Printer Friendly

Key crop nutrient management issues in the Western Australia grains industry: a review.

Introduction

Fertiliser costs represent a significant part of the variable costs of growing grain crops in the Mediterranean-type environment of Western Australia (WA). For example, the State average expenditure on fertilisers (S/crop ha) from the BankWest benchmarks increased from $62 to $83 over the period 2001-2006, and the Australian Bureau of Agricultural and Resource Economics (ABARE) farm surveys also indicated that fertiliser cost as a percentage of the total cash costs increased from 11 to 20% over 1990-2006 in the central and southern wheatbelt of WA (Allen Herbert, pers. comm.). The recent gross margins estimated by Allen Herbert (pers. comm.) for Merredin, WA, indicated that fertiliser cost was around $80/ha, representing 32% of the total variable costs, and this value could increase up to $120/ha in the high rainfall areas. The increased focus by grain growers on nutrient management due to rising fertiliser prices was also clearly reflected in a recent Grains Research and Development Corporation (GRDC) farm survey (NMS 00002).

During the early stages of agricultural development in WA, understanding and correcting deficiencies of phosphorus (P), copper (Cu), zinc (Zn), and molybdenum (Mo) through development of fertiliser practices and soil and plant testing calibrations dominated the research and extension, when wheat was the dominant crop in rotations with pastures, and regular cultivation was a key soil management practice. Much has changed in the WA wheatbelt since then. For some nutrients (P and Cu, Zn, Mo), soil reserves have been increased due to past fertiliser application. For others, such as potassium (K) and sulfur (S), exports in grain due to increasing yields have not been matched by inputs--for example, K fertiliser was not used until very recently and S inputs decreased as a result of the switch from using superphosphate to high analysis P fertilisers-leading to frequent deficiencies of these nutrients and increased risk of deficiencies of others, e.g. magnesium (Mg). Large areas of diverse crops are now sown usually with minimum-tillage (to replace a wheat pasture rotation). Economic pressures have pushed the cropping areas further west into the high rainfall zone (HRZ, >450mm). A wide range of fertiliser products is now available, and use of precision agriculture (PA) technology could add opportunities and complexity to the crop nutrient management (Chen et al. 2007).

With many changes in farming systems and farm production economy, it is now important to review assumptions about relevant nutrient management issues. The objectives of this work were to conduct a comprehensive analysis of the opportunities and threats associated with nutrient management in the WA grain-growing region, and provide a source of information from which decisions can be made about future directions of research, development, and extension.

Identifying the key crop nutrient management issues

A broad consultation with farmer groups, farm advisors, fertiliser companies, and researchers was undertaken to identify the key nutrient management issues for the WA grains industry. A detailed description of the methodology is provided in Chen et al. (2007). The consultation process involved face-to-face discussion, informal phone conversations and focus group discussion, and assessment of questionnaires. The objective of the face-to-face discussion, informal telephone conversations, and focus group discussions was to identify significant nutrient management issues to the grains industry for the follow-up survey assessment.

In the survey, 17 issues were outlined (Chen et al. 2007) for participants from different sectors associated with production, extension, research, and education of the grains industry, and they were asked to rate each in terms of impact and influence according to the following definitions:

* impact on the grains industry in the Western region if the issue was addressed successfully;

* influence that a project is likely to have in achieving that impact.

The 4 review topics rated in the survey (see Chen et al. 2007 for details) as having high impact and influence on the grains industry were:

* improving soil testing and interpretation;

* role of fluid fertilisers in the WA grains industry;

* use of spatial and temporal information to improve crop nutrient management, particularly for nitrogen (N);

* developing recommendations for managing emerging nutrient deficiencies.

Improving soil testing and interpretation

In WA, soil testing was initially developed primarily for the prediction of P requirements of wheat using soil surface (0-0.10m) sampling, at a time when thorough tillage of the 0-0.10+m layer was the common practice in grain cropping (Bolland and Gilkes 1990). A large body of soil analysis results has built up around sampling to 0.10 m depth during the summer, assuming that nutrients are well mixed in this layer, that most nutrients are acquired by plants from this part of the soil profile, and that a strong correlation exists between topsoil and subsoil nutrient availability.

More recently, however, it has been demonstrated that the soil test calibrations based on surface soils may not be reliable for nutrients such as K (Brennan and Bolland 2006a) and S (Anderson et al. 2006; Brennan and Bolland 2006b). In earlier cropping systems, fertilisers were applied and mixed with surface soils through 2-3 cultivations before crop establishment. When broadcast application of fertiliser was applied during the pasture phase, cultivation for the following crop mixed the fertiliser residues in the ploughed layer. However, in minimum tillage cropping which now occupies about 80% of the cropped area in WA (Crabtree 2004), fertilisers are usually banded in or near the row used to sow crops and are no longer thoroughly mixed with the top 0.10 m of soils. This shift in cropping practice has important implications for soil sampling and soil test interpretation. Similarly, adoption of disc openers for fertiliser placement, tramlining (controlled traffic farming), and wider row spacing have the potential to markedly change the horizontal distribution of nutrients in the top 0.10 m soil layer.

Various agencies make fertiliser recommendations based on soil testing information for the 0-0.10 m depth, irrespective of tillage system and row spacings. As farming practices and technologies are constantly changing, it is important to review soil testing in relation to the current farming systems, and identify any 'gaps' for further improvement in soil test based recommendation systems. Brown (1993) conducted a comprehensive review of soil sampling procedures in Australia. The review covered subjects related to soil sampling and sources of soil sampling variation, soil sampling equipment, soil sampling design and pattern, sample handling, etc. However, the implications of minimum tillage for soil sampling were not considered in that review. The present review focuses on those soil-test related questions identified as being important to the grains industry through consultation with stakeholders.

Sampling depth

In WA, the standard soil sampling depth to estimate plant available P, K and S is 0-0.10m (Chen et al. 2007). There has been increasing evidence (from published literature and local reports) suggesting that the optimum sampling depth needs to be reviewed, particularly on duplex soils. In reviewing effects of heterogeneous nutrient supply on root growth and nutrient uptake, Robson et al. (1992) suggested that on duplex soils, there could be advantages in using an increased sampling depth, and on these soils, more information on the distribution of nutrients available to plants with depth would be required to develop soil testing further. Subsoil sampling could also potentially identify soil constraints to root growth, as demonstrated in earlier studies where high levels of extractable aluminium in the 0.15-0.20m depth was identified as a constraint to wheat root growth (Carr et al. 1991).

Phosphorus

The Colwell (1963) sodium bicarbonate soil test procedure is widely used in WA. Many factors affect the Colwell soil P test calibrations. In early studies, effects of factors such as crop species, fertiliser type and history, soil cultivation, spatial soil variation, and temporal variation (sampling in November v. March) were evaluated in field studies (Bolland and Gilkes 1990; Bolland 1992, 1995a; Bolland and Wilson 1994; Chen et al. 2007). However, little specific information exists about the effects of soil sampling depth on soil P test values and calibration and on the accuracy of crop response predictions. Due to limited mobility of P down the soil profile, particularly in soils with high P sorption, it is generally assumed that the soil P test based on the top 10cm soil sampling depth adequately indicates soil P supply (Fig. 1) except on very sandy soils. Bolland (1995b) pointed out that soil P could leach on very sandy soils with a P retention index (PRI) of <2 mL/g soil in areas with average annual rainfall >450 mm. In these areas, over time, enough P applied in previous years could leach so that P levels increase at depth and could potentially be accessed by plants. For deep sandy soils, Colwell soil test P values may increase in years after P application to at least 0.40 m (Bolland and Brennan 2006). Bolland and Jarvis (1996) reported that on sandy soils where soil P could be leached down the soil profile, lupin yield response from banding fertiliser P below the seed could be reduced. While P leaching into the subsoil on very sandy soils is well established, the significance of subsoil P for crop P uptake is not well understood.

Apart from the Colwell P soil test procedure, the Mehlich 3 soil test package (Mehlich 1984) has been evaluated by the Chemistry Centre, WA (David Allen, pers. comm.). The test is designed to extract plant available forms of major soil cations (Ca, Mg, Na, and K), trace elements (B, Cu, Fe, Mn, and Zn), phosphate, and other elements such as cadmium and S. The evaluation studies (Bolland et al. 2003, David Allen, pers. comm.) showed that the Mehlich 3 results were highly correlated with the current soil test procedures. The Mehlich 3 soil test procedure could be adopted by commercial soil test laboratories if they are willing to invest in start-up costs particularly when a wide range of nutrients and elements needs to be determined in soil analysis. The current methods (Colwell, Olsen P, and Mehlich 3) for the assessment of plant-available P involve shaking and extracting the soil sample with a salt solution. The method of diffuse gradient in thin films (DGT, see Zhang et al. 1998 for details) may provide an alternative soil testing procedure for in situ soil assessment. In Australia, the DGT method was evaluated to measure soil available-P in agricultural soils (Mason et al. 2008). The initial pot trials using soils collected from South Australia and WA suggested that the DGT technique was more accurate in assessing early dry matter response of wheat to P application than the Colwell test (Sean Mason, pers. comm.). However, more research needs to be done before the DGT method could be used for routine soil testing and fertiliser recommendations for crop production in Australia. Although continuous research and development of new soil test procedures will help improving fertiliser recommendations for crop production, we believe that improving soil sampling and test interpretation based on the current soil test procedures will be more important for future research and development as it can deliver more immediate impacts to the industry.

[FIGURE 1 OMITTED]

Potassium

The method for making K fertiliser recommendations in WA is based on bicarbonate-extractable K (Colwell extraction) of the surface soil (0-0.10m) (Edwards 1998a). Although surface soil testing is a useful tool for monitoring soil K levels, the accuracy of prediction of yield responses may vary, particularly in situations where plants obtain a significant part of their total K requirements below the sampling depth. Potassium supply from the subsoil can be important, as root distribution down the soil profile will vary with plant species, subsoil constraints, and subsoil K contents. Hamblin and Hamblin (1985) reported that lupins and wheat had <50% of their total root length in the top 0.20 m, while annual pasture species such as Trifolium and Medicago had 70% of roots in that layer. In soils with subsoil K supply, spring wheat obtained 34% of its total K requirements from the subsoil (Kuhlmann 1990). The contribution of the subsoil to uptake of K increased with the development of the crop, from 8% at first node stage to 35% at ear emergence, as the proportion of total root length in the subsoil increased (Kuhlmann 1990).

Assessing soil capacity to supply K based on the soil samples taken from the 0-0.10m layer, may work well on deep sands where most extractable K is associated with organic matter and there are often low levels of extractable K in the subsoil (Wong and Wittwer 1997). However, the 0-0.10m sampling depth on duplex soils may be less reliable. Duplex soils form a large part of the soils used for cropping in WA (Edwards 1997). Potassium concentrations in the heavier textured subsoils vary greatly due to differences in clay content and mineralogy (Edwards 1997). In the early 1990s, soil sampling indicated that many of the duplex soils used for wheat production had low soil K levels in both surface and subsurface soil layers, because on these soils, the dominant clay mineral is kaolinite, which naturally has a low K content (Edwards 1997). However, for some duplex soils that have subsoil clay enriched with K due to the presence of K-bearing minerals, published information has been inconsistent with regard to interpretation of the soil K test. In field experiments conducted in the central wheatbelt from Beverley to Katanning (1993-1996), Wong et al. (2000a) reported that subsoil K was not important in determining wheat yield response to K fertiliser and concluded that K deficiency could be predicted in these soils using 0-0.10 m samples only. In the study of Wong et al. (2000a), no grain yield response to K fertiliser application was recorded at 2 sites (Kweda and Meckering), with both having high extractable K below the 0.10 m soil layer. In a more extensive study where field experiments were conducted at 56 sites across south-western Australia, Brennan and Bolland (2006a) reported that soil testing based on the 0.10 m sampling depth accurately predicted deficiency or sufficiency of soil K for canola grain production at 50 of the 56 sites. For the remaining 6 sites, K was inadequate in the topsoil but adequate in the subsoil and the current soil test procedure incorrectly predicted K deficiency (Fig. 1). Therefore, it is clear that further investigation is needed to define soil types and locations where subsoil K levels need to be determined by soil analysis before recommending K fertiliser.

Sulfur

Probert and Jones (1977) found that soil test S in Australian tropical soils was often lower in the topsoil than in the subsoil. Subsoils had greater capacities to sorb sulphate, thereby retaining S in a soil layer that could be explored by plant roots. Hence, the S status of the subsoil was important for growth of tropical legumes. This may also be the case for duplex soils in WA. The duplex soils typically have a higher clay and aluminium oxide content in the subsoils (McArthur 1991) and thus have a capacity to sorb sulfate, depending on pH. Anderson et al. (2006) found that wheat and particularly lupin grown on grey duplex soils derived a large amount of S[O.sub.4.sup.2-] from the soil layers below 0.10m. Brennan and Bolland (2006b) analysed data from 59 field experiments and found that in about half of the sites, soil test S increased with soil depth. For soils with low soil test S in the top 0.10m but high test S in the subsoils, canola plants showed S deficiency symptoms during early growth, but the S deficiency disappeared as roots grew deeper into the soil layer with adequate S, and final canola yield was not affected by the temporary S deficiency. Shallow soil sampling (top 0.10m of soil) overestimated the prevalence of S deficiency for canola grain production; thus, deep soil sampling needs to be considered for canola (Fig. 1). Recovery from early S deficiency was also observed for wheat (N. Edwards, unpublished data) and hence suggests a more general need for subsoil sampling to improve prediction of S deficiency.

Soil sampling methods in zero-till cropping systems

Soil sampling techniques have been based on the assumption that nutrients and other soil characteristics are randomly distributed properties that can be evaluated by random sampling (Kitchen et al. 1990). This assumption may not be valid in zero-till cropping systems due to nutrient stratification resulting from lack of soil mixing, fertiliser banding in seeding rows, and stubble retention. Trends towards wider row spacing, the use of disc openers for seed and fertiliser placement, and tramlining may further exacerbate the horizontal heterogeneity of nutrient levels in the 0-0.10 m soil layer.

Nutrient stratification

Significant vertical stratification of soil extractable nutrients such as P and K in long-term no-till fields has been reported in North America (Howard et al. 1999). Lack of soil mixing, surface broadcasting of fertilisers, high crop residue concentrations at the soil surface, and limited mobility of P and K in soils are all contributing factors to vertical soil stratification of nutrients in no-till systems (Howard et al. 1999; Selles et al. 1999). Long-term no-till management has also resulted in pronounced horizontal heterogeneity of P and K (Mallarino and Borges 2006), particularly where crop rows are repeatedly planted on or close to preceding rows.

Crops in WA are often sown with much narrower row spacing (~0.20m or less) compared with crops grown in North America. This may be why little research on the potential impact of no-till on soil nutrient distribution and its implications for soil testing and interpretation has taken place in WA. However, surface soil nutrient stratification (extractable P and K, total N and organic C) was reported in reduced tillage systems of WA (White 1990) and New South Wales (Cornish 1987) in earlier studies. In recent field experiments where crops had been sown using no-till for 7-11 years, Bolland and Brennan (2006) reported that soil nutrient values for in-row samples increased as the amount of fertiliser applied increased and were about double the values for the samples collected half-in and half-between rows (Fig. 2) in the second year after 3 different methods of applying P, Cu, and Zn (drilled, deep banded, top 0.10m of soil cultivated before fertiliser drilled) on wheat (sown in 24-cm rows) at 16 locations. Similar soil nutrient distribution patterns were also observed in the long term no-till cropping soils of the northern wheatbelt of WA (Thompson 2003).

[FIGURE 2 OMITTED]

Root growth and activity in the surface layer are more vulnerable to drought than those in subsurface layers (Cornish 1987). Hence, surface nutrient stratification could result in less plant nutrient uptake and thus increase the likelihood of nutrient deficiency in crop tissues, and may also cause yield loss in growing seasons with periods of drought that coincide with periods of high crop demand for nutrients. Nutrient stratification both vertically and horizontally resulting from no-tillage and fertiliser row placement as discussed above will have important implications for obtaining a representative soil sample from which to determine soil nutrient levels. This effect may become more pronounced if wide row spacing for lupin (>0.25 m) is adopted by growers in WA.

Soil sampling

Soil sampling techniques that account for band placement of fertiliser in assessing field nutrient status have been evaluated with a focus on less mobile nutrients such as P and K. Kitchen et al. (1990) used 2 sampling methods to collect soil samples where fertiliser was banded in rows. The techniques were based on the knowledge, or lack of knowledge, of band location within the field. A modified exponential decay model was used to describe residual P bands:

Y = A exp(-Bx) + C

where Y is the soil-test value, X is lateral distance from the band, A and B are curve-fitting parameters, and C is soil-test concentration when unaffected by the band.

Using the above procedures, Kitchen et al. (1990) suggested that in a situation where a residual P band was obvious, a ratio of 1 : 20, 1 : 16, and 1 : 8 in-the-band cores to between-the-band cores could be considered for 0.76, 0.61, and 0.30m band spacing, respectively. However, in a situation where the location of the P bands was unknown, random sampling was the only alternative. Tyler and Howard (1991) reported that the representative nutrient status of a soil having banded fertiliser could be evaluated using a random sampling technique, although random sampling of a band (a non-random variable) would require an impractically high number of soil cores to accurately and precisely represent soil test P levels (Ashworth et al. 1994; James and Hurst 1995; Stecker and Brown 2001). The transect method has also been suggested by several researchers (Ashworth et al. 1994; James and Hurst 1995) as a means to overcome the large variability and uncertainty of band location that results from banded P. It involves sampling a slice of soil across the row to include banded and non-banded soil. However, the reliability of the transect method of soil sampling for predicting fertiliser responses relative to other soil sampling methods has not yet been evaluated.

In summary, there are gaps in our knowledge of optimal soil sampling procedures when plants can acquire significant amounts of major nutrients (P, K, and S) in subsoils, and where nutrient stratification resulting from adoption of zero-till cropping systems produced horizontal and vertical heterogeneity of nutrient levels.

Role of fluid fertilisers in the WA grains industry

The use of fluid fertilisers is increasing in Australia. The Fertiliser Industry Federation of Australia estimates that fluid fertilisers represented 2% of total fertiliser use (2002 data) but about 8% in WA (Wylie et al. 2003). By 2005 in WA, the use of fluid fertilisers had increased to 16% of the total N fertilisers (FIFA, unpublished data). In WA, the main interest has been in the application of fluid N. The use of fluid N fertilisers is increasing because the cost per kg N in fluid N fertilisers in WA is only marginally higher than solid (granular) fertilisers, unlike in other States where the cost of N in fluid formulations is much higher than urea (Wylie et al. 2003). In South Australia, research demonstrated more efficient uptake of P and improved yields of wheat using fluid P fertiliser on calcareous soils (Wylie et al. 2003). Currently, about 80 farmers use fluid P fertilisers and this number could expand to 3000 farmers, who crop 2 Mha of calcareous soils in South Australia, Victoria, and small areas of WA.

Crop response to applications of fluid P across different soil types in Australia

The use of fluid fertilisers at seeding as sources of N, P, and Zn increased grain yield in the upper Eyre Peninsula, South Australia, on a calcareous red brown sandy clay loam during 1993-1995 (Holloway et al. 2001). In South Australia, 92 studies/field experiments compared fluid with granular fertilisers using 3 basic groups of fluid fertilisers--ammonium polyphosphates (phosphoric acid-based products) and technical grade mono-ammonium phosphate (MAP) or di-ammonium phosphate (DAP)--dissolved in water. That research was conducted on 3 soil types, grey highly calcareous sandy loams with 15-70% calcium carbonate, red-brown calcareous sandy loams with 5-15% calcium carbonate, and red-brown loamy sand with 1-5% carbonate and at low N fertility. Sixty of these comparisons showed grain yield increases due to fluid P applications, 4 showed yield decreases, and 27 had no differences (Holloway et al. 2004). Using the soils collected from Victoria and South Australia, McBeath et al. (2005) reported that fluid P fertilisers were likely to produce greater wheat biomass than granular products on calcareous soils. Overall, CaC[O.sub.3] content was a significant factor influencing whether fluid fertiliser would be more effective in improving crop dry matter production and P uptake than a granular form of P. Recent analysis (McBeath et al. 2007) further confirmed that fluid P fertiliser could produce greater wheat biomass than granular P in any soils with >5% CaC[O.sub.3] in the topsoils.

In WA, there have been fewer field studies to compare fluid with granular sources of P in improving crop yield. However, unlike South Australia, grain yield advantages using fluid P have not been consistently observed in WA. CSBP Ltd conducted 28 trials to compare fluid with granular sources of P over the period 2000-2005; Loss (2006) reported that fluid P fertilisers resulted in an 8-31% yield increase compared with conventional granular fertilisers at only 6 sites of the 28 trials (Table 1). All responsive sites' soils had high P sorption capacity. Field trials conducted by Summit Fertilizers (S. Alexander, pers. comm.) also showed grain yield benefits of using fluid P over granular P at equivalent P rates on a high P-sorbing soil. The WA work suggests that on soils with low Colwell P and high P sorption, use of fluid P fertilisers could produce greater crop yield than granular fertilisers. As little soil information is presented in these field studies, it is not yet possible to draw clear conclusions on key soil properties controlling crop yield response to fluid P in WA soil conditions.

Glasshouse experiments were undertaken to understand the mechanisms by which acidic fluid fertilisers (phosphoric acid) enhance P uptake (McLaughlin 2002; Lombi et al. 2004; Bertrand et al. 2006). The results indicated that the greater ability of fluid forms of P to improve the growth and P uptake of wheat was mainly due to decreased rates of chemical 'fixation' of P in the soil (available chemical form) and increased mobility (diffusion rate) of P compared with granular forms (see Chen et al. 2007 for details). Although crop yield benefits of using fluid P in calcareous soils have been well documented, our understanding of crop response to fluid P in acid soils is still inadequate. Further research is needed if fluid P technology is to be adopted by WA growers under mostly acid soil conditions, but preliminary results suggest that this work should focus on P sorption characteristics of soils.

Grain yield and protein benefits of using fluid v. granular N

The trials conducted by CSBP (Loss and Appelbee 2006) showed that in 107 of the 132 comparisons of urea ammonium nitrate and urea granules at the same rate of N and time of application, there was no difference between fluid N and urea for wheat and canola grain yields (Fig. 3). The results from the 22 trials conducted by CSBP suggested that compared with boom spray application of fluid N, banded fluid N improved N use efficiency, which could lead to a decrease in N application rate while still maintaining the same grain yield (CSBP 2005).

Many areas in WA produce wheat grain with low protein content due to factors such as decreased proportion of legume crops in the rotation, leaching of N applied at seeding (especially in wet years), and decline of soil N reserves (CSBP 2003). The availability of fluid N products offers new opportunities for farmers to effectively manipulate grain protein content based on seasonal conditions, particularly in WA where the price of the fluid N products is low compared with other States of Australia. Late applications of fluid N could be used to enhance grain protein content, as this generally increases protein more than yield since there is limited dilution of the protein produced in grain. Nitrogen accessed late in the season may be more effectively translocated to the grain rather than to vegetative parts (Grant et al. 2001). However, timing of N supply must still be early enough in plant development to allow sufficient time for plant uptake and translocation to grain. The trials conducted at Buntine and Kojonup showed that in high-yielding situations (4-4.5 t/ha), late application of fluid N could be used to boost wheat protein content (CSBP 2003). However, late applications of urea or urea ammonium nitrate at flowering to boost wheat crop grain protein content could be risky in WA climatic conditions where flowering often coincides with periods of water deficit that could affect the capacity of crops to utilise N applied (Fillery 2004). Loss (2004) suggested that in dry conditions, fluid N could be absorbed through the leaves and should provide a more reliable protein increase than top-dressed urea on soil, as N could only be utilised to increase protein content when significant rain leaches the urea into the rooting zone. It is very clear that more work is needed to better understand the impact of late N application on N dynamics, crop protein content, and financial return.

[FIGURE 3 OMITTED]

Using spatial and temporal information to improve crop nutrient management, particularly for nitrogen

Role of precision agriculture research in understanding crop yield variability and improving crop nutrient management with particular reference to Western Australia

Yield variabiliy, causes, and delineating management zones

Yield monitoring was introduced to Australia in 1993. Initial yield maps revealed a degree of variation in grain yield that surprised farmers but also encouraged more to buy yield-monitoring equipment (Dobermann et al. 2004). Mapping through yield monitors revealed the variation and promoted the possibility of controlling it through the development of PA technologies (Cook and Bramley 1998; Price 2005).

Subsequently, there has been little uptake of PA in the Australia grains industry. For example, it is estimated that only 3% of Australian grain growers use some form of PA technology (Price 2004). Of the growers who operate this technology, not all practice spatially variable management. In WA, some 20-25 farmers use variable rate fertiliser applications (Ian Maling, pers. comm.). Some of the uncertainty surrounding its adoption is whether enough within-paddock variation in yield exists to justify the investment. Robertson et al. (2006a) surveyed 200 yield maps of wheat obtained from a wide range of locations and seasons in the northern sandplain region of WA. Yield varied from 0.5 to 4.9t/ha due to within-paddock and seasonal variations. Similarly, Wong and Asseng (2004) reported that yield mapping for >5 years in Three Springs, WA showed large spatial variability with grain yield, ranging typically from 0.4 to 4 t/ha within a paddock. Both studies in WA suggest that variation in grain yields is sufficient to justify variable rate fertiliser applications.

Diagnosis of causes of yield variability is important for growers to make management decisions on how to respond to the variability. There are many causes for crop yield variability in a field, these include inherently variable water-holding capacity of the soil, subsoil constraints (compaction and B toxicity), soil acidity (A1 toxicity), nutrient deficiency, pests and diseases, and soil and landscape features interacting with seasonal conditions (frost and waterlogging) (Patabendige et al. 2003). In the Mediterranean-type environment of WA, plant-available water storage capacity (PAWC) of a soil is the most important yield determinant and it can vary from 30 to 150 mm across a field due to differences in soil texture and root distribution (Wong et al. 2006). Wong and Asseng (2006) reported that the main cause of temporal and spatial wheat yield variability within the field was due to interactions of seasonal rainfall, PAWC, and N fertiliser applications. Spatial variability was low in low rainfall years when yields across the field were generally low. With adequate N, spatial variability increased with seasonal rainfall as sites with higher PAWC conserved more water in wet seasons to give higher yield response than sites with low PAWC. Knell and Slade (2005) also found that unstable yield zones were associated with soil conditions such as waterlogging, subsoil acidity, and subsoil compaction.

Whelan and McBratney (2000) reported that temporal variability in crop yield at the within-field scale was often larger in magnitude than spatial variability, and thus suggested that this could increase the risk of economically and environmentally inappropriate actions if differential treatments were solely based on spatial information. This highlights the importance of considering both spatial and temporal variability when adopting PA technologies (McBratney et al. 2005). Robertson et al. (2006a) found that crop yield potential and soil fertility status within a paddock were driving the potential economic benefits that can be gained from zone management through use of PA technologies. With the aid of the APSIM simulations, Robertson et al. (2006a) reported that in an 'average' season, potential yield might vary by 3 t/ha (0.9-3.8 t/ha) across zones with different PAWC, but in a below-average season, the variation might only be 1.3 t/ha, thus diluting the potential benefits for zone management. In above-average seasons the range might increase to 3.6t/ha (Table 2). Clearly, using a modelling approach could offer an opportunity to take into account the effect of seasonal variability on yield performance when defining management zones.

Principal spatial data layers used in delineating management zones include one or more of the following: yield maps, soil maps (including electrical conductivity), remote sensing image, digital elevation models, farmers' knowledge, and management history. In WA, Silverfox Solutions Pty Ltd has been providing commercial services to researchers and farmers to identify management zones using normalised difference vegetation index (NDVI) or yield maps. This requires 3-5 years' data for the crop of interest (mainly cereals) (Ian Maling, pers. comm.). Before a paddock is considered suitable for zoning, at least 60% of the paddock must have behaved in a stable/ consistent manner through time. Adams and Maling (2005) indicate that in situations of consistent performance of NDVI, it can be used with confidence over a large area of a farm to define management zones. The zones have different crop yield potential, which is then modified by seasonal conditions.

Applications of PA technologies in nutrient management

Nitrogen is mostly applied at a uniform rate across an entire field even though actual requirements are known to vary substantially at this scale due to differences in potential yield (demand), soil N status, N mineralisation, and the efficiency of fertiliser use (supply). One way to increase N-use efficiency is therefore to adjust fertiliser inputs according to site-specific conditions, taking into account the within-field variability of supply and demand. Adams et al. (2000) reported that yield maps in WA provided useful information about spatial variation of expected achievable yield, and yield maps together with computer modelling could be used for deriving site-specific N recommendations.

In WA, spatial technologies have also been used for mapping risk of nutrient deficiencies (boron, B) and site-specific nutrient requirements (K). Wong et al. (2005) mapped B deficiency risk in soils of WA using a weight-of-evidence model. This work provides useful information on B deficiency risk in soils of WA, especially those acid and sandy soils developed on sandstones, and suggests that the risk of B deficiency will become important if these soils continue to be cropped with oilseeds and legumes and without B input from fertilisers. The scale of this mapping is smaller than required for PA but the approach could be applied to map deficiency risk of other nutrients (e.g. Mg). Wong et al. (2001) used soil sampling and yield information from yield monitors and NDVI measurements for mapping the spatially variable K requirements of crops. The work has not yet been adopted in the existing commercial fertiliser recommendation systems. However, it provides a good example of developing cheap methods for mapping soil nutrient supply at a paddock scale that could be used in PA.

Techniques for mapping field variability and crop monitoring

The field factors affecting crop yield variation can be broadly grouped into 4 categories. They are soil-water factors, topographic factors, soil nutrition factors, and crop growth and health factors (Table 3).

The PAWC is an important yield determinant in the Mediterranean environment of WA, but direct measurement of spatial variation of PAWC is time-consuming and costly. Apparent electrical conductivity (ECa) values of land by EM38 are related to clay content, organic matter content, and soil water content (Godwin and Miller 2003). When soils are near their field capacity, soil moisture content and bulk density are largely governed by clay content. Thus, EM38 measurements when soils are near field capacity could be used indirectly to assess PAWC. Wong et al. (2006) conducted a survey using a Geonics electromagnetic induction instrument (EM38) when soil was wet, and PAWC was also directly measured at the survey locations. They reported that there was a significant linear relationship between ECa and PAWC, and the established relationship between PAWC and ECa was then applied to obtain a high spatial resolution (5-m grid) PAWC map of the field. In another study, Wong et al. (2006) reported that salt-affected areas resulted in larger ECa values that were not related to PAWC. In these areas, gamma-emission from [.sup.40]K could be used to estimate relative PAWC. These studies suggest that it is likely that a combination of ECa and gamma-emission (or elevation) will be required to relate sensing data to targeted PAWC measurements on a farm. Research conducted in Europe (Godwin and Miller 2003) also indicted that EM surveys could provide rapid, non-invasive information on the variation in soil texture and available water. In the WA wheatbelt, farmers currently make little use of soil sensing technologies, partly due to lack of infrastructure and support, but the mining industry has a lot of experience using soil sensing and this could provide a service for agriculture (Robertson et al. 2006b).

Soil nutrition problems at paddock scale could be addressed by designing the soil sampling regime based on maps of crop performance or on information from soil sensing technologies. In Europe, the NDVI data derived from the red and near-infrared bands of a SPOT satellite image provide useful information to design targeted soil sampling (Godwin and Miller 2003). In WA, Silverfox Solutions has been working with growers using NDVI-derived management zones to direct targeted soil sampling (Ian Maling, pers. comm.). The company has also been working with CSIRO and DAFWA to explore the possibility of identifying areas where crops may be affected by specific nutrient deficiencies, frost, and subsoil problems by closely studying the relationships between yield maps and NDVI image maps. Robertson et al. (2006b) reported that the use of mid-season NDVI and yield mapping could provide useful information on the causes of low- or high-yielding areas. For example, low-yielding areas with high mid-season biomass might suggest a subsoil constraint problem or rapid depletion of stored soil water, and this information will be useful to conduct targeted soil sampling to confirm the exact cause. In WA, soil sensing technologies have not yet been widely used for mapping soil available nutrient status. In one example, Wong and Harper (1999) used gamma-ray spectrometry to estimate plant-available K (Colwell-K) but they found that the success of using spectrometry for predicting available K depended on a strong relationship ([r.sup.2] = 0.9) between total K and available K, which may not hold in all areas since it depends on soil mineralogy and organic matter levels.

Monitoring the physiological status of crops spatially and their potential response to tactical interventions such as in-season N application is essential for growers to adopt site-specific nutrient management and variable rate technologies. In WA, little research has been conducted to evaluate the usefulness of crop monitoring techniques that could be used for site-specific crop management. Rodriguez et al. (2005) reported that digital thermal imaging could be used for assessing the physiological status of wheat crops as affected by water and N. Developing methods for crop quality assessment is equally important to those who monitor crop growth. For example, estimating grain protein at growth stages early enough for effective intervention would be an incentive for growers to adopt PA. Skerritt et al. (2002) confirmed that within-field variation in protein content and protein quality of wheat in WA was often very significant, and they also found that areas of higher yield usually did not have lower protein content or protein quality. Potential for predicting final grain protein using reflectance measurements obtained with a hand-held instrument over the wheat canopy around flowering time was reported in Queensland (Kelly et al. 2005). Further research and development of the methods for crop monitoring and crop quality assessment will certainly help the adoption of PA.

Tactical nitrogen management in cereal crops

There has been renewed interest in improving fertiliser N use efficiency in WA due to the rising cost of N fertiliser, coupled with the availability of new fertiliser technologies for growers (fluid N fertiliser and PA). High input of N is reported to be critical to achieve targeted crop yield potential in the HRZ of WA (Hill 2005). The different soil N dynamics and crop yield potential in the HRZ means that N management practices developed in the traditional wheatbelt may not be directly applied to the HRZ (Chen et al. 2007). Fertiliser N use efficiency is still low in Australia. In response to the overall problem of low fertiliser N use efficiency, several comprehensive reviews have been published recently to critically review the existing literature and explore options for improving fertiliser N use efficiency in cereal crops (Raun and Johnson 1999; Cassman et al. 2002; Dobermann and Cassman 2004; Giller et al. 2004; Dobermann 2005; Fageria and Baligar 2005; Ladha et al. 2005; Rengel 2005; Chen et al. 2008). These reviews concur on the importance of using advanced spatial and temporal information and advanced fertiliser technologies to improve N use efficiency.

Nitrogen management practices and recommendation systems for cereal crops in WA

Nitrogen fertiliser is now considered one of the most costly inputs into cereal production systems in Australia. According to the national assessment of fertiliser inputs in Australian agriculture (National Land and Water Resources Audit 2001), annual application rates range from 1 to 80 kg N/ha across the dryland cereal production areas in southern Australia, and in some areas under irrigation, application rate could be up to 200 kg N/ha. In WA, all soils are potentially N-deficient, thus use of N fertilisers is important to sustain crop yields (Mason 1998). Use of N fertiliser on cereal crops in WA began in the late 1960s and the use grew with the development of NP compound fertilisers in the 1970s. Urea replaced ammonium nitrate as the major N source in the mid 1970s. Granular urea is now the most commonly applied N fertiliser and it is also the least expensive in cost per unit N (compared with MAP and DAP). Recently, more growers have started using fluid N. Rate of N application varies according to many factors such as crop species, rainfall, soil type, paddock history, and time of sowing. In WA, the rates used by growers generally range from 0 to 30kgN/ha in the low rainfall areas (<300 mm), from 20 to 40 kg N/ha in the medium rainfall areas (300-400 mm), and from 40 to 80 kgN/ha in the high rainfall areas (>400mm) (Anderson and Hoyle 1999). However, in the HRZ of WA, 160kgN/ha was required in experimental plots to achieve targeted wheat yield of 5-6 t/ha (Hill 2005).

Early work in WA (Halse et al. 1969; Mason et al. 1972) indicated that time of fertiliser N application significantly affected the potential for wheat grain yield by altering the production and survival of tillers. Similarly, McDonald (1989) in a review of Australian crop N nutrition research concluded that the most appropriate time for post-sowing N applications was during the period from tillering to the commencement of stem elongation. Later applications (after stem elongation) resulted in little or no grain yield response but usually increased grain protein content. In WA, splitting N fertiliser application has been practiced in the medium and high rainfall zones since the introduction of NP compound fertilisers in the 1970s, which allowed growers the flexibility to drill the compound fertiliser at seeding and then apply urea after seeding (Chen et al. 2007). The wide adoption of minimum tillage and introduction of early maturity varieties in WA have increased the opportunity for obtaining economic response to N applied late in the season (Bowden 1999). Although splitting N was adopted early in WA, there has been little research on the information growers are currently using to make the decisions on splitting N applications, particularly in response to changing cropping systems and new fertiliser and application technologies.

Mason (1994) developed an N recommendation system for wheat based on 1000 trials completed over many years. In this N recommendation system, Mason (1994) used rainfall zone (A <330mm, B=330-460mm, C >460mm), soil type (heavy or light land), and paddock management history (cropping or pasture history, stubble management, etc.) to categorise N requirement situations. Burgess et al. (1991, 1992) developed a simple computer program (NPDECIDE) based on many years' field studies, to help farmers and advisors to make decisions on N and P fertiliser use for cereals. The concept and framework adopted by NPDECIDE has led to further development of the N decision tool 'Select Your Nitrogen' (SYN) (Diggle et al. 2003). SYN calculates the N available to the crop from different N sources (organic N, residue organic N, and fertiliser N) using NAVAIL. It then calculates actual yield and grain protein based on the available N and the potential yield of the crop (Diggle et al. 2003). The development of computer-based and quantitative N decision-support systems have helped extension of research information on crop N nutrition and management to the grain industry, and the development of a commercial computer-based tool 'SoilMat' in WA for making pre-seeding N recommendations through the CSBP Nulogic service (W. Pluske, pers. comm.).

In WA, the availability of fluid N could increase the opportunity to manipulate crop protein in response to seasonal and crop conditions, particularly in the HRZ. Furthermore, while the above N fertiliser recommendations on cereals were based on an extensive experimental program, cereal cropping systems have recently undergone substantial changes including 80% adoption of minimum tillage and stubble retention by growers, decreased use of pastures in rotations, earlier sowing and harvest dates, higher yield potential, and increased emphasis on grain quality. It remains unclear how significantly these changes are affecting the accuracy of current N fertiliser recommendations.

Nitrogen management in high rainfall zone

The HRZ in WA is defined as the cropping areas where annual average rainfall is 450-800 mm with a growing season length of 7-10 months (Zhang et al. 2006). In these areas, 864 000 ha is sown to annual grain crops. Potentially, about 2.1 Mha could be used for annual cropping in this zone (Hill 2005). In the HRZ, average crop yield from 1996 to 2001 was 2.7, 2.4, and 1.4t/ha, respectively, for wheat, barley, and canola (Hill 2005). These yields are only about 50% of the potential yields (defined as water-limited yield potential) estimated using the APSIM model (5-6 t/ha for wheat, 3-4 t/ha for canola; Zhang et al. 2006). Hill (2005) reported that the constraints for crop production in these areas were poor soil drainage, inappropriate crop rotations, poor weed control, inadequate nutrition, ineffective control of insects and disease, inappropriate sowing time and seeding rate, and poorly adapted varieties. The interest in cropping in the HRZ has been growing in WA, and recent work in the HRZ reveals that the major knowledge gaps are subsoil constraints and poor crop nutrition and their impact on crop yields (Hill 2005).

Better N management in the HRZ is important to achieve a potential yield of 6t/ha in wheat and 3.5 t/ha in canola. A baseline study for the HRZ showed that farmers applied only 50-70 kg N/ha for wheat, barley, and canola, representing only about half of the N required for the targeted crop yields (Hill 2005), notwithstanding the fact that N demands for different crops can vary significantly among years due to the variation in rainfall. Zhang et al. (2006) indicated that wheat and canola could respond up to 100-120 kg N/ha in the HRZ, beyond which the response of yield was minimal. High N application can stimulate early vegetative growth and water use, deplete soil water reserves, induce water stress during grain filling, and thus lead to 'haying off of wheat crops, in low and medium rainfall areas (van Herwaarden et al. 1998). However, in the HRZ, Zhang et al. (2006) examined the existing research data and concluded that targeting higher optimum dry matter at anthesis using moderate or high levels of N application was unlikely to cause yield penalties by haying off. Splitting N applications according to soil water and weather conditions compared with all N applied at seeding could improve crop yield (Table 4) and eliminate the effect of waterlogging, which has been perceived to be a major limitation to crop production (Hill 2005). Nitrogen fertiliser application after waterlogging events has been shown to reduce the detrimental effects of waterlogging (McDonald 1989). Application of some N after waterlogging is one of the methods to increase yield and correct N deficiency by replacing N lost during the waterlogging period as result of denitrification and leaching (Fillery and McInnes 1992).

Opportunities for improving nitrogen use efficiency for cereal crops in WA

Agronomic efficiency, the conversion of applied fertiliser N into grain yield, is a common benchmark for assessing the efficiency of N fertiliser in cereal production systems, as is apparent recovery, the proportion of applied fertiliser recovered in grain. McDonald (1989) reported that agronomic efficiency for rainfed wheat in Australia ranged from -4 to 43 kg grain yield/kg N applied, with apparent recovery from 6 to 78%. In WA, Mason et al. (1972) reported that agronomic efficiency of rainfed wheat was only 3-13 kg grain yield/kg N with apparent recovery of 21-48%. Ladha et al. (2005) reviewed N-use efficiency on the global scale for 3 major cereal crops--maize, rice, and wheat--and reported that in Australia, average agronomic N-use efficiency was 8 kg grain yield/kg N applied, with apparent recovery of 54% for cereal crops. Low N-use efficiency is a common problem for many cereal crop production systems in the world, thus it is important to understand factors affecting fertiliser N-use efficiency and explore options for improving it.

The factors controlling fertiliser N-use efficiency can be classified into 3 groups: (i) factors controlling crop demand for N, (ii) factors controlling supplies of plant available N from soil and fertilisers, and (iii) factors controlling N losses from soil-plant systems (Ladha et al. 2005). Improvement of N-use efficiency in crop production systems needs to focus on achieving synchrony between crop N demand and the N supply from all sources throughout the growing season and at the same time reducing N losses. Cassman et al. (2002) suggested that developing better crop and soil management practices would be more important than crop genetic improvement in improving N-use efficiency. They further suggested that 'precision management' in time and space of the N demand/supply/loss factors, to maximise the synchrony between crop-N demand and the supply of mineral N (from soil reserves and N inputs) while minimising N losses, was critical to achieve N use efficiency in cereal production systems.

Foliar N application (supplementing soil applications) could be considered to improve synchrony between crop-N demand and the supply of N in time. For example, Bly and Woodard (2003) reported a large positive wheat grain yield response to foliar urea application in the situation where the targeted yield exceeded the capacity of soil to meet N demand by crops. Varga and Sevenjak (2006) found that late-season urea spraying consistently improved grain yield and quality when a low basal N rate was used at seeding. However, the fate of foliar applied N compared with N applied as urea to soil in soil-plant systems is not well quantified under variable climatic conditions. Rawluk et al. (2000) reported that in growth chamber studies, when soil water conditions were optimal for the growth and N uptake, recovery of foliar-applied 15N-labelled urea by wheat was significantly lower than urea applied to soil at the same growth stage, suggesting that the majority of N was absorbed through the soil root pathway or that less foliar-applied N was translocated to the grain than soil-applied N as was also suggested in early studies in Australia (Strong 1982). In WA, most of the current fluid N technology is aimed at early applications (when essentially all the fluid N is available for uptake by roots) rather than late applications (when foliar uptake would be more important). Field data on the relative effectiveness of fluid v. granular N for post-seeding applications is clearly a knowledge gap for cereal and oilseed production in WA.

Controlled or slow release N fertilisers may offer an option to reduce N losses if fertiliser products improve synchronisation between the release of N and the demand of the crops (Giller et al. 2004). The use of controlled or slow release fertiliser products is still limited in agricultural crops due to the high costs of these products. However, their cost has been gradually decreasing over time, and recently Agrium developed a new controlled release fertiliser product (Environmentally Smart Nitrogen, ESN) for corn, wheat, canola, barley, and potato crops in the USA and Canada (Agrium 2006). The product has been under testing in these countries but it is now commercially available. Use of controlled release fertiliser for crop production is still very limited in Australia. In WA, experiments were conducted to test plastic- or sulfur-coated urea products and nitrification inhibitors but the results were not sufficiently cost-effective for growers to adopt these technologies (Chen et al. 2007). However, given the development of ESN for many cereal crops in the USA and Canada, there is a need to evaluate the products in Australia (UM 00023, Chen et al. 2008).

Site-specific N management strategies also need to be considered. Depending on when decisions are made and what information is used, they can be (i) prescriptive, (ii) corrective, or (iii) a combination of both (Dobermann and Cassman 2002). With the prescriptive approach, the amount and timing of N applications are prescribed before sowing. Thus, site-specific spatial information from various sources such as maps of soil properties, yield maps, terrain attributes, EC maps, etc., are essential to determine crop response to N (Giller et al. 2004). With the corrective approach, different diagnostic tools can be used to assess soil or crop N status during the growing season as the basis for making decisions on N applications at certain growth stages. Several promising technologies have emerged in recent years, with particular emphasis given to real-time measurements of crop greenness using different tools (Ladha et al. 2005). Use of the above tools to help make in-season N decisions is just starting in Australian agriculture and more research is needed to confirm their effectiveness here. There are benefits if prescriptive and corrective approaches are integrated to quantify how much, where, and when N must be added. This integration reduces uncertainties because a variety of information sources is used, including pre-season assessment of soil N supply and in-season assessment of crop N demand (Giller et al. 2004). Dobermann et al. (2004) reviewed many studies conducted in the USA, Europe, and Asia, and found that in maize, wheat, and rice production systems, use of a site-specific N management approach did not always increase grain yield but it reduced N rates application while maintaining crop yields, and hence consistently improved fertiliser N use efficiency in these systems (Table 5).

Developing recommendations for managing emerging nutrient deficiencies

In the WA wheatbelt, for some nutrients (P, Zn, Cu, and Mo), soil reserves have been increased due to past fertiliser applications but for others (K, S, Mg, and B), exports of these nutrients in grain due to increasing yields have not been matched by inputs (Chen et al. 2007). The importance of S nutrition in canola, wheat, and grain legumes has also been addressed in several projects (Anderson et al. 2006; Brennan and Bolland 2006b). Boron deficiency risk in WA soils has been studied by Wong et al. (2005); the study showed that on acid sandy soils with low B, continuous cropping would put further pressure on soil B reserves, and as planting areas and yields of canola and grain legumes increase, the risk of B deficiency is likely to increase. The potential risk of soil Ca deficiency has also been raised recently, due to a change of P fertiliser type used in WA (Chen et al. 2007); however, Ca budgets in the National Land and Water Resources Audit (2001) were positive for both grazing and cropping systems. This review will focus on Mg as Mg deficiency seems to be most likely based on nutrient budgeting studies in WA and the fact that very little information exists on potential risk of Mg deficiency across different soils in WA.

Magnesium balance in the WA cropping systems

Farm-gate nutrient balances conducted through the National Land and Water Resources Audit for 2 broad land uses in WA suggested Mg balance was negative to neutral in the cropping zone (NLWRA 2001). In a study of nutrient balance for typical WA wheatbelt farms, Wong et al. (2000b) reported a strong negative K and Mg balance as a result of low addition and continuous export of these nutrients in grain (Table 6). These 2 nutrient balance studies reveal that intensification of grain cropping would result in further depletion of soil K and Mg unless there are changes in nutrient management practices. The requirements for K in cropping systems have been addressed in research (Edwards 1997; Wong et al. 2000a; Brennan and Bolland 2006a), but for Mg it is not clear what levels of risk exist at present and how this would vary with soils and regions for crop production in WA. The major Mg inputs to soils are from the breakdown of certain minerals, and input from rain and fertiliser. Losses include crop removal and leaching. Generally, Mg input from rain is small except for the coastal areas of the WA wheatbelt, where Mg inputs from rain may meet crop requirements (Scott 1990). The major controls of Mg supply to crops are current exchangeable Ms, the release from mineral breakdown and loss through leaching.

Soil Mg content, soil test calibration and crop response to Mg application

Soil Mg content in WA generally ranges from 0.05 to 0.5%. Exchangeable Mg represents ~5% of the total soil Mg and normally occupies 4-20% of the cation exchange capacity (McArthur 1991). In a study of 91 acidic Queensland soils, Brace et al. (1989) reported that 52% of surface soils and 59% of subsoils had Mg concentrations <1 cmol(+)/kg. In a similar study of 60 highly weathered soils, Menzies et al. (1994) reported median exchangeable Mg values of 0.91 and 0.3 cmol(+)/kg for surface and subsoils, respectively. McArthur (1991) reported that many Reference soils in WA had exchangeable Mg <0.2 cmol(+)/kg. The data from the Chemistry Centre, WA (G. Anderson, pers. comm.) indicated a positive relationship between clay content and soil exchangeable Mg levels for surface soils in WA (Fig. 4), and the data also indicate that 8% of all sandy soil samples had exchangeable Mg levels <0.2 cmol(+)/kg and 19% of all clay soil types had exchangeable Mg levels <0.5 cmol(+)/kg, suggesting greater risk of Mg deficiency in light-textured soils.

Studies in other countries have demonstrated that Mg deficiency in plants generally occurs in soils that are low in exchangeable Mg, light-textured, highly leached, acidic, and low in cation exchange capacity (Aitken and Scott 1999). In Australia, Mg deficiency in sugarcane was observed, with yield responses to the application of Mg being recorded in a field trial (Ridge et al. 1980). On acidic soils in Queensland, field and glasshouse work also reported several maize yield responses to applied Mg (Hailes et al. 1994, 1997; Aitken et al. 1999). In WA, Mg deficiency has been reported in fruit crops and vegetable crops but not in commercial cereal and grain legume crops (Edwards 1998b). Soil exchangeable Mg provides a good estimate of plant-available Mg (Aitken and Scott 1999). Little soil test calibration has been conducted for Mg in Australian soils. For New South Wales soils, Cumming and Elliot (1991) considered that Mg deficiency in sandy soils occurs at exchangeable Mg concentrations <0.2 cmol(+)/kg, and in clay soils at <0.5 cmol(+)/kg. Hailes et al. (1997) reported a critical exchangeable Mg concentration of 0.21 cmol(+)/kg for maize in a glasshouse study. In a field study, Aitken et al. (1999) reported a critical Mg value of 0.27 cmol(+)/kg, corresponding to 90% of relative yield. Studies in other countries suggested that Mg deficiency was generally associated with exchangeable Mg concentrations below about 0.2-0.3 cmol(+)/kg (Aitken and Scott 1999). However, exchangeable Mg measurements based on surface soil sampling may not indicate risk of Mg deficiency if there is abundant Mg in subsoils. For example, soil analysis data from the Chemistry Centre, WA (G. Anderson, pers. comm.), indicate that on duplex soils, subsoil Mg was high, and could be an important source of plant-available Mg. Scott and Conyers (1995) suggested that Mg deficiency observed in wheat in new South Wales was transitory as the soils had low exchangeable Mg in the upper soil layer but high Mg at depth.

[FIGURE 4 OMITTED]

General discussion

Inputs from a wide range of stakeholders in the grains industry suggest that the most important nutrient management issues requiring further consideration are: (a) improving soil testing and interpretation; (b) the role of fluid fertilisers in crop nutrient management; (c) using spatial and temporal information to improve crop nutrient management, particularly for N; and (d) managing emerging nutrient deficiencies (Chen et al. 2007).

With respect to further improving soil testing and interpretation, this review suggests that future research should focus on addressing soil sampling and interpretation questions, as they are important factors affecting the accuracy of fertiliser recommendations in response to changing farming practices (see Chen et al. 2007 for detailed recommended R&D activities).

The review indicates the inconsistent crop response to fluid P application observed in WA and this will hamper further extension and adoption of this technique by grain farmers. Thus, more work may be needed to understand the mechanisms involved in different crop response to fluid P fertiliser additions, particularly on low pH soils in WA where lime and solid P fertilisers are often used to maintain crop production (Chen et al. 2007). An understanding of the long-term performance of fluid P will require an assessment of its residual value compared to granular P under field conditions. The reviewed information also suggests that fluid N is as effective as granular urea. Post-seeding application of fluid N fertilisers could become an important tool to manage N input tactically to improve crop yield and grain protein in WA. Thus, there is a need to quantify the fate of granular urea or fluid N applied to dryland wheat crops to determine the agronomic efficiency of fluid N application at different crop growth stages and potential losses compared to granular N fertiliser (Chen et al. 2007).

The PA technology has potential to improve crop nutrient management and farm profitability in WA. Overall, within-paddock yield variability exists and is large enough (0.5-4.9t/ha) to justify the investment in using PA technology at farm scale. The review indicates that understanding both spatial and temporal yield variation is critical for the successful adoption of PA technologies by growers. In WA, there is also a need to explore the use of different layers of spatial information for determining management zones, and conduct field validation studies comparing different zoning methods (Chen et al. 2007). A recent review of potential technologies that could be used for rapid soil measurements (McKenzie et al. 2003) indicates that mid-infrared instruments, EM survey instruments, and ion-selective field transistors have potential for rapid soil measurements. In Australia, using mid-infrared diffuse reflectance spectroscopy as a tool for rapid soil sample analysis (rather than 'wet' chemistry methods) has been investigated extensively (Bramley and Janik 2005). It is hoped that such development could result in a rapid method for soil nutrient assessment spatially in fields. Crop monitoring is clearly an important means to optimise inputs during a growing season. In addition, developing crop-quality assessment methods could result in the possibility of in-field separation of grain into quality classes bringing greater competitive advantage from the adoption PA.

In response to the widespread adoption of minimum tillage and stubble retention cropping systems, gradually increased use of liquid fertilisers together with advanced application technologies, and increased interest in cropping in the HRZ, there is a need to better understand growers' practices and attitudes to crop N management, and to better position research and extension activities. New fertiliser technologies and site-specific N management could play a key role in reduced N losses in soil-plant systems and improve N use efficiency (Chen et al. 2007).

The negative balance of Mg outputs relative to inputs observed in WA cropping systems, together with the information reviewed on factors affecting soil Mg content and crop response to Mg application across Australia, suggests that there is a real need to evaluate the risk of Mg deficiency in WA, particularly on acidic sandy soils (see Chen et al. 2007 for details).

Acknowledgments

We thank all people from Grower Groups, CSBP, Summit Fertilisers, United Farmers, Nutrient Management Systems Pty Ltd, Silverfox, University of Western Australia, CSIRO, Chemistry Centre of WA, Department of Agriculture and Food Western Australia (DAFWA), and independent advisers, who participated in face-to-face discussions, informal phone conservations, focus group discussions and a written survey, and provided their professional views on a diversity of issues associated with crop nutrient management. We thank Mr. Patrick Page (DAFWA) for his help in organising focus group discussion. We thank Dr Robert Belford for his help in accessing final reports of GRDC funded projects in WA; Mr. Martin Shafron (Fertilizer Industry Federation of Australia) for his help in accessing unpublished data; Dr Mike Wong, Dr Mike McLaughlin, Dr Geoff Anderson, Dr Stephen Loss, Mr. Sandy Alexander, Ms. Alison Slade, and Dr Cindy Grant (Canada) for their help in providing research reports and published or unpublished information. We also thank Mr. Ian Mating, Mr. Wayne Pluske, Mr. Eddy Pol, Mr. Patrick Gethin, Mr. Jeremy Lemon, Ms. Caroline Peek for discussions and information on the current crop nutrient management practices in WA. We thank Dr Dave Allen (section 3), Dr Mike Bolland (section 3), Dr Stephen Loss (section 4), and Ms. Bindi Isbister (section 5), who reviewed draft sections and provided comments that improved the draft. Finally we thank the GRDC for financial support.

Manuscript received 28 April 2008, accepted 15 September 2008

References

Adams ML, Cook SE, Bowden JW (2000) Using yield maps and intensive soil sampling to improve nitrogen fertiliser recommendations from a deterministic model in the Western Australia wheatbelt. Australian Journal of Experimental Agriculture 40, 959-968. doi: 10.1071/ EA99129

Adams ML, Maling I (2005) Simplifying management zone--a pragmatic approach to the development and interpretation of management zones in Australia. In 'Proceedings of the 7th International Conference on Precision Agriculture'. 25-28 July 2004. (Ed. Mulla) (Precision Agriculture Centre, University of Minnesota: St. Paul, Mn)

Agrium (2006) Agrium Inc. www.agrium.com.

Aitken RL, Dickson T, Hailes KJ, Moody PW (1999) Response of field-grown maize to applied magnesium in acidic soils in north-eastern Australia. Australian Journal of Agricultural Research 50, 191-198. doi: 10.1071/A98149

Aitken RL, Scott BJ (1999) Magnesium. In 'Soil analysis: an interpretation manual'. (Eds KI Peverill, LA Sparrow, DJ Router) pp. 255-262. (CSIRO Publishing: Collingwood, Vic.)

Anderson GC, Fillery IR, Ripper FH, Leach BJ (2006) Sulfur mineralization in a coarse-textured soil atter different sulfate fertilisation histories, and yield responses of wheat and lupin. Australian Journal of Soil Research 44, 165-174. doi: 10.1071/SR05080

Anderson WK, Hoyle FC (1999) Nitrogen efficiency of wheat cultivars in a Mediterranean environment. Australian Journal of Experimental Agriculture 39, 957-965.

Ashworth J, Crepin JM, Repski R (1994) Soil sampling with modified chainsaw. Agronomy Journal 86, 736-740.

Bertrand I, McLaughlin M J, Holloway RE, Armstrong RD, McBeath TM (2006) Changes in P bioavailability induced by the application of liquid and powder sources of P, N and Zn fertilisers in alkaline soils. Nutrient Cycling in Agroecosystems 74, 27-40. doi: 10.1007/s10705-005-4404-3

Bly AG, Woodard HJ (2003) Foliar nitrogen application timing influence on grain yield and protein concentration of hard winter and spring wheat. Agronomy Journal 95, 335-338.

Bolland MDA (1992) Effect of sampling depth on bicarbonate soil phosphorus test values. Fertilizer Research 32, 121-124. doi: 10.1007/ BF01054403

Bolland MDA (1995a) Studies on temporal variation of Colwell soil-test phosphorus measured on soil samples collected from November to March in south-western Australia. Communications in Soil Science and Plant Analysis 26(11&12), 1947-1961.

Bolland MDA (1995b) Problems for soil-testing for phosphorus for narrow-leafed sweet lupins (Lupinus angustifolius). In 'Spotlight on lupins'. Department of Agriculture Western Australia Publication 2/95. pp. 191-193. (Department of Agriculture Western Australia: Perth, W. Aust.)

Bolland MDA, Allen DG, Walton KS (2003) Soil testing for phosphorus: comparing the Mehlich 3 and Colwell procedure for soils of southwestern Australia. Australian Journal of Soil Research 41, 1185-1200. doi: 10.1071/SR02153

Bolland MDA, Brennan RF (2006) Phosphorus, copper and zinc requirements of no-till wheat crops and methods of collecting soil samples for soil testing. Australian Journal of Experimental Agriculture 46, 1051-1059. doi: 10.1071/EA05024

Bolland MDA, Gilkes RJ (1990) Cultivation reduces fertiliser residual effectiveness and affects soil testing for available phosphorus. Fertilizer Research 24, 33-46. doi: 10.1007/BF01073144

Bolland MDA, Jarvis RJ (1996) Effectiveness of different methods of applying superphosphate for lupins grown on sandplain soils. Australian Journal of Experimental Agriculture 36, 707-715. doi: 10.1071/EA9960707

Bolland MDA, Wilson IR (1994) Soil phosphorus testing: 1. Studies on spatial variation of Colwell soil test phosphorus. Communications in Soil Science and Plant Analysis 25(13&14), 2371-2384.

Bowden JW (1999) A decision aid to increase farm profit through better grain quality decisions. A final report prepared for the Grains Research and Development Corporation. Department of Agriculture, Western Australia, Perth.

Bramley RGV, Janik LJ (2005) Precision agriculture demands a new approach to soil and plant sampling and analysis--examples from Australia. Communications in Soil Science and Plant Analysis 36, 9-22. doi: 10.1081/CSS-200042958

Brennan RF, Bolland MDA (2006a) Soil and tissue tests to predict the potassium requirements of canola in south-western Australia. Australian Journal of Experimental Agriculture 46, 675-679. doi: 10.1071/ EA04218

Brennan RF, Bolland MDA (2006b) Soil and tissue tests to predict the sulfur requirement of canola in south-western Australia. Australian Journal of Experimental Agriculture 46, 1061-1068. doi: 10.1071/ EA04206

Brennan RF, Bolland MDA (2007) Soil and tissue tests to predict the phosphorus requirements of canola in south-western Australia. Journal of Plant Nutrition 30, 1767-1777.

Brown AJ (1993) A review of soil sampling for chemical analysis. Australian Journal of Experimental Agriculture 33, 983-1006. doi: 10.1071/EA9930983

Bruce RC, Warrell LA, Bell LC, Edwards DG (1989) Chemical attributes of some Queensland soils. 1. Solid and solution phase compositions. Australian Journal of Soil Research 27, 333-351. doi: 10.1071/ SR9890333

Burgess SJ, Bowden JW, Diggle AJ (1991) NPDEC1DE User's Guide: A computer program to help with nitrogen and phosphorus fertiliser decisions for cereals. Agriculture Western Australia Bulletin No. 4343. pp. 164-167.

Burgess SJ, Diggle AJ, Bowden JW, Fillery IR (1992) NAVAIL Spreadsheet. Miscellaneous Publications 27/92, Department of Agriculture, Western Australia.

Carr SJ, Ritchie SP, Porter WM (1991) A soil test for aluminium toxicity in acidic subsoils of yellow earths in Western Australia. Australian Journal of Agricultural Research 42, 875-892. doi: 10.1071/AR9910875

Cassman KG, Dobermann A, Waiters DT (2002) Agroecosystems, nitrogen-use efficiency, and nitrogen management. Ambio 31(2), 132-140. doi: 10.1639/0044-7447(2002)031[0132:ANUEAN]2.0.CO;2

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

Chen W, Bell RW, Bowden JW, Porter W (2007) Review of nutrient management issues in the Western Australia grains industry. Final project (UWA 00084) report to Grains Research & Development Corporation, Canberra.

Colwell JD (1963) The estimates of the phosphorus fertiliser requirements of wheat in southern New South Wales by soil analysis. Australian Journal of Experimental Agriculture and Animal Husbandry 3, 191-197. doi: 10.1071/EA9630190

Cook SE, Bramley RGV (1998) Precision agriculture--opportunities, benefits and pitfalls of site-specific crop management in Australia. Australian Journal of Experimental Agriculture 38, 753-763. doi: 10.1071/EA97156

Cornish PS (1987) Effects of direct drilling on the phosphorus uptake and fertiliser requirements of wheat. Australian Journal of Agricultural Research 38, 775-790.

Crabtree B (2004) The development and extension of no-till farming systems in WA. Final report to Grains Research & Development Corporation (GRDC).

CSBP (2003) Late Flexi-N applications can boost wheat protein. Productivity Highlights, Number 3, October 2003.

CSBP (2005) Applying Flexi-N-boom sprayer or seeder? Productivity Highlights, Number 2, August 2005.

Cumming RW, Elliot GL (1991) Soil chemical properties, in 'Soils: Their properties and management'. (Eds PEV Charman, BW Murphy) pp. 193-205. (Sydney University Press: Sydney, NSW)

Diggle AJ, Bowden JW, Burgess SJ (2003) Users manual of 'Select your Nitrogen': A decision tool for quantifying nitrogen availability and crop response in broadacre farming systems. Department of Agriculture, Western Australia, Perth.

Dobermann A (2005) Nitrogen use efficiency-state of the art. In 'The IFA International Workshop on Enhanced-Efficiency Fertilizers'. Frankfurt, Germany, 28-30 June 2005.

Dobermann A, Blackmore S, Cook SE, Adamchuk VI (2004) Precision farming: challenges and future directions. In 'Proceedings of the 4th International Crop Science Congress'. 26 Sept.-1 Oct. 2004, Brisbane, Qld. (CD-ROM) Website: www.cropscience.org.au

Dobermann A, Cassman KG (2002) Plant nutrient management for enhanced productivity in intensive grain production systems of the United States and Asia. Plant and Soil 247, 153-202. doi: 10.1023/ A:1021197525875

Dobermann A, Cassman KG (2004) Environmental dimensions of fertilizer nitrogen: what can be done to increase nitrogen use efficiency and ensure global food security? In 'Agriculture and the nitrogen cycle: Assessing the impacts of fertilizer use on food production and environment'. SCOPE 56. (Eds AR Mosier, JK Syers, JR Freney) pp. 261-278. (Island Press: Washington, DC)

Edwards NK (1997) Potassium fertiliser improves wheat yield and grain quality on duplex soils. In 'Proceedings of the First Workshop on Potassium in Australian Agriculture'. 3-4 October 1997, Geraldton, Western Australia. (Eds MTF Wong, Y Pal, NK Edwards) pp. 69-75.

Edwards NR (1998a) Potassium. In 'Soil Guide: a hand book for understanding and managing agricultural soils'. (Ed. G Moore) Agriculture Western Australia Bulletin No. 4343.

Edwards NR (1998b) Other nutrients magnesium. In 'Soil guide: a hand book for understanding and managing agricultural soils'. (Ed. G Moore) Agriculture Western Australia Bulletin No. 4343.

Fageria NK, Baligar VC (2005) Enhancing nitrogen use efficiency in crop plants. Advances in Agronomy 88, 97-185. doi: 10.1016/S0065-2113 (05)88004-6

Fillery IR, McInnes KJ (1992) Components of the fertilizer nitrogen balance for wheat production on duplex soils. Australian Journal of Experimental Agriculture 32, 882-899. doi: 10.1071/EA9920887

Fillery IRP (2004) Issues with the application of foliar N fertilisers. In 'First Fluid Fertiliser Workshop Proceedings (ARRIS Pty Ltd)'. Adelaide, S. Aust. (Eds J Kelly, N Wojcik, MJ McLaughlin) pp. 159-168.

Giller KE, Chalk P, Dobermann A, Hammond L, Heifer P, Ladha JK, Nyamudeza P, Maene L, Ssali H, Freney J (2004) Emerging technologies to increase the efficiency of use of fertilizer nitrogen. In 'Agriculture and the nitrogen cycle: Assessing the impacts of fertilizer use on food production and environment'. SCOPE 56, (Eds AR Mosier, JK Syers, JR Freney) pp. 35-51. (Island Press: Washington, DC)

Godwin RJ, Miller PCH (2003) A review of the technologies for mapping within-field variability. Biosystems Engineering 84(4), 393-407. doi: 10.1016/S1537-5110(02)00283-0

Grant CA, Rawluk CDL, Lukow O, Johnston A, McKenzie R (2001) Optimizing wheat yield and protein on the Canadian prairies. Final report to the Canada-Manitoba Agric.-food Research and Development Initiative (ARDI).

Hailes KJ, Aitken RL, Dickson T, Moody PW, Bell LC (1994) Maize response to applied magnesium in some krasonzems. In 'Soils ain't soils: What's special about Krasnozems'. (Eds A Fulton, L Sparrow) pp. 126-127. (Australian Institute of Agriculture Science: Tasmania)

Hailes KJ, Aitken RL, Menzies NW (1997) Magnesium in tropical and subtropical soils from north-eastern Australia. II. Response by glasshouse-grown maize to applied magnesium. Australian Journal of Soil Research 35, 629-641. doi: 10.1071/S96082

Halse N J, Greenwood EAN, Lapins P, Boundy CA (1969) An analysis of the effects of nitrogen deficiency on the growth and yield of a Western Australian wheat crops. Australian Journal of Agricultural Research 20, 987-998. doi: 10.1071/AR9690987

Hamblin AP, Hamblin J (1985) Root characteristics of some temperate legume species and varieties on deep, free-draining entisols. Australian Journal of Agricultural Research 36, 63-72. doi: 10.1071/ AR9850063

Hill N (2005) Increasing productivity of southern high rainfall cropping systems. A final report prepared for the Grains Research and Development Corporation. Department of Agriculture and Food Western Australia, South Perth.

Holloway RE, Bertrand I, Frischke AJ, Brace DM, McLaughlin M J, Shepperd W (2001) Improving fertiliser efficiency on calcareous and alkaline soils with fluid sources of P, N and Zn. Plant and Soil 236, 209-219. doi: 10.1023/A: 1012720909293

Holloway RE, Frischke BM, Frischke A J, Brace DM, McLaughlin M J, Lombi E, Armstrong RD (2004) Field evidence for efficiency of fluid fertilisers. In 'First Fluid Fertiliser Workshop Proceedings (ARRIS Pty Ltd)'. Adelaide, S. Aust. (Eds J Kelly, N Wojcik, MJ McLaughlin) pp. 11-22.

Howard DD, Essington ME, Tyler D (1999) Vertical phosphorus and potassium stratification in no-till cotton soils. Agronomy Journal 91, 266-269.

James DW, Hurst RL (1995) Soil sampling techniques for band-fertilized, no-till fields with Monte Carlo simulations. Soil Science Society of America Journal 59, 1768-1772.

Kelly R, Strong W, Jensen T, Butler D, Wells N, Apan A (2005) Predicting protein levels is part of the future toolkit. Ground Cover October/ November 2005: Precision Agriculture Supplement. Grains Research and Development Corporation, Canberra. pp. 22-23.

Kitchen NR, Havlin JL, Westfall DG (1990) Soil sampling under no-fill banded phosphorus. Soil Science Society of America Journal 54, 1661-1665.

Knell G, Slade A (2005) Using precision agriculture and soil inoculants to improve crop performance and grower returns. Report to the Grains Research and Development Corporation. Corrigin farm improve group WA.

Kuhlmann H (1990) Importance oftbe subsoil for the K nutrition of crops. Plant and Soil 127, 129-136. doi: 10.1007/BF00010845

Ladha JK, Pathak H, Krupnik TJ, Six J, van Kessel C (2005) Efficiency of fertilizer nitrogen in cereal production: retrospects and prospects. Advances in Agronomy 87, 85-156. doi: 10.1016/S0065-2113(05) 87003-8

Lombi E, McLaughlin M J, Johnston C, Armstrong RD, Holloway RE (2004) Mobility and lability of phosphorus from granular and fluid monoammonium phosphate differs in a calcareous soil. Soil Science Society of America Journal 68, 682-689.

Loss S (2004) Agronomic performance of fluid fertilisers in WA. In 'First Fluid Fertiliser Workshop Proceedings'. (Eds J Kelly, N Wojcik, MJ McLaughlin) pp. 187-189. (ARRIS Pty Ltd: Adelaide, S. Aust.)

Loss S (2006) Producing profits with phosphorus. In 'Proceedings of 2006 Agribusiness Crop Updates--Cereal Updates Section'. 15-16 February 2006, Western Australia. pp. 66-67.

Loss S, Appelbee M (2006) Agronomic performance and adoption of liquid fertilisers in Western Australia. In 'Proceedings of the 13th Australian Agronomy Conference'. 10-14 September 2006, Perth, W. Aust. (Eds NC Turner, T Acuna, RC Johnson) Website: www.agronomy. org.au

Mallarino AP, Borges R (2006) Phosphorus and potassium distribution in soil following long-term deep-band fertilization in different tillage systems. Soil Science Society of America Journal 70, 702-707. doi: 10.2136/sssaj2005.0129

Mason MG (1994) Nitrogen fertilizers for cereal production. Agriculture Western Australia Bulletin No. 4283, pp. 1-19.

Mason MG (1998) Nitrogen. In 'Soil guide: A handbook for understanding and managing agricultural soils'. (Ed. G Moore) pp. 164-167. Agriculture Western Australia Bulletin No. 4343.

Mason MG, Rowley AM, Quayle DJ (1972) The fate of urea applied at various intervals after the sowing of a wheat crop on a sandy soil in Western Australia. Australian Journal of Experimental Agriculture and Animal Husbandry 12, 171-175. doi: 10.1071/EA9720171

Mason S, Hamon R, Zhang H, Anderson J (2008) Investigating chemical constraints to the measurement of phosphorus in soils using diffusive gradients in thin films (DGT) and resin methods. Talanta 74, 779-787. doi: 10.1016/j.talanta.2007.07.005

McArthur WM (1991) 'Reference soils of south-western Australia.' (Australian Soil Science Society Inc. (WA Branch): Perth W. Aust.)

McBeath TM, Armstrong RD, Lombi E, McLaughlin M J, Holloway RE (2005) Responsiveness of wheat (Triticum aestivum) to liquid and granular phosphorus fertilisers in southern Australian soils. Australian Journal of Soil Research 43, 203-212. doi: 10.1071/SR04066

McBeath TM, McLaughlin M J, Armstrong RD, Bell M, Bolland DA, Conyers MK, Holloway RE, Mason SD (2007) Predicting the response of wheat (Tritium aestivum L.) to liquid and granular phosphorus fertilisers in Australian soils. Australian Journal of Soil Research 45, 448-458. doi: 10.1071/SR07044

McBratney AB, Whelan BM, Ancev T, Bouma J (2005) Future directions of precision agriculture. Precision Agriculture 6, 7-23. doi: 10.1007/ s11119-005-0681-8

McDonald KG (1989) The contribution of nitrogen fertilizer to the nitrogen nutrition of rainfed wheat crops in Australia: a review. Australian Journal of Experimental Agriculture 29, 455-481. doi: 10.1071/ EA9890455

McKenzie N J, Bramley RGV, Farmer T, Janik LG, Murray W, Smith C J, Mclaughlin MJ (2003) Rapid soil assessment--a review of potential benefits and opportunities for the Australian grains industry. Report for the Grains Research and Development Corporation, Canberra.

McLaughlin MJ (2002) New fertiliser formulations for highly calcareous soils of South Australia. Final report to the South Australia Grain Industry Trust, CSIRO Land and Water.

Mehlich A (1984) Mehlich-3 soil test extractant: a modification of Mehlich-2 extractant. Communications in Soil Science and Plant Analysis 15, 140-1416.

Menzies NW, Bell LC, Edwards DG (1994) Exchangeable and solution phase chemistry of acid, highly weathered soils. 1. Characteristics of soils and effects of lime and gypsum amendments. Australian Journal of Soil Research 32, 251-267. doi: 10.1071/SR9940251

National Land and Water Resources Audit (2001) 'Australian Agricultural Assessment. Vol. 1.' (NLWRA: Canberra)

Patabendige D, Wong MTF, Bowden JW (2003) Zone management in precision agriculture by matching fertiliser input to crop demand. Bulletin 4584, Department of Agriculture Western Australia, Perth.

Price P (2004) Spreading the PA message. Ground Cover, August 2004. Grains Research and Development Corporation, Canberra.

Price P (2005) Is it worth mapping your paddock? Ground Cover, October/ November 2005: Precision Agriculture Supplement. Grains Research and Development Corporation, Canberra.

Probert ME, Jones PK (1977) The use of soil analysis for predicting the response to sulphur of pasture legumes in the Australian tropics. Australian Journal of Soil Research 15, 137-146. doi: 10.1071/ SR9770137

Raun WR, Johnson GV (1999) Improving nitrogen use efficiency for cereal production. Agronomy Journal 91, 357-363.

Rawluk CDL, Racz GJ, Grant CA (2000) Uptake of foliar or soil application of [sup.15]N-labelled urea solution at anthesis and its affect on wheat grain yield and protein. Canadian Journal of Plant Science 80, 331-334.

Rengel Z (2005) Breeding crops for adaptation to environments with low nutrient availability. In 'Abiotic stresses: Plant resistance through breeding and molecular approaches'. (Eds M Ashraf, PJC Harris) pp. 239-276. (The Haworth Press: New York)

Ridge DR, Hurney AP, Haysom MBC (1980) Some aspects of calcium and magnesium nutrition in north Queensland. Proceedings' of the Australian Society of Sugar Cane Technologists 2, 55-61.

Robertson MJ, Isbister B, Maling I, Oliver Y, Wong MTF, Adams ML, Bowden JW, Tozer P (2006b) Managing spatial and seasonal variability within field can improve the profitability of WA grain production. In 'Proceedings 13th Australian Society of Agronomy'. 10-14 September 2006, Perth, Western Australia. www.agronomy.org.au

Robertson MJ, Lyle G, Bowden JW, Brennan L (2006a) How much yield variation do you need to justify zoning inputs. Agribusiness Crop Updates, 15-16 February, Perth, Western Australia.

Robson AD, Longnecker NE, Osborne LD (1992) Effects of heterogenous nutrient supply on root growth and nutrient uptake in relation to nutrient supply on duplex soils. Australian Journal of Experimental Agriculture 32, 879-886. doi: 10.1071/EA9920879

Rodriguez D, Sadras VO, Christensen LK, Belford R (2005) Spatial assessment of the physiological status of wheat crops as affected by water and nitrogen supply using infrared imagery. Australian Journal of Agricultural Research 56, 983-493. doi: 10.1071/AR05035

Scott BJ (1990) Diagnosis of magnesium deficiency by plant analysis. PhD Thesis, The University of Western Australia, Australia.

Scott BJ, Conyers MK (1995) Magnesium nutrition and lime movement down the profile. In 'Making better fertiliser, lime and gypsum recommendations'. (Eds IG Fenton, PW Orchard) (NSW Agriculture: Wagga Wagga, NSW)

Selles F, McConkey BG, Campbell CA (1999) Distribution and forms of P under cultivator- and zero-tillage for continuous- and fallow-wheat cropping systems in the semi-arid Canadian prairies. Soil & Tillage Research 51, 47-59. doi: 10.1016/S0167-1987(99)00027-6

Skerritt JH, Adams ML, Cook SE, Naglis G (2002) Within-field variation in wheat quality: implications for precision agriculture management. Australian Journal of Agricultural Research 53, 122-1242. doi: 10.1071/AR01204

Stecker J, Brown JR (2001) Soil phosphorus distribution and concentration from repeated starter phosphorus band application. Communications in Soil Science and Plant Analysis 32(5&6), 803-819. doi: 10.1081/ CSS-100103910

Strong WM (1982) Effect of late application of nitrogen on the yield and protein content of wheat. Australian Journal of Experimental Agriculture and Animal Husbandry 22, 54-61. doi: 10.1071/EA9820054

Thompson G (2003) Nutrient distribution patterns of 3 no-till soils. In 'Proceedings of Agribusiness Crop Updates 2003'. Western Australia.

Tyler DD, Howard DD (1991) Soil sampling patterns for assessing no-tillage fertilization. Journal of Fertilizer Issues 8(3), 52-56.

van Herwaarden AF, Farquhar GD, Angus JF, Richard RA, Howe GH (1998) 'Haying-off, the negative grain yield response of dryland wheat to nitrogen fertilizer 1. Biomass, grain yield, and water use. Australian Journal of Agricultural Research 49, 1067-1081. doi: 10.1071/A97039

Varga B, Sevcnjak Z (2006) The effect of late-season urea spraying on grain yield and quality of winter wheat cultivars under low and high basal nitrogen fertilization. Field Crops Research 96, 125-132. doi: 10.1016/j. fcr.2005.06.001

Whelan BM, McBratney AB (2000) The "null hypothesis" of precision agriculture management. Precision Agriculture 2, 265-279. doi: 10.1023/A:1011838806489

White PF (1990) The influence of alternative tillage systems on the distribution of nutrients and organic carbon in some common Western Australian wheatbelt soils. Australian Journal of Soil Research 28, 95-116. doi: 10.1071/SR9900095

Wong MTF, Asseng S (2004) Fluctuations in spatial variability of wheat yields. In 'Proceedings of the 4th International Crop Science Congress'. 26 September-1 October 2004, Brisbane, Australia. www.cropscience. org.au

Wong MTF, Asseng S (2006) Determining the causes of spatial and temporal variability of wheat yields at sub-field scale using a new method of upscaling a crop model. Plant and Soil 258, 1-13.

Wong MTF, Bell RW, Frost K (2005) Mapping boron deficiency risk in soils of south-west Western Australia using a weight of evidence model. Australian Journal of Soil Research 43, 811-818. doi: 10.1071/SR05022

Wong MTF, Corner RJ, Cook SE (2001) A decision support system for mapping the site-specific potassium requirement of wheat in the field. Australian Journal of Experimental Agriculture 41, 655-661. doi: 10.1071/EA00191

Wong MTF, Edwards NK, Barrow NJ (2000a) Accessibility of subsoil potassium to wheat grown on duplex soils in the south-west of Western Australia. Australian Journal of Soil Research 38, 745-751. doi: 10.1071/SR98117

Wong MTF, Harper RJ (1999) Use of on-ground gamma-ray spectrometry to measure plant-available potassium and other topsoil attributes. Australian Journal of Soil Research 37, 267-277. doi: 10.1071/S98038

Wong MTF, Oliver Y, Robertson M J, Wittwer K (2006) Technologies to estimate plant available soil water storage capacities at high spatial resolution. In 'Proceedings 13th Australian Society of Agronomy'. 10-14 September 2006, Perth, Western Australia. www.agronomy.org.au

Wong MTF, Wittwer K (1997) Aspects of potassium impacting on the sustainability of Western Australian agriculture. In 'Proceedings of the First Workshop on Potassium in Australian Agriculture'. 3-4, October 1997, Geraldton, Western Australia. (Eds MTF Wong, Y Pal, NK Edwards) pp. 9-14.

Wong MTF, Wittwer K, Zhang H (2000b) Historical nutrient balance at paddock and whole farm scales for typical wheatbelt farms in the Dowerin-Wongan Hills area. Agribusiness Crop Updates 2000, Perth, Western Australia.

Wylie P, Mayfield A, Harrop F, Roe A (2003) The potential of fluid fertilisers for broadacre cropping in Australia. A review prepared for the Grains Research and Development Corporation, Horizon Rural Management.

Zhang H, Davison W, Gadi R, Kobayashi T (1998) In situ measurement of dissolved phosphorus in natural water using DGT. Analytica Chimica Acta 370, 29-38. doi: 10.1016/S0003-2670(98)00250-5

Zhang H, Turner NC, Poole ML, Simpson N (2006) Crop production in the high rainfall zones of southern Australia--potential, constraints and opportunities. Australian Journal of Experimental Agriculture 46, 1035-1049. doi: 10.1071/EA05150

W. Chen (A,B,E), R. W. Bell (A), R. F. Brennan (B), J.W. Bowden (B), A. Dobermann (C), Z. Rengel (D), and W. Porter (B)

(A) School of Environmental Science, Murdoch University, 90 South Street, Murdoch, WA 6150, Australia.

(B) Department of Agriculture and Food, Western Australia, Locked Bag 4, Bentley Delivery Centre, WA 6983, Australia.

(C) International Rice Research Institute, Los Banos, Laguna, Philippines.

(D) School of Earth & Geographical Sciences, University of Western Australia, Crawley, WA 6009, Australia.

(E) Corresponding author. Email: wchen@agric.wa.gov.au
Table 1. Summary of the sites where liquid P fertilisers produced
significantly greater wheat yield response than equivalent rates
of granular fertilisers in Western Australia (after Loss 2006)

 Yield
 increase
 (t/ha)
 (% yield
Site Year Soil type Soil pH increase in
 (Ca[Cl.sub.2]) parentheses)

Dandaragan 2000 Sandy loam 5.5 - (9) (^)
Salmon Gums 2001 Clay loam 8.0 - (11) (^)
Mukinbudin 2001 Loam 7.5 - (8) (^)
Southern Cross 2003 Sandy loam 4.4 0.4 (31)
West Dale 2003 Gravelly loam 5.4 0.3 (29)
Newdegate 2005 Loamy sand 5.2 0.2 (8)

(^) Only % increases were reported for these sites.

Table 2. APSIM-simulated estimates of potential yield (t/ha)
for wheat on 3 soils contrasting in plant available water
capacity (PAWC) using Buntine rainfall data (Robertson et
al. 2006a)

 Soil
 PAWC (mm)

Rainfall 50 90 130

Above average 1.4 4.5 5.0
Average 0.9 3.0 3.8
Below average 0.6 1.5 1.9

Table 3. Soil/crop variability affecting crop performances in fields
and appropriate survey methods to assess variability
(after Godwin and Miller 2003)

Group Factor Method

Soil water Soil texture, Soil mapping, profile
 structure, description,
 available water electromagnetic
 and water logging induction
Topographic Topography and Topographic surveys,
 micro-climate digital elevation
 models
Soil nutrition Major nutrients, pH, Targeted sampling,
 and trace elements canopy density
 assessment via
 yield maps, image
 analysis
Crop growth Nutrient deficiency, Thermal or
 and health water stress, weeds, reflectance
 condition insects, and diseases imaging

Table 4. Impact of different N management practices on
dry matter and grain yield (t/ha) of Calingiri wheat
at Cranbrook in the high rainfall zone

N application rate: 160 kg N/ha (Hill 2005)

 Dry Grain
Treatments matter yield

Control 4.5 2.2
All N at seeding 4.0 2.2
33% N at seeding, rest 5.3 2.7
 at 1st node
33% N at seeding, rest 7.5 3.5
 after waterlogging
33% N at seeding and 8.9 3.5
 at 1st node and rest
 after waterlogging
l.s.d. (P=0.05) 1.0 0.6

Table 5. Examples of different site-specific management strategies
implemented infield studies (Dobermann etal. 2004) Conventional
treatment: uniform N rate and fixed splitting of N (existing best
management recommendations or farmers' practice). Decision tools
used in N management: S, pre-sowing assessment of soil N supply
using soil sampling and test; Mr, soil/crop model to predict N rate;
Mt, soil/crop model to predict splitting/timing of N applications;
D, in-season diagnosis and adjustment of plant N using sensing
tools

 Decision tools
 N applied
Crop, location N treatment S Mr Mt D (kg/hg)

Maize, USA Conventional x -- -- -- 142
 Site-specific 1 x -- -- -- 141
 Site-specific 2 x -- -- -- 113
Wheat, UK Conventional x -- -- -- 174
 Site-specific x -- -- -- 155
Wheat, Germany Site-specific 1 x -- -- x 178
 Site-specific 2 x x x -- 138
Rice, China Conventional -- -- -- -- 171
 Site-specific x x -- x 126

 N use
 Yield efficiency
Crop, location N treatment (t/ha) (kg/kg)

Maize, USA Conventional 10.3 73
 Site-specific 1 10.4 74
 Site-specific 2 10.2 90
Wheat, UK Conventional 7.4 43
 Site-specific 7.2 46
Wheat, Germany Site-specific l 6.3 35
 Site-specific 2 6.3 46
Rice, China Conventional 6.0 37
 Site-specific 6.4 52

Table 6. Examples of on-farm nutrient balance (kg/ha.year)
based on continuous individual paddock records for 1963-1999

Note: records were of varying length of time to take into
account of acquisition and development of new
paddocks during 1963-99 (Wong et al. 2000b)

 With new
 Continuous paddocks

N-input from legumes
 not accounted for -13.5 -21.0
N-legume inputs
 accounted for 17.6 -2.4
Phosphorus 6.5 4.6
Potassium -4.0 -5.1
Calcium 13.2 12.0
Sulfur 10.0 5.4
Magnesium -1.0 -1.3
COPYRIGHT 2009 CSIRO Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2009 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Chen, W.; Bell, R.W.; Brennan, R.F.; Bowden, J.W.; Dobermann, A.; Rengel, Z.; Porter, W.
Publication:Australian Journal of Soil Research
Article Type:Report
Geographic Code:8AUST
Date:Feb 1, 2009
Words:15521
Previous Article:Preface: nutrient management by the Australian grains industry.
Next Article:The effectiveness of deep placement of fertilisers is determined by crop species and edaphic conditions in Mediterranean-type environments: a review.
Topics:

Terms of use | Privacy policy | Copyright © 2021 Farlex, Inc. | Feedback | For webmasters |