Printer Friendly

Nitrogen balance in Australia and nitrogen use efficiency on Australian farms.

Introduction

International research on nitrogen (N) in agricultural systems combines the goals of managing N to profitably increase production and nitrogen use efficiency (NUE), with reducing environmental damage when surplus reactive N, particularly from fertiliser, flows from agricultural land to the natural environment. The purpose of this paper is to summarise both strands of research in Australia, including revision of a previous national N balance, which can indicate current and future areas of surplus reactive N (Galloway et al. 2004). We distinguish between the N balances at the scales of the continental landmass and agricultural land because the N balance of the vast nonagricultural zones may disguise or even swamp that of crops and pastures. In the agricultural zones, the goals of improving NUE and reducing leakage of reactive N are compatible. We discuss prospects for reducing surplus N in regions where it is causing environmental damage and improving fertiliser NUE for dryland cropping.

Australian N balance

The land area of Australia is the sixth largest of the ~200 countries but crops and improved pastures make up only 13% of the total area and intensive animal industries are relatively small compared with other developed countries (Table 1). The estimates of the N balance of the Australian continent follow the work of Denmead (1990), Galbally et al. (1992) and McLaughlin et al. (1992), updated by the area and yield of crops and numbers of livestock. The land area is separated into four regions (Fig. 1) to provide context for the agricultural land, based on the methods and maps of ABARES (2010). These regions are (1) non-agricultural land, mainly in the arid zone, as well as forests and woodlands, conservation areas, lakes, cities and defence areas, (2) pastoral rangeland in semiarid and tropical regions, (3) dryland farmland consisting of permanent pastures (i.e. pastures containing annual or perennial grasses or legumes growing on land that has seldom, if ever, grown crops), continuous dryland cropping and phased crop-pasture systems, and (4) intensive agriculture consisting of dairying, irrigated horticultural and field crops, and intensive animal industries. The basis for the estimates in Table 1 are presented as supplementary material. Changes made to the methods used in the three original papers were to (1) decrease the N loss from burning vegetation estimated by Galbally et al. (1992), based on satellite estimates of nitrogen oxide emissions by Jaegle et al. (2005), (2) increase the estimated ammonia loss from vegetation in the arid zone from those proposed by Denmead (1990) to provide consistency with ammonia (N[H.sub.3]) loss in other regions, (3) estimate the spatial transfer of organic N in dust storms, and (4) estimate the mining of soil organic N by crops. From Table 1 the estimated annual input of N in Australian ecosystems is 8.9 Mt, equivalent to 3% of the global input of 300 Mt (Galloway et al. 2004) but the annual consumption of fertiliser N, 1.5 Mt, is 1% of the world consumption.

Total inputs and outputs of N are almost balanced within each of the individual regions of non-agricultural, pastoral rangelands and intensively farmed land (Table 1). The components of the N balance differ between regions because of the N concentration and the size of the region. For example the pastoral zone receives more rainfall than the non-agricultural zone so the N input in rain is greater; the dryland farming zone receives even more rainfall but across a relatively small area.

Within the dryland farming zone the estimated N balance is positive, mainly due to N-fixation by legumes in permanent pastures. The other land uses in this region consist of phased crop-pasture sequences and continuous crops. There are no estimates of the areas of each system but the area of dryland crop increased, sheep numbers decreased and cattle numbers increased since the 1980s (Kirkegaard et al. 2011). It is clear that many previous crop-pasture farms now grow only crops but there are no estimates of numbers. Inputs and outputs of N in the regions arc discussed in the following sections.

Arid non-agricultural land and grazed rangelands

The inland arid and semiarid land make up most of the national estate. Part of these zones has no agricultural production and the rest consists of pastoral properties that produce cattle and sheep with the only N input from atmospheric deposition and biological N-fixation. The N inputs on a per hectare basis are small but the amount over the whole area far exceeds the N offtake in meat and wool. There is no evidence of soil-N accumulation or depletion within historical time in these regions. The contribution from rain in the arid zones (and in the agricultural zones) is calculated from the land area, mean annual rainfall and an N concentration in rainwater of 0.5 mg [L.sup.-1], based on measurements of Wetselaar and Hutton (1963) and Crockford and Khanna (1997). The main source of N deposited in rain is not related to local thunderstorm activity (Wetselaar and Hutton 1963) but is likely to be mainly recycled N from biomass burning and N[H.sub.3] released from soil, plants and the urine voided by grazing animals (Denmead 1990; Galbally et al. 1992). There is also likely to be biological N-fixation from various symbiotic (e.g. Acacia and other leguminous shrubs and forbs, Casuarina, Macrozamia and lichens) and non-symbiotic associations (Gupta et al. 2006), cyanobacterial soil crusts, free-living microbes and the gut bacteria of cellulose-digesting termites (Evans et al. 2011). Non-symbiotic N-fixation is likely to be greater in northern than in southern Australia because of generally lower soil-N status and high temperatures during the wet season.

A significant amount of N is lost and redistributed in dust storms that occur most frequently when strong winds coincide with a period of drought. Most of the soil particles are lifted from the arid inland and some is from farmland. McTainsh et al. (2005) reported an average of 62 dust storms each year from 1960 to 2002 in Australia, most of which remain within the arid zone and only the largest reach the coast. They estimated that a large dust storm in 2002 contained 3.4-4.9 Mt of particulates.

A later large dust storm reported by Aryal et al. (2012) contained 10.6% organic matter. Assuming an 'average' dust storm half the size estimated by McTainsh et al. (2005) and an organic matter content of 10.6%, of which an estimated 4% was N, we come up with an annual estimate 0.5 Mt of N redistributed within the Australian continent or blown into the sea. This represents redistribution of 0.7 kg N ha 1 across the arid zone.

Dryland farming

Fanning in this zone consists of permanent pasture grazed by sheep and cattle, continuous cropping and phased crop-pasture sequences. From 1850 to 2000, crop area grew at an average annual rate of 3.2% (Angus and Good 2004) as crops replaced, and continue to replace, pastures (Fig. 2). The expansion of cropping has been partially within established farming regions as well as expansion into new regions in Western Australia, Queensland and along the low-rainfall and high-rainfall boundaries of the cropping region (Kirkegaard et al. 2011). The only source of N for the first century of crop production was from mining the soil, initially through continuous cropping and then through fallow-crop sequences (Donald 1965). Soil N was replenished by biological N-fixation by legumes in improved pastures after the 1950s, triggered by a high wool price and encouraged by two decades of a subsidy by the Australian Government for phosphate fertiliser (Henzell 2007). Soil total N is maintained when legume-based pastures make up about half the farm area (Angus and Peoples 2012). The area and quality of pastures that significantly contribute to the N balance is difficult to estimate; the total area defined by ABARES (2010) as Grazing Modified Pastures is 72 Mha, but pastures in the north-eastern part of this zone contain relatively few legumes; a major source of N for livestock in the region is urea blocks, which represent much of the estimated 25 000 t of N used as licks and stockfeed (Table 3; CN Walker, pers. comm.). A more realistic estimate of the area of pastures that contribute to the N balance is 50 Mha and the estimated biological N-fixation is 60 kg N [ha.sup.-1] (Peoples et al. 2012). This is consistent with the estimate of McDonald (1989), who found annual increments in soil N from legume-based pastures ranging within 19-117 kg N [ha.sup.-1] (average 63 kg N [ha.sup.-1]) from 15 studies across southern Australia. Crop legumes also contribute to the N balance of dryland farms but make up only ~5% of the crop area.

Rainfall is variable in most of the dryland farming zone, particularly in the north-east. Yield of dryland crops can be partly buffered by stored soil moisture associated with fallows, stubble retention and direct drilling (Hunt and Kirkegaard 2011), but yield is still strongly tied to rainfall during the growing season. The variable yield potential presents a challenge for N management. Applying an average amount of N-fertiliser can result in underfertilisation in favourable seasons and overfertilisation in droughts, leading to reduced yield due to 'haying off (van Herwaarden et al. 1998). Many grain growers adopt a tactical approach to N management, aiming to delay application of fertiliser until the yield potential is more predictable during the stem-elongation phase. This is followed by topdressing an amount of N, often based on a simple budget as described below, and timed to synchronise with peak crop-N demand. Allowance for an estimate of mineralised N in such a budget is essential for matching N supply and demand, but also recognises the annual decrease in soil organic N (Table 2). Crops and livestock are integrated on mixed farms, not just through the sequence of crops and pastures, but also through grazing vegetative crops by sheep and cattle, the effect of which is for greater reliance on fertiliser N than that biologically fixed by pasture (Virgona et al. 2006). Livestock also graze crop stubble and increase the accumulation of soil mineral N, because the consumption of high-carbon (C) residues reduces N immobilisation (Hunt et al. 2016). Crops grown after canola also benefit from increased N mineralisation (Ryan et al. 2006).

Irrigation and high-rainfall farming

Most of the irrigated land lies on semiarid plains along the inland rivers of eastern and south-eastern Australia. The yields and N input to crops--cotton, rice, maize, wheat, as well as to dairy pastures and horticultural and viticulture crops--are high by international standards. The area of irrigated land is limited by the amount of irrigation water which, during drought, is conserved for perennial crops and milking cattle. The area of annual crops and hence the amount of N-fertiliser used are variable but the application rate is high because water is assured. Thus NUE is generally greater for irrigated than for dryland crops because N-fertiliser can be washed into the root zone soon after application. Poor irrigation management leads to denitrification losses due to prolonged periods of soil saturation (Mathers et al. 2007). There are no reports of leaching below the root zone but Weaver et al. (2013) measured accumulation of nitrate (N[O.sub.3.sup.-]) at the bottom of the cotton root zone.

High-rainfall fanning land is found along the south-west, and the south-east and east coasts where it is concentrated between the Great Dividing Range and the ocean. The largest users of N-fertiliser in this zone are dairying, sugarcane and horticulture. Gourley et al. (2012) surveyed Australian dairy farms and found an average NUE of 26% and an N-surplus > 100 kg N ha , based not only on fertiliser but also on fodder imported to the farm. Bell et al. (2016) report similar NUE for sugarcane but no comparable data are available for horticulture. Raised beds are increasingly employed to manage waterlogged cropping soils and increase NUE across southern Australia (MacEwan et al. 2010).

Soil N levels

The conventional wisdom is that Australian soils are ancient and infertile (PMSE1C 2010) and this generalisation certainly applies to many soils that arc deeply weathered and located in regions that are climatically unsuited for farming (McKenzie et al. 2004). Soils in some marginal agricultural regions are naturally acidic, sodic or saline, but it is not clear whether these constraints limit productivity more than climate or farm management. The generalisation about ancient and infertile soils does not apply to the extensive areas of agricultural soils in south-eastern Australia that were enriched by Quaternary aeolian deposits (McKenzie et al. 2004) as well as productive alluvial soils in eastern Australia with minor weathering and even small areas of soil formed on basalt flows that post-date human occupation. Elsewhere in southern Australia, soil N levels increased from pre-farming levels due to biological N-fixation from extensive legume-based pasture (Ladd and Russell 1983; Grace and Oades 1994). Clearing native vegetation for farms in South Australia and Western Australia after the 1950s was often followed by a period of continuous legume-based pasture that increased soil N before commencement of ley farming or continuous cropping.

In many undisturbed Australian agricultural soils the total N content was, by international standards, consistent with their water balance, temperature and texture. For example, the average N content of one of the most widespread agricultural soil types, the red-brown earths (Chromosols, Dermosols, Kandosols and Sodosols) was ~1.5 g [kg.sup.-1] before intensive agriculture (Stace et al. 1968), comparable with undisturbed soils in parts of the United States (Arkansas and Mississippi) with a similar mean annual temperature of ~15[degrees]C (Jenny 1941). The original nutrient content of soil in Australia before agriculture, as in agricultural soils everywhere, becomes decreasingly relevant to production and offsite effects as fertiliser supplies more of the nutrients removed in crops and livestock.

N mining

Long-term experiments in Australia show that the total N (and organic C) content of soils decreases with continuous cropping and crop--fallow systems. Clarke and Russell (1977) reviewed many experiments that quantified the amount of soil N removed by crops during the first half of the twentieth century. The experimental crops received no N-fertiliser and the low yield levels and rates of soil-N depletion were unrepresentative of current cropping systems. Table 2 reports more recent observations and experiments where yields were representative of current crops. In some of these cases the decrease in soil total N with continuous cropping and fallowing appears to be linear when measured over periods of several decades but is non-linear over a longer period, falling to a new equilibrium. The pattern of decrease is expressed as a half-life of total N in the soil based on the amount of N measured at the first and last observation. Averaged over 15 experiments and observations reviewed in Table 2, the half-life of total N in the soil was 23 [+ or -] 12 years.

In order to estimate the average extent of soil-N mining on cropping land, we combine the estimated half-life with an estimate of the number of non-legume crops harvested from an average field since 1850. The small area where crops were grown in 1850 could possibly have grown 165 crops, whereas fields that first grew crops in 2014 could have grown no more than one crop. The area of crops has increased at the expense of pastures, and doubled since the 1970s (Fig. 2), so the average number of crops is at the lower end of the possible range. Equation 1 provides an estimate of the average number of crops by weighting the number of years since crops were first grown with cropping area, A, in each year, y, normalised to the area in 2014 (Fig. 2). The remaining term in Eqn 1 is which represents the frequency of cropping after the first crop.

N = f/165 [y=2014.summation over (y=1850)] (2015 -y) [A.sub.y]/[A.sub.2014] (1)

There are no statistical or survey estimates of f, so a sensitivity analysis is justified. Solving Eqn 1 provides an estimate that 16 crops had been grown on an average cropping field by 2014, assuming continuous cropping since the first crop, or eight crops based on the more realistic assumption of one crop every two years after the first crop. Combining the estimate of eight crops with a half-life for total soil N of 23 years provides an estimate that 22% of the total N has been mined from the average field. This is a surprisingly small number and reflects the relatively recent expansion of cropping. For the large proportion of land where cropping started in the 1960s and 1970s (Fig. 2), it is likely that ~20 crops have been grown and these have mined ~45% of soil total N.

There are few comparable estimates from long-established farming regions internationally, where early farmers undoubtedly mined soil N. In the warm environment of Tanzania, the topsoil N under maize receiving no N inputs over 15 years (Solomon et al. 2000) had a half-life of 15 years. In the cooler North Dakota environment, topsoil N under long-term wheat-fallow receiving no N inputs for 45 years (Schimel 1986) had a half-life of 65 years.

A notable result among the long-term experiments is that applying N-fertiliser to continuous crops had little effect on the depletion rate of soil total N (Russell 1981 ; Heenan et al. 2004). Soil N (and C) can be replenished by pastures (e.g. Ladd and Russell 1983; Grace and Oades 1994; Helyar et al. 1997), and phased crop-pasture systems dominated the dryland farming systems in southern Australia from the 1950s to the early 1990s. These systems maintained soil N and C with little N-fertiliser, provided that about half the farm grew pastures (Angus and Peoples 2012). It is possible, but expensive, to replenish soil N (and C) in a continuous cropping system when stubble retention is combined with applied fertiliser N, phosphorus (P) and sulfur to maintain the ratios of these nutrients in soil (Kirkby et al. 2016).

Factors influencing the ability of soil to supply N to crops include the amount and quality of soil organic matter and residues, disturbance, moisture and temperature regimes (Campbell et al. 1981). An indicator of soil-N supply is the mineral N content in agricultural soils (typically to a depth of 60 cm) before sowing winter crops. Fillery (2001) reported a mean value of 98 kg N [ha.sup.-1] from a survey of experiments after pasture in Western Australia. Results from the laboratory of Incitec-Pivot Ltd since 2010, representing hundreds of farm samples in eastern Australia, indicate a mean value of-80 kg N ha in the top 60 cm. These values are higher than comparable measurements in Western Europe and North America, probably because of the relatively high levels of total N and because the relatively high temperatures in Australia promote mineralisation. Potentially mineralisable N stores in south-eastern Australia range from 8% of the total N in cropping systems, where residues are burnt, to 22% after 15 years of residue retention (Gupta et al. 1994).

A conclusion from this section is that growing continuous crops is effectively mining the non-renewable resource of soil N. The most common alternative is applying N-fertiliser, which, when produced from natural gas, is also produced from a non-renewable resource. The sustainable alternatives are to invigorate the ley-farming system to provide more biologically-fixed N or produce N-fertiliser using a renewable energy source such as solar or wind power.

Fertiliser N use

Before the mid-1990s most of the N-fertiliser used in Australian agriculture was for high-value crops such as horticulture and sugarcane and little was applied to dryland crops (Angus 2001). The reason for the generally low rate was not lack of research and extension, but because wheat yield did not reliably respond profitably to applied N at the time (Colwell and Morton 1984).

The growth in N-fertiliser usage in Australia was slow compared with the rest of the world before the mid-1990s (Fig. 3) but for the rest of that decade there was a boom in N-fertiliser use, mostly as inputs to wheat and other dryland crops. This boom coincided with increased area of canola production and lime application. Canola was the first widely grown non-legume break crop in Australia and wheat grown after canola responded more reliably to N-fertiliser than wheat after wheat (Angus 2001). Lime was needed because canola is acid sensitive. Liming also enabled other acid-sensitive crops and pasture species to be grown. Other factors that encouraged farmers to apply N at this time were the availability of efficient fertiliser spreaders and increased premiums for high grain protein. The use of N-fertiliser stabilised during the millennium drought of 2002-2009, after which usage resumed its upward course (Fig. 3). Most of the N-fertiliser applied to dryland crops is topdressed and at the relatively low rate of 45 kg N [ha.sup.-1] (Table 3). Intensive crops and pastures occupy a relatively small area of land but receive larger application rates.

Myers (1984) proposed a simple budget of N inputs and output to estimate N-fertiliser requirement for a single field, in this case a wheat crop which represents the largest crop and consumer of N (Table 4). When this topic was visited previously (Angus 2001), the average N-fertiliser application to wheat was 30 kg N [ha.sup.-1], which represented about one-third of the N total supply, the remainder coming from depletion of native soil total N and the N fixed by pasture legumes and a small contribution from crop legumes. At that time, the supply of N-fertiliser worldwide provided about half of the supply to world agriculture (Jenkinson 2001). Applying the same approach to update estimates for all Australian dryland crops in 2014, we estimate that fertiliser provides ~40% of the total N input. The rates of fertiliser N applied in 2001 and 2014 suggest that N application is increasing at an annual rate of ~1 kg ha -1.

In Table 4, mineralisation is partitioned into the contributions from mining the soil and N-fixation by previous pastures. The procedure was to first estimate N-fixation, assuming 50% pasture on the farm from 1950 until 2014 and mineralisation of legume residues according to the rates estimated by Angus and Peoples (2012). These estimates include only the N contributions of legumes and do not include other benefits of legumes to the following crops, for example, stimulated mineralisation, reduced immobilisation and carry-over of spared soil mineral N (Peoples et al. 2017).

The contribution from soil mining was then estimated from the difference between total N-mineralisation and the contribution from N-fixation. Other estimates in Table 4 are the amounts of N retained in the soil or lost. Both are based on the fate of [sup.15]N fertiliser reviewed in Fig. 4, increased to account for the flow of non-fertiliser N to these pathways. This allocation is based on measurements showing that NUE of native soil N was similar to NUE of fertiliser for dryland wheat (Angus et al. 1998).

N balance in ley farming and continuous cropping

The combined supply of mined soil-N, biologically-fixed N from ley pasture and fertiliser are simulated for a pasture-crop sequence, starting in 1950, about the time when improved pastures were first widely adopted in southern Australia (Fig. 5). The equations used to generate the annual changes in Fig. 5 were, for the cropping phase:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)

and for the pasture phase:

[DELTA][N.sub.t], = [N.sub.fix] (3)

where [N.sub.t] is total amount of soil N, and the annual change, [DELTA][N.sub.t], is net mineralisation. The parameter [r.sub.y] is the rate of net mineralisation of the native N in year y of the cropping phase, expressed as a proportion of [N.sub.t]. the variable [N.sub.fix] is biological N-fixation during the pasture phase and mineralisation of the N fixed one year earlier ([N.sub.fix,y-1]) and two years earlier ([N.sub.fix,y-2]), with the mineralisation rates of the fixed N represented by [r.sub.1] and [r.sub.2] respectively; N-fertiliser is not included as a contributor to [N.sub.t] based on the observation of Heenan et al. (2004) that N, was little affected by long-term annual applications of N-fertiliser.

Based on the assumptions in the caption of Fig. 5, the amount of N fixed in the pasture phase was not enough to offset the decrease in native soil total N. However, starting in the mid-1970s, the supply of N-fertiliser began to compensate for the decreased soil N, in terms of soil mineral N supply for crops. The use of average fertiliser data should be interpreted cautiously since N application to crops grown in pasture rotations is presumably less than to continuous crops. Within pasture-crop systems more N-fertiliser is presumably applied in the later years of a cropping phase, and Fig. 5a suggests that second and third crops require ~10 kg N [ha.sup.-1] more than first crops.

The emphasis on the N balance of ley farming in this section is because of its complexity and not necessarily because of its future importance. For the increasingly popular continuous cropping system, the balance of total N is described simply as the half-life (Table 2).

NUE and losses

The NUE is a generic term expressed in many ways but in this case we refer mainly to apparent aboveground recovery of fertiliser N (AARFN) because this enables comparisons between species and takes some account of grain protein. There are more complete and complex assessments of NUE that consider both yield and grain protein responses to fertiliser N, and their relative profitability (Fischer et al. 1993; Angus 1995).

For irrigated crops in the semiarid zone, AARFN averages ~40% for well managed cotton (Macdonald et al. 2017) but can be higher in favourable seasons (Rochester and Bange 2016) and 76% for well managed medium-grain rice (Angus et al. 2016), both at yield levels on farms that are high by world standards. AARFN in dryland crops can be lower. From a survey of 60 commercial dryland wheat crops in south-eastern Australia (Angus et al. 1989, those that were most likely to give large N responses (mid-season topdressing, early sowing, low N status, following a break crop and with adequate soil water) had an average AARFN of 36%. This value does not include rhizodeposition, which is the fertiliser N contained in the roots and excreted by roots into the soil. Wichern et al. (2008) reported that rhizodeposited N represented about one-third of total crop N, so the real fertiliser recovery was greater than the apparent recovery of 36%. For the 60 commercial crops, the average agronomic efficiency (the additional grain per unit of additional fertiliser N) was 13 but the N : grain price ratio was 6 (i.e. the price of N was six times the price of grain). At this price ratio it was profitable for farmers to apply N-fertiliser to grain crops, despite the relatively low AAFN.

These results are consistent with experiments that traced [sup.15]N fertiliser applied to 74 wheat crops in Australia (Fig. 4). At maturity, 44 [+ or -]14% of the fertiliser was in aboveground crop tissue, 34 [+ or -]14% in soil and 22 [+ or -]16% was not recovered, presumably lost by one or more of the processes of denitrification, leaching and N[H.sub.3] volatilisation. The greater crop-N recovery in the [sup.15]N experiments than in commercial crops reflects more favourable experimental conditions.

Although these single-year experiments mostly showed N loss, there was no evidence of N loss in a 15-year, continuous-crop experiment in western Victoria (Norton et al. 2015). In this experiment the amount of N removed in grain equalled the amount of fertiliser N removed plus the depletion in soil N. The cause of the variability of the results in Fig. 4 has not been explained and their difference from the long-term experiment suggests that the question of N loss is not well understood and needs further research.

National statistical data show mixed evidence for the relationship between wheat production and N-fertiliser use (Fig. 6). Of the fertiliser N applied in Australia, about half is applied to wheat (Table 3), so any association between national N use and wheat production should be interpreted cautiously. The large increase in wheat production in the 1990s followed a rapid rise in fertiliser use, as well as the first significant adoption of break crops and lime application (Angus 2001). The millennium drought from 2002 to 2009 led to decreased yields and less use of N-fertiliser, so these seasons are not relevant to this discussion. However, after the drought the upward trend in fertiliser-N use resumed but was not accompanied by increased wheat production until the exceptionally wet 2016 growing season. The lack of yield response in the 10-15 years before then may have been due to unfavourable seasons, particularly low spring rainfall (Hochman et al. 2017), or because there were no other innovations like the use of break crops and lime in the 1990s. The national wheat yield of 2.61 [ha.sup.-1] in 2016 was more than 20% greater than any previous season. It is possible that the combination of adequate water and improved management, including the increased amount of N-fertiliser, was responsible for the increased yield. Future trends in N-fertiliser and wheat production may resolve this issue.

The loss of fertiliser N indicated by the [sup.15]N studies on wheat are reflected in other agricultural systems. Losses exceeding 20% of applied N are reported for dairy farms (Rowlings et al. 2016; Stott and Gourley 2016) and sugarcane production systems (Bell 2014). Pilbeam (1996) showed that amount of fertiliser N remaining in the soil at maturity of wheat was greater in the generally dry environments in Australia than in wetter environments where more of the labelled N was present in the crop.

Previous improvements in NUE of Australian dryland crops have come about mostly by synchronising fertiliser application with crop N demand and by increasing crop-N demand through early sowing, controlling root disease, correcting soil acidity and micronutrient deficiency, and by increasing the yield potential by plant breeding (Angus 2001). Increasing N demand by the crop leads to more rapid N uptake and so less exposure of the fertiliser to the immobilisation and loss pathways. New methods to increase NUE are discussed in the Conclusions.

Denitrification

Studies reviewed by Chen et al. (2008) and Grace (2015) showed that loss of fertiliser N by denitrification is common in Australian crops. Since most of the fertiliser N is applied to crops and pastures during winter and spring, the rate of denitrification is probably limited by temperature. Assuming that most of the N loss estimated from Fig. 4 represents denitrification, the annual N loss from the 1.59 Mt of fertiliser N (Table 3) would be <0.35 Mt. Less research has been conducted on losses from non-fertiliser sources of N, but the results of Pu et al. (1999) show large losses of N mineralised from organic matter in eastern Australia when the soil is warm and wet. The most widespread occurrence of warm and wet soil is during floods in summer and autumn. One extensive flood on the plains of eastern Australia in January 2011 covered 1.7 x [10.sup.8] ha, mostly in cropping and grazing land. This area is equivalent to the combined areas of France, Germany, the Netherlands, Belgium, Denmark and Norway, and much of this land remained inundated for over a month in mid-summer. No measurements of denitrification are reported for this event, but it would be reasonable to assume that all the soil N[O.sub.3.sup.-] was denitrified. A conservative estimate of the soil N[O.sub.3.sup.-] is 10 kg N [ha.sup.-1] in the top 0.6 m, based on the lowest values of recent soil samples analysed at the Incitec-Pivot laboratory (CN Walker, pers. comm.). Assuming this amount, the total denitrification from this one flooding event would be 1.3 Mt of N. Such flooding events are infrequent but still represent a loss of N comparable with, or greater than, the loss from fertiliser. Denitrification is also significant in water bodies and Harris (2001) concluded that >75% of dissolved N could be lost through this pathway.

Ammonia and nitrogen oxide exchange with the atmosphere

Ammonia is emitted to the atmosphere from soil, plants, burning biomass and animal excreta (Denmead 1990). Soils and plants can also capture NH3 from the atmosphere. The extent of NH3 loss from native vegetation is uncertain, but Denmead (1990) considered this pathway to be the main net source of N[H.sub.3] to the atmosphere. The second largest agricultural source of atmospheric N[H.sub.3] is from the urine of grazing cattle, sheep and kangaroos. The amounts are estimated following the methods of Denmead (1990) with contemporary livestock numbers, which in 2014 amounted to ~300 million sheep equivalents. The loss of N[H.sub.3] from the extensive arid zone dominates these losses. The amounts partly balance N deposition in rainfall and may represent part of the same material. In the dryland farming zone, annual N deposition from all sources is ~5 kg N [ha.sup.-1] less than one-fifth of the amounts in farming regions of North America and Western Europe (Burkart and James 1999; Addiscott 2005). Ammonia deposition downwind of large Australian cattle feedlots is comparable with those in Western Europe and North America (Denmead et al. 2014), contributing free fertiliser to downwind farms but environmental damage to water bodies and native vegetation growing on poorly buffered soil.

The loss of N[H.sub.3] to the atmosphere from dryland crops is mostly from hydrolysis of urea applied to the surface of moist alkaline soil which contains urease, typically in plant residues. In Australia the situations in which this is most likely to occur are dairy pastures (Eckard et al. 2003) or crops growing on alkaline soils or with heavy residues, where >20% of fertiliser can be lost as N[H.sub.3] (Turner et al. 2012). The model of Fillery and Khimashia (2016) shows that drilling urea into the soil or topdressing before rain can reduce these losses to zero. In our experience most farmers use rainfall forecasts in planning to topdress urea and minimise losses.

Burning biomass releases nitrogen oxide and nitrogen dioxide (N[O.sub.x]) that, like N[H.sub.3], are mostly returned to the land in rain (Galbally et al. 1992). Fires in the arid zone were traditionally managed by the indigenous practice of firestick farming (Gammage 2011). This traditional practice of relatively frequent 'cool' burns in patches of tens of hectares protected the landscape from less frequent 'hot' wildfires over areas of many thousands of hectares, mostly started by lightning, which are now the main form of biomass burning in the nonagricultural zone (Burrows et al. 2006). On grazed rangeland, the loss of N[O.sub.x] by burning vegetation has been partly suppressed by graziers and replaced by N[H.sub.3] loss from the urine of grazing animals.

In the dryland farming zone, burning cereal stubbles is usually by 'cool' fires and certainly results in loss of N to the atmosphere. However the long-term reduction in total soil N with burnt stubble is not significantly greater than with retained stubble, which decomposes within months or years (Angus et al. 2006). No-till is increasing on Australian grain farms (Llewellyn et al. 2012) and this practice reduces the need for the time-consuming process of burning stubbles.

Leaching of N[O.sup.-.sub.3] and runoff of particulate N

The N[O.sup.-.sub.3] leaches from the topsoil into the subsoil or groundwater when there is N[O.sup.-.sub.3] in the soil profile and the water supply exceeds the water-holding capacity. The largest losses are on coarse-textured soil in regions and seasons when the water balance is positive for part of the year, such as in winter-rainfall regions of southern Australia and in the wet season in northern Australia. There are relatively few observations of N[O.sup.-.sub.3] leaching in Australia and most of those reported in Table 5 are on the wet fringe of the dryland farming region or in an intensive farming region. Anderson et al. (1998) measured up to 59 kg N [ha.sup.-1] leached from a coarse-textured soil in a relatively high-rainfall part of the dryland farming region of Western Australia. Extrapolations to other parts of the region using a simulation model suggested that the long-term mean quantity of leaching varied from zero on a loamy sand in a dry environment to 50 kg N [ha.sup.-1] [year.sup.-1] on a sand in a wet environment (Milroy et al. 2008).

Along the wetter fringe of the dryland farming regions of eastern Australia, winter rainfall is less intense and N[O.sup.-.sub.3] leaching is about half the values in Western Australia. In the examples for eastern Australia listed in Table 5, N03 leaching could be reduced with more intensive management, for example growing perennial rather than annual pastures (Ridley et al. 2001), by earlier sowing of crops and splitting N-fertiliser (Anderson et al. 1998) or by double cropping (Turpin et al. 1998). By contrast, improved management had less effect on leaching in Western Australia. The leaching of N[O.sup.-.sub.3] occurs infrequently on most dryland cropping farms in eastern Australia because the soil water-holding capacity is generally sufficient to contain the surplus of rainfall over potential evapotranspiration.

Leached N[O.sup.-.sub.3] and particulate N contribute to N pollution of surface water bodies. The concentration of N in most streams that drain areas of permanent pasture in Victoria exceed official guidelines (Ridley et al. 2004), and there is evidence of leaching into shallow groundwater. There are few measurements of N03~ leaching to the watertable, which is normally tens of metres below the surface in the dryland farming region, but because of the surplus N in the permanent pasture and dairy-farming regions the problems of leaching and runoff are likely to get worse unless more N is removed.

High N[O.sup.-.sub.3] concentration in groundwater is not of much public concern in Australia because relatively small amounts of groundwater are used for human consumption. In fact the largest areas of groundwater affected by high N[O.sup.-.sub.3] levels are not due to fertiliser. One is in arid central Australia due to leaching of N that had been biologically fixed by termite gut bacteria in geological time (Barnes et al. 1992). Another large area of high-N03 groundwater, in a winter-rainfall region of South Australia, is of agricultural origin but from mineralisation of organic N derived from biological N-fixation by clover (Dillon 1988).

Offsite damage from leaching and runoff has been extensively studied in Australia. Damage to the Great Barrier Reef is the most grievous consequence of leaching and runoff of inorganic N (and P) from near-coastal intensive agriculture sugarcane fields and particulate erosion from extensive inland grassy woodlands grazed by cattle, discussed by Bell et al. (2016).

There is also N (and P) runoff into the estuaries and coastal lagoons along the south-western, southern and eastern coastline (Harris 2001). These water bodies periodically become eutrophied by nutrients sourced from diffuse and point sources in cleared catchments. Eutrophication is not exclusively a result of farming and there are records of algal blooms in rivers and lagoons before white settlement, presumably because of concentration of nutrients during drought. There is also eutrophication due to treated and untreated human sewage. Harris (2001) concluded that the N and P discharge to Australian coastal waters was small compared with those in the northern hemisphere because of less atmospheric N deposition, lower population densities and less fertiliser use. Eutrophication in Australian coastal lagoons is normally N-limited and there are frequent N-fixing cyanobacterial blooms. It thus becomes important to minimise movement of P into the watercourses. For the Gippsland Lakes, Roberts et al. (2012) concluded that the most cost-effective methods to reduce contamination were by enforcing existing regulations on the large sources of P from the dairy industry.

Another serious effect of N[O.sup.-.sub.3] leaching is acidification of poorly buffered soil in southern Australia (Helyar and Porter 1989). In this case N[O.sup.-.sub.3] in the topsoil, mostly originating from biological N-fixation by pastures, and more recently from Nfertiliser, leaches into the subsoil along with alkali and alkaliearth metals, which arc then replaced by protons adsorbed onto clay minerals in the topsoil. The lime needed to neutralise the acidity caused by the N cycle is an additional cost of N from both biological N-fixation and fertiliser. Lime is applied to crops growing on acidified soils but the amount is probably less than the amount needed to reverse the acidification. Much less lime is applied to permanent pasture than to crops.

Conclusions

The amount of reactive N is fairly stable over most of the Australian continent but the stability of N in the vast arid zone buffers changes in the agricultural zones. Natural processes of N redistribution in episodic windstorms and N loss by denitrification during flood events are comparable to the N inputs from fertiliser. The large uncertainties of N fluxes during such episodes contrast with precise estimates of the amounts in fertiliser and agricultural offtake and show a need for more research. However, it is clear from the N balance that Australian agricultural ecosystems are less dependent on fertiliser N than other developed countries. Non-fertiliser sources of N provide a significant amount of N input for agricultural production and also contribute to N-related environmental pollution.

Within the agricultural regions there is a net accumulation of reactive N, mainly because of N-fixation by legumes in permanent pastures in the dryland farming zone and dairy regions. This in turn causes N03~ leaching into groundwater and soil acidification. The N surplus in permanent pastures is partly offset by the mining of soil N by dryland crops but not on the same land. The fertiliser that replaces the N depicted by dryland crops has a low NUE. We offer suggestions for further research to improve the efficiency of N-fertiliser for dryland crops and the environmental problems caused by excess N.

N-fertiliser for dryland crops

Because of N-mining by continuous cropping, the current application rate of N-fertiliser to dryland crops, ~45 kg N ha ', will need to double in about five decades if yields are to be maintained and NUE is not improved. Even more N-fertiliser will be required if yields are to achieve their water-limited potential (Sadras and Angus 2006). If the current value of NUE (44%, Fig. 4) is not improved, the cost of inefficient fertiliser use will increase in proportion to the amount applied, and will increase relatively more than the total cost of crop production, even if the per-unit price of fertiliser does not increase. It is impossible to predict the future price of fertiliser N or the closely related natural gas price, which has been relatively low since development of non-conventional gas extraction (EIA 2017). Factors that are likely to increase the price of N-fertiliser, or at least prevent a price decrease, are the eventual depletion of natural gas reserves and the approaching limits to the efficiency with which N[H.sub.3] is synthesised from natural gas (Smil 2001).

Research to increase the low NUE of Australian dryland crops is therefore a high priority. Many decades of research on N-fertiliser rates, dates, forms and additives have led to an inconsistent but generally low NUE and it is unlikely that more of the same will lead to improvement. Better understanding of the N-fertiliser processes and new application methods may lead to increased NUE. The loss processes of leaching and N[H.sub.3] volatilisation have been extensively investigated but less is known about denitrification and the fate of fertiliser N that is retained in the soil at harvest (Fig. 4). The amount of N-fertiliser retained in the inorganic and organic pools is not known. Within the organic pool we do not know how much fertiliser N is directly immobilised by microbes and how much is taken up by the crop and later returned to the soil by the process of rhizodeposition (Wichern et al. 2008).

The process of N immobilisation proceeds at the same time as gross mineralisation (Myrold and Bottomley 2008) but can be spatially separated in the soil. The balance tips towards immobilisation when residues are added to the soil but the rate of immobilisation can still be significant even with no additions. The microbes responsible for immobilisation (and denitrification) compete for mineral N with roots (Myrold and Bottomley 2008), so it may be possible to increase NUE by increasing uptake by roots, suppressing uptake by microbes or by reducing N[O.sup.-.sub.3] loss potential through the use of Enhanced Efficiency Fertilisers (EEFs) (Lester et al. 2016). As discussed above, increasing crop N demand has been responsible for previous increases in N-uptake (Angus 2001) but there has been little attention paid to suppressing the assimilation of fertiliser N by microbes. The fate of immobilised fertiliser N is also unknown. If it is remineralised within a few months it may contribute profitably to the following crop, but if remineralisation approaches the average rate of soil organic matter decomposition (3% per year, Table 2) then the delayed return on investment represents a significant cost. There are examples of crops responding to N-fertiliser applied to the previous crop and even to the crop before that, but many crops do not show responses to N-fertiliser applied in the previous year. Presumably soil properties, environment and management are responsible for the variation. Research to clarify the processes may help liberate some of the fertiliser N remaining in the soil at harvest.

There may be opportunities to suppress immobilisation of fertiliser N by increasing the spatial separation between N-fertiliser and microbes. Current methods of application such as broadcasting and banding solid or fluid fertiliser in narrow rows provides little advantage to crops competing for mineral N with microbes in the topsoil. A possible method to provide an advantage to plants is to inject N into subsoil where microbial populations are low. Another is to inject urea or anhydrous N[H.sub.3] in shallow mid-row bands as part of the sowing operation to produce a sufficiently high ammonium concentration to temporarily inactivate the microbes through elevated pH. Yet another, made possible by precision guidance, is by in-crop mid-row banding of urea as an alternative to topdressing. Banding between every second crop row can double the urea concentration in the rows compared with banding between every row, and still provides each crop row with access to one fertiliser band. Concentrating ammonium in bands is known to delay nitrification and hence reduce the risk of losses by denitrification and N[O.sup.-.sub.3] leaching (Wetselaar et al. 1973; Angus et al. 2014). The slow release of N[O.sup.-.sub.3] from urea or anhydrous N[H.sub.3] bands helps prevent excessive vegetative crop growth. Sandral et al. (in press) reported an experiment with dryland wheat where the apparent grain recovery of N from mid-row banded urea applied at the most profitable rate (80kg N [ha.sup.-1]) was 52%, significantly greater than current practices of broadcasting at stem elongation (45%) or broadcasting and incorporation by sowing (36%). Further research on mid-row urea banding is needed and combining banded urea with EEFs (urease inhibitors, nitrification inhibitors and polymer coatings) may increase NUE more than when these products are used alone.

Environmental damage from N

Environmental damage from excess N identified in this review is caused by N[O.sup.-.sub.3] leaching and particulates entrained in runoff from pastures and crops in the winter wet season in southern Australia and the summer wet season in northern Australia. When the leaching and runoff is connected to estuaries and lagoons the N (and P) causes eutrophication. The sources of much of the excess is N-fixation by permanent pastures in the dryland farming region and dairy farms, and also N-fertiliser and imported feed on dairy farms (Aarons et al. 2017). Leakage of excess N from permanent pastures is a widespread problem that requires a reduction in soil N levels as well as less N input. A possible solution that deserves research is the introduction of occasional grain or hay crops into current permanent pastures and export of the produce from the farm. Although such a practice could reduce the N surplus there is the risk of disrupting the existing grazing system and releasing reactive N to the atmosphere (Mielenz et al. 2017). Additionally, the small paddock size suitable for grazing livestock is potentially incompatible with the large size needed for efficient crop production. Electronic systems of livestock management such as virtual fences (Butler et al. 2006) are a possible replacement for fences and enable efficient crop production to be reintroduced to land growing permanent pastures. One of the consequences of N[O.sup.-.sub.3] leaching is rapid soil acidification on poorly buffered soils. Liming to reverse acidification is increasing on cropping soils but there is little evidence of such increases on permanent pastures. Introducing occasional crops could also lead to more liming.

The N (and P) pollution of the Great Barrier Reef is a similar problem to estuary eutrophication but at a larger scale and with greater adverse consequences. Here, as with nutrient loss from permanent pastures, the main contributor to N runoff and leaching may not be fertiliser. In the case of the Great Barrier Reef, a major source of N pollution may be N and P adsorbed on particulates eroded from overgrazed grassy woodlands in the catchments by occasional intense and widespread rainfall over inland catchments (Bell et al. 2016). A promising way to reduce this erosion is accelerated destocking of cattle when the vegetation cover is low. However, it is not clear whether nutrients adsorbed on particulates contribute as much pollution as the more frequent release of inorganic nutrients from intensive agriculture. Evidence that intensive agriculture is a major source of nutrient pollution comes from correlations between excess fertiliser N input and export of inorganic N from catchments (Thorburn and Wilkinson 2013). Methods proposed to abate pollution from intensive agriculture include changes in crop type, reduction of N-fertiliser and restoration of riparian native vegetation (Kroon et al. 2016). Identifying the major sources of pollution would hasten and reduce the cost of abatement.

Denitrification causes not only a production loss but also contributes to climate change through nitrous oxide emissions. The methodological problems in measuring dinitrogen gas emissions, normally orders of magnitude larger than for nitrous oxide, hinders measurement of denitrification and development of strategies to increase NUE, particularly on heavier textured soils. Improved application methods and use of EEFs, as discussed in the previous section, may reduce nitrous oxide emissions along with other N-fertiliser losses. Further research is required on measuring and suppressing denitrification from both fertiliser and non-fertiliser sources on agricultural and non-agricultural land. Research on non-fertiliser sources of denitrification should be linked to measurement of mineralisation and nitrification of soil organic matter and residues.

It is clear from the N balance that Australian agricultural ecosystems are less dependent on fertiliser N than other developed countries. Non-fertiliser sources of N provide most of the input for agricultural production and are also the source of much of the N-related environmental damage. In many developed countries such damage is mainly due to intensive production methods (Galloway et al. 2004) but in Australia the problems arc mainly from extensive methods that could be improved with sustainable intensification of farming systems (Godfray and Garnett 2014).

http://dx.doi.org/10.1071/SR16325

Conflicts of interest

The authors declare no conflicts of interest.

Acknowledgements

We are grateful to Lucy Randall for preparing the map and to Rob Norton, Mark Peoples and Charlie Walker for helpful discussions.

References

Aarons SR. Gourley CJP, Powell M, Hannah MC (2017) Estimating nitrogen excretion and deposition by lactating cows in grazed dairy systems. Soil Research 55. doi: 10.1071/SR17033

ABARES (2010) Land use in Australia--at a glance, www.agriculture. gov.au/abares/aclump/Documents/Land_use_in_Australia_at_a_glance_ 2006.pdf

ABARES (2015) Australian commodity statistics 2015. www.agriculture. gov.au/abares/publications/display ?url=http://l 43.188.17.20/anrdl/ DAFFService/display.php?fid=pb_agcstd9abcc0022015_11a.xml

Addiscott TM (2005) 'Nitrate, agriculture and the environment.' (CABI Publishing: Wallingford, UK)

Anderson GC, Fillery IRP, Dunin FX, Dolling PJ, Asseng S (1998) Nitrogen and water flows under pasture-wheat and lupin-wheat rotations in deep sands in Western Australia 2. Drainage and nitrate leaching. Australian Journal of Agricultural Research 49, 345-361. doi:10.1071/A97142

Angus JF (1995) Modeling N fertilization requirements for crops and pastures. In 'Nitrogen fertilization and the environment'. (Ed. PE Bacon) pp. 109-127. (Marcel Dekker: New York)

Angus JF (2001) Nitrogen supply and demand in Australian agriculture. Australian Journal of Experimental Agriculture 41, 277-288. doi: 10.1071 /EA00141

Angus JF, Good AJ (2004) Dryland cropping in Australia. In 'Challenges and strategies of dryland agriculture.' (Eds S Rao, J Ryan) pp. 151-166. (American Society of Agronomy: Madison Wl, USA)

Angus JF, Peoples MB (2012) Nitrogen from Australian dryland pastures. Crop & Pasture Science 63, 746-758. doi: 10.1071/CPI2161

Angus JF, van Herwaarden AF, Fischer RA (1989) Predicting the yield response of wheat to topdressed nitrogen, p. 550 In: Proceedings of the 5th Australian Agronomy Conference, http://www.regional.org.au/ au/asa/1989/contributed/plant-nutrition/p-18.htm

Angus JF, van Herwaarden AF, Fischer RA, Howe GN, Heenan DP (1998) The source of mineral nitrogen for cereals in southeastern Australia. Australian Journal of Agricultural Research 49, 511 -522. doi: 10.1071 / A97125

Angus JF, Boiger TP, Kirkegaard JA, Peoples MB (2006) Nitrogen mineralisation in relation to previous crops and pastures. Australian Journal of Soil Research 44, 355-365. doi: 10.1071/SR05138

Angus JF, Gupta VVSR, Pitson GD, Good AJ (2014) Effects of banded ammonia and urea fertilizer on soil and the growth and yield of wheat. Crop & Pasture Science 65, 337 352. doi: 10.1071/CP13337

Angus JF, Bacon PE, Reinke RF (2016) Temporary immobilisation promotes nitrogen use efficiency of irrigated rice. In 'Proceedings of the International Nitrogen Initiative Conference, Melbourne, Australia'. (http://ini2016.com/conference-proceedings-2)

Armstrong RD, Halpin NV, McCosker K, Standley J, Lisle AT (1996) Applying nitrogen to grain sorghum in central Queensland: residual value and effect of fallowing and tillage practice. Crop & Pasture Science 47, 81-95. doi:10.1071/AR9960081

Armstrong RD, Probert ME, McCosker K, Millar G (1998) Fluxes of nitrogen derived from plant residues and fertiliser on a cracking clay in a semi-arid environment. Crop & Pasture Science 49. 437-449. doi:10.1071/A97069

Aryal R, Kandel D, Acharya D, Chong MN, Beecham S (2012) Unusual Sydney dust storm and its mineralogical and organic characteristics. Environmental Chemistry 9, 537 -546. doi: 10.1071/EN12131

Bacon PE, Freney JR (1989) Nitrogen loss from different tillage systems and the effect on cereal grain yield. Fertilizer Research 20, 59-66. doi: 10.1007/BF01055429

Barnes CJ, Jacobsen G, Smith GD (1992) The origin of high-nitrate ground waters in the Australian arid zone. Journal of Hydrology 137, 181-197. doi : 10.1016/0022-1694(92)90055-Z

Bell MJ (2014) 'A review of nitrogen use efficiency in sugar cane.' (Sugar Research Australia Ltd: Brisbane, Queensland) 344 pp.

Bell M, Lester D, De Antoni Migliorati M, Rowlings D, Grace P (2015) Nitrogen use efficiency in summer sorghum grown on clay soils. In 'Proceedings 17th Australian Society of Agronomy Conference, Hobart, Australia', (http://www.agronomyaustralia.org/conference-proceedings)

Bell M, Schaffelke B, Moody P, Waters D, Silburn M (2016) Tracking nitrogen from the paddock to the reef- a case study from the Great Barrier Reef. In 'Proceedings of the International Nitrogen Initiative Conference, Melbourne, Australia', (http://ini2016.com/conferenceproceedings-2).

Burkart MR, James DE (1999) Agricultural-nitrogen contributions to hypoxia in the Gulf of Mexico. Journal of Environmental Quality 28, 850- 859. doi : 10.2134/jeq 1999.00472425002800030016x

Burrows ND, Burbidge AA, Fuller PJ, Behn G (2006) Evidence of altered fire regimes in the Western Desert region of Australia. Conservation Science Western Australia 5, 272-284.

Butler Z, Corke P, Peterson R, Rus D (2006) From robots to animals: virtual fences for controlling cattle. The International Journal of Robotics Research 25, 485-508. doi:10.1177/0278364906065375

Campbell C, Myers R, Weier K (1981) Potentially mineralizable nitrogen, decomposition rates and their relationship to temperature for five Queensland soils. Australian Journal of Soil Research 19, 323-332. doi:10.1071/SR9810323

Chen D, Suter H, Islam A, Edis R, Freney JR, Walker CN (2008) Prospects of improving efficiency of fertiliser nitrogen in Australian agriculture: a review of enhanced efficiency fertilisers. Australian Journal of Soil Research 46, 289-301. doi:10.1071/SR07197

Clarke AL, Russell JS (1977). Crop sequential practices. In 'Soil factors in crop production in a semi-arid environment'. (Eds JS Russell, EL Greacen) pp. 279-300. (University of Queensland Press: St. Lucia)

Colwell JD, Morton R (1984) Development and evaluation of general or transfer models of relationships between wheat yields and fertilizer rates in southern Australia. Australian Journal of Soil Research 22, 191-205. doi: 10.1071 /SR9840191

Crockford RH, Khanna PK (1997) Chemistry of throughfall, stemflow and litterfall in fertilized and irrigated Pinus radiata. Hydrological Processes 11, 1493-1507. dot: 10.1002/(SICI)1099-1085(199709)11:11<1493:: AID-HYP475>3.0.CO;2-P

Dalal RC (1992) Long term trends in total nitrogen of a vertisol subjected to zero-tillage, nitrogen application and stubble retention. Australian Journal of Soil Research 30, 223-231. doi:10.1071/SR9920223

Dalal RC, Mayer RJ (1986) Long-term trends in fertility of soils under continuous cultivation and cereal cropping in southern Queensland. V Rate of loss of total nitrogen from the soil profile and changes in carbon: nitrogen ratios. Australian Journal of Soil Research 24, 493-504. doi : 10.1071 /SR9860493

Dalal RC, Thornton CM, Cowie BA (2013) Turnover of organic carbon and nitrogen in soil assessed from 813C and 5I5N changes under pasture and cropping practices and estimates of greenhouse gas emissions. The Science of the Total Environment 465, 26-35. doi:10.1016/ j.scitotenv.2013.04.101

De Antoni Migliorati M, Bell MJ, Grace PR, Rowlings DW, Scheer C, Strazzabosco A (2014) Assessing agronomic and environmental implications of different N fertilisation strategies in subtropical grain cropping systems on Oxisols. Nutrient Cycling in Agroecosystems 100, 369-382. doi: 10.1007/s 10705-014-9655-4

Denmead OT (1990) An ammonia budget for Australia. Australian Journal of Soil Research 49, 511-522.

Denmead OT, Chen D, Rowell D, Loh Z, Hill J, Muir S, Griffith DWT, Naylor T, Bai M, Phillips F, McGinn S (2014) Gaseous nitrogen emissions for Australian cattle feedlots. In 'Nitrogen deposition, critical loads and biodiversity'. (Eds MA Sutton et al.), pp. 23-29. (Springer: Dordrecht)

Dillon PJ (1988). An evaluation of the sources of nitrate in groundwater near Mount Gambier, South Australia. CSIRO Water Resources Series 1. CSIRO Land and Water, Canberra.

Donald CM (1965) The progress of Australian agriculture and the role of pastures in environmental change. Australian Journal of Science 27, 187-198.

Dunsdorfs E (1956) 'The Australian wheat-growing industry 1788-1948.' (Melbourne University Press)

Eckard RJ, Chen D, White RE, Chapman DF (2003) Gaseous nitrogen loss from temperate perennial grass and clover dairy pastures in south-eastern Australia. Australian Journal of Agricultural Research 54, 561-570. doi: 10.1071 /AR02100

EIA (2017). U.S. Energy Information Agency, http://www.eia.gov.

Evans TA, Dawes TZ, Ward PR, Lo N (2011) Ants and termites increase crop yield in a dry climate. Nature Communications 2,262. doi: 10.1038/ ncomms1257

Fillery IRP (2001) The fate of biologically fixed nitrogen in legume-based dryland farming systems: a review. Australian Journal of Experimental Agriculture 41, 361-381. doi: 10.1071/EA00126

Fillery IRP, Khimashia N (2016) Procedure to estimate ammonia loss after N fertiliser application to moist soil. Soil Research 54, 1-10. doi: 10.1071/ SR 14073

Fischer RA, Howe GN, Ibrahim Z (1993) Irrigated spring wheat and timing and amount of nitrogen fertilizer. 1. Grain yield and protein content. Field Crops Research 33, 37-56. doi: 10.1016/037841290(93)90093-3

Freney JR. Smith CJ, Mosier AR (1992) Effect of a new nitrification inhibitor (wax coated calcium carbide) on transformations and recovery of fertilizer nitrogen by irrigated wheat. Fertilizer Research 32, 1-11. doi: 10.1007/BF01054388

Galbally IE, Fraser PJ, Meyer CP, Griffith DWT (1992) Biosphereatmosphere exchange of trace gasses over Australia. In 'Australia's renewable resources: sustainability and global change'. (Eds RM Gifford, MM Barson) pp. 117 149. (Bureau of Rural Resources: Canberra)

Galloway JN, Dentcner FJ, Capone DG (2004) Nitrogen cycles: past, present, and fitture. Biogeochemistry 70, 153-226. doi:10.1007/s10533004-0370-0

Gammage B (2011) 'The biggest estate on earth.' (Allen and Unwin: Sydney)

Godfray HCJ, Garnett T (2014) Food security and sustainable intensification. Philosophical Transactions of the Royal Society of London. Series B. Biological Sciences 369, 20120273. doi:10.1098/rstb.2012.0273

Gourley CJP, Dougherty WJ, Weaver DM, Aarons SR, Awty IM, Gibson DM, Hannah MC, Smith AP, Peverill Kl (2012) Farm-scale nitrogen, phosphorus, potassium and sulfur balances and use efficiencies on Australian dairy farms. Animal Production Science 52, 929-944. doi:10.1071/AN11337

Grace P (2015) National coordination of an integrated, data synthesis and modelling network for reducing N20 emissions from Australian soils. Final Report to the Australian Department of Agriculture. Queensland University of Technology, Brisbane.

Grace PR, Oades JM (1994) Long-term field trials in Australia. In 'Longterm experiments in agricultural and ecological sciences'. (Eds RA Leigh, AE Johnson) pp. 53-81. (CAB International, Wallingford)

Gupta VVSR, Grace P, Roper MM (1994) Carbon and nitrogen mineralization as influenced by long-term soil and crop residue management systems in Australia. In 'Defining soil quality for a sustainable environment'. (Eds J Doran, D Bezdicek, DC Coleman) pp. 193-200. (Soil Science Society of America Special Publication No. 35: Madison, WI)

Gupta VVSR, Roper MM, Roget DK (2006) Potential for non-symbiotic N2fixation in different agroecological zones of southern Australia. Australian Journal of Soil Research 44, 343-354. doi: 10.1071/SR05122

Harris GP(2001) Biogeochemistry of nitrogen and phosphorus in Australian catchments, rivers and estuaries: effects of land use and flow regulation and comparisons with global patterns. Marine and Freshwater Research 52, 139-149. doi: 10.1071/MF00031

Harris R, Armstrong R, Wallace A (2015) Recovery of [sup.15]N urea fertiliser applied to wheat under different management strategies in the High Rainfall Zone of south western Victoria. In 'Proceedings 17th Australian Society of Agronomy Conference, Hobart, Australia' (http://www. agronomyaustralia.org/conference-proceedings)

Heenan DP, Chan KY, Knight PG (2004) Long-term impact of rotation, tillage and stubble management on the loss of soil organic carbon and nitrogen from a Chromic Luvisol. Soil & Tillage Research 76, 59-68. doi: 10.1016/j.still.2003.08.005

Helyar KR, Porter WM (1989) Soil acidification, its measurement and the processes involved. In 'Soil acidity and plant growth'. (Ed. AD Robson) pp. 61-101. (Academic Press: Sydney)

Helyar KR, Cullis BR, Furniss K, Kohn GD, Taylor AC (1997) Changes in the acidity and fertility of a red earth soil under wheat-annual pasture rotations. Australian Journal of Agricultural Research 48, 561-586. doi: 10.1071 /A96069

Henzell T (2007) 'Australian agriculture: its history and challenges.' (CSIRO: Collingwood)

Hochman Z, Gobbett DL, Horan H (2017) Climate trends account for stalled what yields in Australia since 1990. Global Change Biology doi: 10.1111/gcb.13604

Hunt JR, Kirkegaard JA (2011) Re-evaluating the contribution of summer fallow rain to wheat yield in southern Australia. Crop & Pasture Science 62, 915-929. doi: 10.1071/CP11268

Hunt JR, Swan AD, Breust PD, Peoples MB, Kirkegaard JA (2016) Sheep grazing on crop residues increase soil mineral N and grain N uptake in subsequent wheat crops. In 'Proceedings of the International Nitrogen Initiative Conference, Melbourne, Australia'. (http://ini2016. com/conference-proceedings-2).

Jaegle L, Steinberger L, Martin RV, Chance K (2005) Global partitioning of NOx sources using satellite observations: relative roles of fossil fuel combustion, biomass burning and soil emissions. Faraday Discussions 130, 407-423. doi: 10.1039/b502128f

Jenkinson DS (2001) Nitrogen in a global perspective, with focus on temperate areas--state of the art and global perspectives. Plant and Soil 228, 3-15. doi : 10.1023/A: 1004870606003

Jenny H (1941) 'Factors in soil formation.' (McGraw-Hill: New York)

Kirkby CA, Richardson AE, Wade LJ, Conyers M, Kirkegaard JA (2016) Inorganic nutrients increase humification efficiency and C-sequestration in an annually cropped soil. PLoS One 11(5), e0153698. doi:10.1371/ journal.pone.0153698

Kirkegaard JA, Peoples MB, Angus JF, Unkovich MJ (2011) Diversity and evolution of rainfed farming systems in southern Australia. In 'Rainfed farming systems'. (Eds P Tow, I Cooper, I Partridge, C Birch) pp. 715-754. (Springer: Dordrecht)

Kroon FJ, Thorburn P, Shaffelke B, Whitten S (2016) Towards protecting the Great Barrier Reef from land-based pollution. Global Change Biology 22, 1985-2002. doi:10.1111/gcb.13262

Ladd JN, Russell JS(1983) Soil nitrogen. In 'Soils: an Australian viewpoint'. pp. 589-607. (CSIRO: Melbourne and Academic Press: London)

Lam SK, Chen D, Norton R, Armstrong R (2012) Nitrogen demand and the recovery of [sup.15]N-labelled fertilizer in wheat grown under elevated carbon dioxide in southern Australia. Nutrient Cycling in Agroecosystems 92. 133-144. doi: 10.1007/s 10705-011 -9477-6

Lester DW, Bell MJ, Bell KL, Migliorati M DA, Scheer C, Rowlings D, Grace PR (2016) Agronomic responses of grain sorghum to DMPP-treated urea on contrasting soil types in north-eastern Australia. Soil Research 54, 565-571. doi: 10.1071/SR 15337

Llewellyn RS, D'Emden FH, Kuehne G (2012) Extensive use of no-tillage in grain growing regions of Australia. Field Crops Research 132, 204-212. doi:10.1016/j.fcr.2012.03.013

Macdonald BCT, Chang YF, Nadelko A, Tuomi S, Glover M (2017) Tracking fertiliser and soil nitrogen in irrigated cotton: uptake, losses and the soil N stock. Soil Research 55, 264-272. doi: 10.1071/SR16167

MacEwan RJ, Crawford DM, Newton PJ, Clune TS (2010) High clay contents, dense soils, and spatial variability are the principal subsoil constraints to cropping the higher rainfall land in south-eastern Australia. Soil Research 48, 150-166. doi:10.1071/SR09076

Mathers NJ, Nash M, Gangaiya P (2007) Nitrogen and phosphorus exports from high rainfall zone cropping in Australia: Issues and opportunities for research. Journal of Environmental Quality 36, 1551-1562. doi: 10.2134/jeq2006.0464

McDonald GK (1989) The contribution of nitrogen fertiliser to the nitrogen nutrition of rainfed wheat crops in Australia: a review. Animal Production Science 29. 455-481. doi:10.1071/EA9890455

McKenzie N, Jacquier D, Isbell R, Brown K (2004) 'Australian soils and landscapes.' (CSIRO: Canberra)

McLaughlin MJ, Fillery IR, Till AR (1992) Operation of the phosphorus, sulphur and nitrogen cycles. In 'Australia's renewable resources: sustainability and global change'. (Eds RM Gifford, MM Barson) pp. 67-116. (Bureau of Rural Resources: Canberra)

McTainsh G, Chan Y, McGowan H, Leys J, Tews K (2005) The 23rd October 2002 dust storm in eastern Australia: characteristics and meteorological conditions. Atmospheric Environment 39, 1227-1236. doi : 10.1016/j.atmosenV.2004.10.016

Mielenz H, Thorburn PJ, Harris RH, Grace PR, Officer SJ (2017) Mitigating [N.sub.2]O emissions from cropping systems after conversion from pasture--a modelling approach. European Journal of Agronomy 82, 254-267. doi : 10.1016/j.eja.2016.06.007

Milroy SP, Asseng S, Poole ML (2008) Systems analysis of wheat production in low water-holding soils in a Mediterranean environment II. Drainage and nitrate leaching. Field Crops Research 107, 211-220. doi: 10.1016/j.fcr.2008.02.008

Myers RJK (1984) A simple method for estimating the nitrogen fertiliser requirement ofacereal crop. Fertilizer Research 5,95-108. doi: 10.1007/ BF01049494

Myrold DD, Bottomley PJ (2008). Nitrogen mineralization and immobilization. In 'Nitrogen in agricultural systems'. (Eds JS Schepers, WR Raun) pp. 157 172. (American Society of Agronomy: Madison WI)

Norton R, Walker C, Farlow C (2015) Nitrogen removal and use on a long-term fertilizer experiment. In 'Proceedings 17th Australian Agronomy Conference, Hobart'. (http://www.agronomyaustralia.org/conference-proceedings)

Peoples MB, Brockwell J, Hunt JR. Swan AD, Watson L, Hayes RC, Li GD, Hackney B, Nuttall JG, Davies SL, Fillery IRP (2012) Factors affecting the potential contributions of N2 fixation by legumes in Australian pasture systems. Crop & Pasture Science 63, 759-786. doi: 10.1071/ CP12123

Peoples MB, Swan AD, Goward L, Kirkegaard JA, Hunt JR, Li GD, Schwenke GD, Herridge DF, Moodie M, Wilhelm N, Potter T, Denton MD, Browne C, Phillips LA, Khan DF (2017) Soil mineral nitrogen benefits derived from legumes and comparisons of the apparent recovery of legume or fertiliser nitrogen by wheat. Soil Research 55. doi: 10.1071/SR16330

Pu G, Saffigna PG, Strong WM (1999) Potential for denitrification in cereal soils of northern Australia after legume or grass-legume pastures. Soil Biology & Biochemistry 31, 667-675. doi: 10.1016/S0038-0717(98) 00154-0

Pilbeam CJ (1996) Effect of climate on the recovery in crop and soil of [sup.15]N-labelled fertilizer applied to soil. Fertilizer Research 45, 209-215. doi: 10.1007/BF00748591

PMSEIC (2010) Australia and food security in a changing world. (Prime Minister's Science, Engineering and Innovation Council, Canberra). www.chiefscientist.gov.au

Poss R, Smith CJ, Dunin FX, Angus JF (1995) Rate of soil acidification under wheat in a semi-arid environment. Plant and Soil 177, 85-100. doi :10.1007/BF00010340

Rasiah V, Armour JF, Menzies N W, Heiner DH, Donn MJ, Mahendrarajah S (2003) Nitrate retention under sugarcane in wet tropical Queensland deep soil profiles. Australian Journal of Soil Research 41, 1145-1161. doi : 10.1071/SR02076

Ridley AM, White RE, Helyar KR, Morrison GR, Heng LK, Fisher R (2001) Nitrate leaching loss under annual and perennial pastures with and without lime on a duplex (texture contrast) soil in southeastern Australia. European Journal of Soil Science 52, 237-252. doi: 10.1046/ j.1365-2389.2001.00387.x

Ridley AM, Mele PM, Beverly CR (2004) Legume-based farming in Southern Australia: developing sustainable systems to meet environmental challenges. Soil Biology & Biochemistry 36, 1213-1221. doi: 10.1016/ j.soilbio.2004.04.007

Roberts AM, Pannell DJ, Doole G, Vigiak O (2012) Agricultural land management strategies to reduce phosphorus loads in the Gippsland Lakes, Australia. Agricultural Systems 106, 11 22. doi: 10.1016/j.agsy. 2011.10.009

Robertson FA, Myers RJK, Saffigna PG (1997) Nitrogen cycling in brigalow clay soils under pasture and cropping. Australian Journal of Soil Research 35, 1323-1340. doi: 10.1071/S97026

Rochester IJ, Bange M (2016) Nitrogen fertiliser requirements of high-yielding irrigated transgenic cotton. Crop & Pasture Science 67, 641-648. doi: 10.1071/CP15278

Rowlings DW, Scheer C, Liu S, Grace PR (2016) Annual nitrogen dynamics and urea fertilizer recoveries from a dairy pasture using 15N; effect of nitrification inhibitor DMPP and reduced application rates. Agriculture, Ecosystems & Environment 216, 216-225. doi: 10.1016/ j.agee.2015.09.025

Russell JS (1981) Models of long term organic nitrogen change. In 'Simulation of nitrogen behaviour of soil-plant systems'. (Eds MJ Frissel, JA van Veen) pp. 222-232. (PUDOC: Wageningen)

Ryan MH, Kirkegaard JA, Angus JF (2006) Brassica crops stimulate soil mineral N accumulation. Australian Journal of Soil Research 44, 367-377. doi:10.1071/SR05143

Sadras VO, Angus JF (2006) Benchmarking water use efficiency of rainfed wheat crops in dry mega-environments. Australian Journal of Agricultural Research 57, 847-856. doi:10.1071/AR05359

Sandrai GA, Tavakkoli E, Harris F, Koetz E, Angus J (in press) Improving nitrogen fertiliser use efficiency in wheat using mid-row banding. In 'Proceedings 18th Australian Society of Agronomy Conference, Ballarat, Australia', (http://www.agronomyaustralia.org/conferenceproceedings)

Schimel DS (1986) Carbon and nitrogen turnover in adjacent grassland and cropland ecosystems. Biogeochemistry 2, 345-357. doi: 10.1007/ BF02180325

Schwenke G, Haigh B (2016) The interaction of seasonal rainfall and nitrogen fertiliser rate on soil N20 emission, total N loss and crop yield of dryland sorghum and sunflower grown on sub-tropical Vertosols. Soil Research 54, 604-618. doi:10.1071/SRI5286

Smil V (2001) 'Enriching the earth.' (MIT Press: Cambridge, MA)

Smith CJ, Whitfield DM, Gyles OA, Wright GC (1989) Nitrogen fertilizer balance of irrigated wheat grown on a red-brown earth in southeastern Australia. Field Crops Research 21, 265-275. doi: 10.1016/0378-4290 (89)90008-7

Solomon D, Lehmann J, Zech W (2000) Land use effects on soil organic matter properties of chromic luvisols in semi-arid northern Tanzania: carbon, nitrogen, lignin and carbohydrates. Agriculture, Ecosystems & Environment 78, 203-213. doi:10.1016/S0167-8809 (99)00126-7

Stace HCT, Hubble GD, Brewer R, Northcote KH, Sleeman JR, Mulcahy MJ, Hallsworth EG (1968) 'A handbook of Australian soils.' (Rellim: Glenside, South Australia)

Stott KJ, Gourley CJ (2016) Intensification, nitrogen use and recovery in grazing-based dairy systems. Agricultural Systems 144, 101 112. doi: 10.1016/j.agsy.2016.01.003

Thorburn PJ, Wilkinson SN (2013) Conceptual frameworks for estimating the water quality benefits of improved agricultural management practices in large catchments. Agriculture, Ecosystems & Environment 180. 192-209. doi: 10.1016/j.agee.2011.12.021

Thorburn PJ, Biggs JS, Weier KL, Keating BA (2003) Nitrate in groundwaters of intensive agricultural areas in coastal northeastern Australia. Agriculture, Ecosystems & Environment 94, 49-58. doi: 10.1016/S0167-8809(02)00018-X

Turner DA, Edis RE, Chen D, Freney JR, Denmead OT (2012) Ammonia volatilisation from nitrogen fertilizers in two cropping areas of southern Australia. Nutrient Cycling in Agroecosystems 93, 113-126. doi: 10.1007/ s10705-012-9504-2

Turpin JE. Thompson JP, Waring SA, MacKenzie J (1998) Nitrate and chloride leaching in Vertosols for different tillage and stubble practices in fallow-grain cropping. Australian Journal of Soil Research 36, 31-44. doi: 10.1071/S97037

van Herwaarden AF, Farquhar GD, Angus JF, Richards RA, Howe GN (1998) 'Haying-off, the negative grain yield response of dryland wheat to nitrogen fertiliser I. Biomass, grain yield, and water use. Australian Journal of Agricultural Research 49, 1067-1081. doi:10.1071/A97039

Virgona JM, Gummer FAJ, Angus JF (2006) Effects of grazing on wheat growth, yield, development, water use and nitrogen use. Australian Journal of Agricultural Research SI, 1307 1319. doi: 10.1071/AR06085

Wallace A, Armstrong R, Harris R, Belyaeva O, Grace P, Scheer C (2016) Nitrous oxide emissions from wheat grown in a medium rainfall environment in Se Australia are low compared to overall nitrogen losses. In 'Proceedings of the International Nitrogen Initiative Conference, Melbourne, Australia', (http://ini2016.com/conferenceproceedings-2)

Weaver TB, Hulugalle NR, Ghadiri H, Harden S (2013) Quality of drainage water under irrigated conditions of vertisols of the lower Namoi Valley, New South Wales, Australia. Irrigation and Drainage 62, 107-114. doi : 10.1002/ird. 1706

Wetselaar R, Hutton JT (1963) The ionic composition of rainwater at Katherine NT, and its part in the cycling of plant nutrients. Australian Journal of Agricultural Research 14, 319-329. doi:10.1071/AR9630319

Wetselaar R, Passioura JB, Rose DA, Jakobsen P (1973) Banding nitrogen fertilizers in soil: principles and practice. Chimie & Industrie - Genie Chimeque 106, 567-572.

Wichern F, Eberhardt E, Mayer J, Joergensen RG, Muller T (2008) Nitrogen rhizodeposition in agricultural crops: Methods, estimates and future prospects. Soil Biology & Biochemistry 40. 30-48. doi: 10.1016/j.soilbio. 2007.08.010

Handling editor: Sharon Aarons

J. F. Angus (A,C) and P. R. Grace (B)

(A) CSIRO Agriculture and Food, GPO Box 1700, Canberra 2601, ACT, Australia and EH Graham Centre, Charles Sturt University, Locked Bag 588, Wagga Wagga, NSW 2678, Australia.

(B) Queensland University of Technology, 2 George St, Brisbane, Queensland 4000, Australia.

(C) Corresponding author. Email: john.angus@csiro.au

Received 20 January 2017, accepted 30 June 2017, published online 21 August 2017

Caption: Fig. 1. Four classes of Australian land use, based on the system of ABARES (2010).

Caption: Fig. 2. Area of Australian dryland crops from 1850 to 2014. Data sources are Dunsdorfs (1956), annual issues of the Commonwealth Yearbook and ABARES publications.

Caption: Fig. 3. Changes in N-fertiliser use in Australia and the world. The sources are Angus (2001), Fertiliser Australia (www.fertilizer.org.au) and FAOSTAT (www.fao.org).

Caption: Fig. 4. Fate of [sup.15N] fertiliser applied in 74 Australian grain crops in aboveground plant parts and soil, both sampled at crop maturity and unaccounted [sup.15]N, reported as loss. The values are (means [+ or -] standard deviations) 44 [+ or -] 14% in crop, 34 [+ or -] 14% in soil and 22 [+ or -] 16% loss. Data sources are Australian experiments reported in Pilbeam (1996) as well as Bacon and Freney (1989), Smith et al. (1989), Freney et al. (1992), Armstrong et al. (1996, 1998), Robertson et al. (1997) and more recent studies by Lam et al. (2012), De Antoni Migliorati et al. (2014), Bell et al. (2015), Harris et al. (2015), Schwenke and Haigh (2016) and Wallace et al. (2016).

Caption: Fig. 5. Simulated N dynamics for dryland, non-legume crops from fertiliser, mined soil N and biological N-fixation by pasture legumes grown in the ley phase of a crop-pasture rotation, (a) N supply to crops (blank sections represent pasture phases (h) Soil total N. The simulations refer to phased rotations of 3 years of pasture and 3 years of crop, starting with total soil N of 3000kg [ha.sup.-1] (Dalai and Mayer 1986), biological fixation of 60 kg N [ha.sup.-1] during each year of pasture, annual net mineralisation of 3% of the native soil N during the cropping years (Table 2) and annual net mineralisation of the biologically-fixed N during the cropping phase, consisting of 17% in the first year after pasture, 6% in the second and 3% in all subsequent years (Angus and Peoples 2012). Fertiliser N amounts are from Table 4 and Angus (2001).

Caption: Fig. 6. Australian wheat production and total N-fertiliser use.
Table 1. Estimated balance (Mt [year.sup.-1]) of N inputs and
outputs in Australian regions in 2014, based on the methods of
Denmead (1990), Galbally et al. (1992) and McLaughlin et al.
(1992), with production and fertiliser quantities updated from
ABARES (2015) and land areas by ABARES (2010)

The estimated transfers represent (1) spatial N movement in dust
storms and (2) conversion of organic to mineral N, representing a
loss from the soil due to mining and input to the crop. Sources of
the estimates are shown in the supplementary information

Zone                                  Non-agricultural    Pastoral
                                         (309 Mha)        (355 Mha)

Input
  N in rain and dust                        0.6              1.2
  Biological N-fixation                     0.8              I.I
  Fertiliser N
N offtake in products
  Crop products
  Animal products (A)                                       <0.1
Losses
  Ammonia (B)                               -1.7            -2.1
  Denitrification
  N03 leaching and runoff
  Biomass burning (net N[O.sub.x])          -0.3            -0.2
Transfers (C)
  N in dust storms                      [+ or -]0.3      [+ or -]0.3
  Soil organic to mineral N
Balance                                     -0.6             0.0

Zone                                  Dryland farming    Intensive
                                         (97 Mha)         (4 Mha)

Input
  N in rain and dust                        0.3            <0.1
  Biological N-fixation                     3.2             0.2
  Fertiliser N                              1.1             0.4
N offtake in products
  Crop products                            -0.9            -0.3
  Animal products (A)                      -0.1            -0.6
Losses
  Ammonia (B)                              -0.6            -0.2
  Denitrification                          -0.3            -0.6
  N03 leaching and runoff                  -0.1
  Biomass burning (net N[O.sub.x])         -0.1
Transfers (C)
  N in dust storms
  Soil organic to mineral N             [+ or -]0.7     [+ or -]0.1
Balance                                     2.5            -1.1

Zone                                  Totals

Input
  N in rain and dust
  Biological N-fixation
  Fertiliser N                         8.9
N offtake in products
  Crop products
  Animal products (A)                  -1.9
Losses
  Ammonia (B)
  Denitrification
  N03 leaching and runoff
  Biomass burning (net N[O.sub.x])     -6.2
Transfers (C)
  N in dust storms
  Soil organic to mineral N
Balance                                0.8

(A) Empty liveweight of slaughter animals, milk and clean wool.

(B) Loss from plant communities, soil, urine and urea fertiliser.

(C) Not included in balance.

Table 2. Decreases in total N in the top 10 cm of soils in dryland
cropping systems in Australia. The half-life is calculated from the
equation [N.sub.t] = [N.sub.0][e.sup.-rt], where [N.sub.0] and
[N.sub.t] are the amounts of N at the start and end of observations
respectively, over a period of t years and r is the annual rate of
change in total N

Old, Queensland; SA, South Australia; NSW, New South Wales

Cropping system            Location                Years of
                                                 observations

Continuous sorghum         Narayan, Qld               10

Fence-line comparisons,    119 farms, six soil       1-70
cereals                    types, south Qld

Continuous wheat, no       Hermitage, Qld             14
N-fertiliser

Fallow-wheat, no           Waite Institute, SA        68
N-fertiliser

Continuous wheat, no       Waite Institute, SA        68
N-fertiliser

Continuous wheat, no       Wagga Wagga, NSW           18
N-fertiliser

Continuous wheat, no       Wagga Wagga, NSW           25
N-fertiliser

Continuous wheat, +50 kg   Wagga Wagga, NSW           25
N [ha.sup.-1]

Wheat-broadleaf with       Harden, NSW                19
tactical N

Continuous cereal, no      Theodore, Qld              23
N-fertiliser

Cropping system              Half-life     Reference
                           of soil total
                             N (years)

Continuous sorghum              18         Russell (1981)

Fence-line comparisons,        3-26        Dalai and Mayer (1986)
cereals

Continuous wheat, no            36         Dalai (1992)
N-fertiliser

Fallow-wheat, no                40         Grace and Oades (1994)
N-fertiliser

Continuous wheat, no            48         Grace and Oades (1994)
N-fertiliser

Continuous wheat, no            27         Helyar et al. (1997)
N-fertiliser

Continuous wheat, no            18         Heenan et al. (2004)
N-fertiliser

Continuous wheat, +50 kg        22         Heenan et al. (2004)
N [ha.sup.-1]

Wheat-broadleaf with            14         Angus et al. (2006)
tactical N

Continuous cereal, no           34         Dalai et al. (2013)
N-fertiliser

Table 3. Estimated N-fertiliser use for Australian agriculture, based
on estimated areas for 2010-2014 (ABARES 2015), national fertiliser
use in 2014 is from Fertilizer Australia (www.fertilizer.org.au) and
industry fertiliser use is from fertiliser-industry estimates

                                            Area         Average
                                            (Mha)     fertiliser use
                                                    (kg N [ha.sup.-1])

Dryland crops (A)                            24             45
Intensive farming
                    Cotton                  0.44           300
                    Dairy pastures            2            100
                    Irrigated cereals (B)   0.31           100
                    Sugarcane               0.36           150
                    Viticulture and
                      horticulture          0.50           100
Other
                    Sports-fields, parks
                      and gardens            0.1           200
                    Licks and stockfeed
Total

                                            Total fertiliser
                                               use (Mt N)

Dryland crops (A)                                 1.08
Intensive farming
                    Cotton                        0.09
                    Dairy pastures                0.20
                    Irrigated cereals (B)         0.03
                    Sugarcane                     0.06
                    Viticulture and
                      horticulture                0.05
Other
                    Sports-fields, parks
                      and gardens                 0.02
                    Licks and stockfeed           0.03
Total                                             1.59

(A) Wheat, barley, canola, sorghum, oats and triticale.

(B) Rice, maize and wheat.

Table 4. Nitrogen budget for an average Australian wheat crop,
updated from Angus (2001)

                                                     kg N [ha.sup.-1]

Crop N demand      Yield 2.0 t [ha.sup.-1], 10.5%
                     grain protein                          37
                   Straw N (one-third of grain N)           12
                   Rhizo deposited N (34% of total
                     plant N) (A)                           25
N supply           Fertiliser                               45
                   Rain and dust                            5
                   Mineralisation
                     Mining soil N                          31
                     N fixed from previous
                       pastures                             31
Soil-N retention                                            24
Losses (B)                                                  14

                                                          Totals
                                                     (kg N [ha.sup.-1])

Crop N demand      Yield 2.0 t [ha.sup.-1], 10.5%
                     grain protein
                   Straw N (one-third of grain N)
                   Rhizo deposited N (34% of total
                     plant N) (A)                            74
N supply           Fertiliser
                   Rain and dust
                   Mineralisation
                     Mining soil N
                     N fixed from previous
                       pastures                             112
Soil-N retention
Losses (B)                                                   38

(A) Wichern et al. (2008).

(B) Leaching, ammonia volatilisation and denitrification of
fertiliser and other N.

Table 5. Australian examples of N[O.sub.3.sup.1] leaching below
the root zone and groundwater contamination

Location                Source of N[O.sub.3]    Quantity

N[O.sub.3] leaching

South-east Queensland   Fallow--wheat with      19 kg N [ha.sup.-1]
                        summer rainfall         [year.sup.-1]

Mallee fallow--wheat    Mineralised N during    Accumulation >500 kg
sequence                fallows                 N [ha.sup-1]

Sugar deltas and        Fertiliser from sugar   30-50 kg N
coast                   land                    [ha.sup.-1]
                                                [year.sup.-1]

Western Australian      Fertiliser and          17-59 kg N
crop land               mineralised N           [ha.sup.-1]
                                                [year.sup.-1]

Southern New South      Fertiliser and          4 kg N [ha.sup.-1]
Wales                   mineralised N           [year.sup.-1]

Southern New South      Annual pasture          9-15 kg N [ha.sup.-1]
Wales                                           [year.sup.-1]

Southern Australia      Pastures where annual   15-35 kg N
                        rainfall >450 mm        [ha.sup.-1]
                                                [year.sup.-1]

Groundwater
contamination

South-east South        Annual pasture          7 mg N[O.sub.3]-N
Australia               legumes                 [L.sup.-1]

Central Australian      Holocene leaching       <80 mg N[O.sub.3]-N
groundwater             from termite mounds     [L.sup.-1]

North Queensland        Sugarcane               86% of wells <20 mg
(1454 wells)                                    N[O.sub.3]-N
                                                [L.sup.-1]

Location                Reference

N[O.sub.3] leaching

South-east Queensland   Turpin et al. (1998)

Mallee fallow--wheat    D Roget (unpublished)
sequence

Sugar deltas and        Quoted by Rasiah et
coast                   al. (2003)

Western Australian      Anderson et al.
crop land               (1998)

Southern New South      Poss et al. (1995)
Wales

Southern New South      Ridley et al. (2001)
Wales

Southern Australia      Ridley et al. (2004)

Groundwater
contamination

South-east South        Dillon (1988)
Australia

Central Australian      Barnes et al. (1992)
groundwater

North Queensland        Thorburn et al.
(1454 wells)            (2003)
COPYRIGHT 2017 CSIRO Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2017 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Angus, J.F.; Grace, P.R.
Publication:Soil Research
Article Type:Report
Geographic Code:8AUST
Date:Aug 1, 2017
Words:14090
Previous Article:Dilemma of nitrogen management for future food security in sub-Saharan Africa--a review.
Next Article:Global animal production and nitrogen and phosphorus flows.
Topics:

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |