Resources Conservation Service: Washington

StatSoft Inc. (2001) STATISTICA (data analysis software system), version 6. www.statsoft.com.

Technicon (1977) Individual/simultaneous determination of nitrogen and/or phosphorus in BC acid digests, Industrial Method No. 334-374 W/B. Technicon Industrial Systems: Terrytown, New York.

Wilson BA (1999) Mulga Lands. In ‘Conservation status of Queensland’s bioregional ecosystems’. Ed. P Sattler, R Williams pp. 6/1-6/10. The State of Queensland, Environmental Protection Agency: Brisbane.