Validation of Nbudget for estimating soil N supply in Australia's Northern grains region in the absence of soil test data.
Farmers in Queensland and the northern half of New South Wales (NSW) produce 20-25% of the nation's grain from approximately 5 Mha cropland (Australian Bureau of Agricultural and Resource Economics (ABARES) 2016). Principal crops are wheat, barley, sorghum, chickpea, mung bean and faba bean. The cereals wheat, barley and sorghum, require nitrogen (N) to be supplied through the in situ mineralisation of soil organic matter and crop residues, as well as from fertiliser N inputs. Matching the supply of N to crop demand remains a challenge for farmers because too little supply reduces crop yields and profits, whereas too much may result in unnecessary costs and potential environmental damage through gaseous and/or leaching losses (e.g. Schwenke et al. 2016).
Crop-soil simulation models and decision support (DS) tools to help farmers and their advisers manage fertiliser N inputs have been developed and promoted during the past 20 years, from the relatively simple paper-based 'Nitrogen in 95/96' (Lawrence et al. 2000) to the more complex computer-based models such as the Agricultural Production Systems sIMulator (APSIM; CSIRO, Toowoomba, Australia), and, most recently, the Better Fertiliser Decisions for Cropping (BFDC) national database (https://www.bfdc.com.au/interroga tor/frontpage.vm, accessed 1 December 2016) and the associated web-based DS tool BFDC Interrogator (Cox and Strong 2008; Carberry et al. 2009; Bell et al. 2013). All of these rely on a budgeting approach in which the supply of plant-available N is determined before sowing, together with crop N demand (i.e. the amount of N required to grow the crop; Myers 1987; Marcellos and Felton 1993). The difference between N supply and demand represents the shortfall that is to be met by fertiliser N inputs. All the tools require soil test data for root zone plant-available N and, in some cases, surface soil organic carbon (C).
Northern grain farmers are encouraged to deep core for soil nitrate late in the precrop fallow to determine supply of plant-available N (e.g. Bell et al. 2013). This is reasonable advice given the large variations in sowing soil nitrate levels that can occur. Surveys of winter cropping paddocks in northern NSW during the past 20 years consistently showed as much as 15-fold variations in soil nitrate at the end of the summer fallow and before sowing a winter crop (e.g. Schwenke et al. 1998; Elias and Herridge 2014). Conversely, coring soils to depths of 1 m or more is time consuming and expensive, with the frequency and intensity of coring likely to be compromised to save both time and money. The net result is that many cropping paddocks in the northern grains region are not deep-cored for N testing each year and decisions about fertiliser N inputs are more often based on the relativity of grain and fertiliser prices, paddock knowledge and past (customary) practice (G. D. Schwenke, pers. comm.). Indeed, surveys of farmers and advisors in Australia's northern grains region and elsewhere indicate only a moderate incidence of soil testing for N and a low use of soil test data for decision making (Armstrong et al. 2016; Lobry de Bruyn and Andrews 2016).
Thus, there appears to be a disconnection between the exemplar approach to grain cropping N management, involving soil testing for root zone plant-available N to be used with some type of DS tool, and what farmers are actually doing. The underlying assumption providing the impetus for the work described herein was that a simple-to-use DS tool that did not require either a high level of soil technical knowledge or soil test data for plant-available N (and organic C) may provide a means of improving N management at the paddock level (see Armstrong et al. 2016). Such a DS tool, NBudget, was developed by the author during 2008-12 and subsequently promoted to approximately 800 potential users in Australia's northern grains region through presentations, workshops and demonstrations. N Budget is currently available on request from the author. A manual providing background information about N cycling in the northern grains region soils and describing details of the science underpinning N Budget was released in 2011 and republished in 2013 (Herridge 2013). Details of NBudget and its operation are presented herein. Sowing soil nitrate values estimated using NBudget are compared with measured data from four northern NSW--southern Queensland experiments to validate a key output of the tool.
Materials and methods
NBudget description and development
NBudget essentially uses the established N budgeting approach in which estimates of crop N demand and soil N supply arc compared in order to calculate fertiliser N requirements (Myers 1987; Marcellos and Felton 1993). NBudget could be described as a static empirical model (Myers 1987), essentially a structured set of linked mathematical equations and values that lead the user to estimates of fertiliser N needed to grow a cereal or oilseed crop in a particular paddock in the coming season. Other endpoint estimates include amounts of N fixed by the crop legumes chickpea, faba bean, mung bean and soybean. Notice was taken when developing the model of the guidelines proposed by Passioura (1996), namely that DS tools should be as simple as possible, require little input data, be based on robust empirical relationships and have use restricted to the conditions under which the empirical relationships were founded. It was considered at the outset that NBudget would be used more by advisors that by farmers themselves given the extensive role played by farm consultants in Australian agriculture (Long and Parton 2012).
The major point of difference between NBudget and the other DS tools is that soil testing for water, nitrate or organic C is not required. Rather, NBudget contains rule-of-thumb values and algorithms for estimating soil water and nitrate at sowing and in-crop N mineralisation. Input data to generate the rule-of-thumb values and algorithms were derived primarily from published and unpublished experiments conducted by the farming systems and plant (N) nutrition programs of the NSW and Queensland state departments of primary industries during the past 35 years (Herridge et al. 2010; Herridge 2013). Use was also made of unpublished data from the northern NSW farming systems sites, details of which can be found in Felton et al. (1995, 1998), Herridge et al. (1995, 1998) and Marcellos et al. (1998). The logic, assumptions and science underpinning NBudget are presented in detail in Herridge (2013). NBudget focuses on N demand and supply for grain cropping and other nutrient constraints are not considered.
NBudget was released in 2011 as an Excel-based calculator and transitioned to the NSW Department of Primary Industries (DPI) CropMate website in 2012 (Mcintosh and Herridge 2012). The CropMate website was closed down in late 2013. NBudget currently exists as an Excel file only.
There are two versions of NBudget, one for winter cropping (bread wheat, durum, barley, canola, chickpea and faba bean) and one for summer cropping (sorghum, sunflower, soybean and mung bean, with 16 localities (12 in NSW and four in Queensland) from Roma and St George in Queensland to Dubbo in the central-west of NSW. Note that all values for soil nitrate and water are to 1.2 m depth.
The location is selected in NBudget from the drop-down list of 16 towns and property and paddock names inserted (Fig. 1). In addition, the fertility status of the paddock (very low, low--medium, medium or high) is selected from a drop-down list according to the short description of each (Table 1). These descriptors are used in the look-up tables providing the estimate of sowing soil nitrate for last season's crop as well as in-crop and fallow N mineralisation rates. For both look-up tables, scalar values were 0.52 for very low fertility, 0.61 for low-medium fertility, 0.75 for medium fertility and 1.00 for high fertility. The scalar values approximate 40, 30, 20 or 10 years respectively of cropping effects on soil organic C and N and N mineralisation rates (Dalai and Probert 1997; Cox and Strong 2008; Herridge 2013; Table 1).
Soil descriptions with options of clay soil (>35% clay), red-brown earth (20-35% clay) or sand or sandy loam (<20% clay) and tillage practice (no-tillage or minimum tillage, cultivated) are also selected from drop-down menus. Scalar values for soil description in the look-up tables of sowing soil nitrate-N for last season's crop and in-crop and fallow N mineralisation rates were 1.00 for clay, 0.77 for red-brown earth and 0.60 for sand or sandy loam (Dalgliesh and Foale 1998). Scalar values for tillage practice in the same tables were 1.00 for cultivated and 0.80 for no-tillage practices (Fig. 2).
Two seasons ago
In NBudget, the user selects the crop grown in the paddock two seasons ago from a drop-down list of seven options: double crop, cereal 0 40N, cereal 50-100N, cereal 100N+, canola +N, pulse, long fallow (no crop). An estimate of soil nitrate-N at the start of the previous season (i.e. 12 months ago) is then automatically inserted from a hidden look-up table. The values in the look-up table are based on 460 observations during 11 years of N experiments across five farming systems sites in northern NSW (Fig. 3) and modified, using the scalars above, to include canola and double cropping and to account for variations in tillage, soil type and fertility level. Values in the look-up table range from a low of 11 kg nitrate-N per ha for a double-cropped, very low-fertility, no-tilled sand or sandy loam to 233 kg nitrate-N per ha for a high-fertility, cultivated clay soil that had been long fallowed.
Note that the user does not select for, or indeed need to know, the paddock's soil organic C concentration (Fig. 1). This is because soil nitrate levels are considered to be far more affected by paddock management, soil texture and previous crop history than soil organic C. In a study involving chickpea and faba bean production in northern NSW during two seasons in the 1990s, Schwenke et al. (1998) showed that there was a very poor association between soil organic C (0-15 cm) and sowing soil nitrate-N (0-90 cm) levels (Fig. 4).
The user selects from the drop-down list the crop that was last grown in the paddock and inserts the yield, protein (in the case of bread wheat, durum, barley and sorghum) and amount of fertiliser N applied (Fig. 1). If the last crop grown was a legume, the program provides an estimate of the amount of N fixed; if not a legume, the value is zero. An estimate of soil nitrate at harvest (i.e. the start of the post-crop fallow) is also provided. If that value is <20, it is adjusted to equal 20. The rationale is that not all nitrate-N to 1.2 m depth is used by crops and there will always be an unused amount. The value for that amount can vary substantially and was set at 20 kg N [ha.sup.-1] in NBudget based on 136 observations from the NSW DPI farming systems experiments in northern NSW, where the range was 7-140 kg N [ha.sup.-1] and the median value was 31 kg N [ha.sup.-1]. The stepwise set of linked algorithms to generate estimates of both fixed N and harvest soil nitrate-N are included as an appendix in Herridge (2013).
Values for in-crop and fallow mineralisation of soil organic N are taken from a look-up table, developed initially from APSlM-generatcd values for cultivated clay soils that had been cropped for varying lengths of time (Dalai and Probert 1997; J. Turpin, pers. comm.) and modified to account for differences in soil texture and fertility, tillage (see scalar values above) and for periods of low soil water (scalar of 0.75). Rates of winter in-crop and summer fallow mineralisation of background (native) soil organic N for no-tilled soils are presented in Fig. 5. Rates are substantially higher during the November-May summer fallow than during the June-October cropping period, ranging from 28-96 N [ha.sup.-1] for the summer fallow to 10-32 N [ha.sup.-1] for the in-crop (winter) period.
With fresh residues of the previous crop, it is assumed that 65-70% of residue carbon (C) is respired during the post-crop summer fallow, with the remaining 30-35% locked into stable soil organic matter with a C : N ratio of 11:1 (Ladd 1987). Depending on the C : N ratio of the fresh residues, mineral N will either be released during the fallow (~C:N<28:1), or immobilised (~C : N>28 :1; Jensen 1997). Simulated amounts of nitrate-N either mineralised or immobilised during the summer fallow from residues of the six crops in the winter NBudget arc shown in Table 2. Values range from 26 kg N [ha.sup.-1] immobilised from barley residues to the net mineralisation of 18 kg N [ha.sup.-1] from faba bean residues. Grain yields in the table are considered as typical yields for the region and were based on the transpiration efficiency values in NBudget.
Functions and values describing harvest index (HI) and N harvest index (NHI) were determined from published (e.g. Felton et al. 1998; Marcellos et al. 1998; Schwenke et al. 1998) and unpublished data from the NSW DPI farming systems experiments in northern NSW. To estimate total crop biomass from shoot biomass, a multiplication factor of 2.0 is used for chickpea (assumes 50% of plant N is below ground), 1.5 for soybean (assumes 33% below-ground N) and 1.4 for the remainder of the grain legumes, cereal and oilseed crops (assumes 30% below-ground N; e.g. Buyanovsky and Wagner 1986; Khan et al. 2003; Unkovich et al. 2010). The same multiplication factors are used to convert shoot N to total crop N.
The function estimating residue %N was derived from more than 200 values for grain protein and estimated aboveand below-ground residue-N concentrations for wheat, barley, chickpea and faba bean (Angus and Fischer 1991; Herridge et al. 1995,1998; Dalai etal. 1997; Felton etal. 1998; Marcellos et al. 1998; Turpin et al. 2002). The relationship between the two is described by the formula: Residue %N = 0.0724 x %grain protein. Residue N is then calculated in two ways: (1) by combining estimated %N with residue biomass; and (2) as the difference between total crop N (using grain N and NHI) and grain N. The final estimate of residue N is calculated as the mean of the two methods.
Current summer fallow: estimating current levels of soil nitrate-N and plant-available water
The end-of-fallow soil nitrate level (i.e. May for winter crops and October for summer crops) is estimated as the sum of adjusted or non-adjusted harvest nitrate-N, the N mineralised from background (native) soil organic N and the N mineralised or immobilised from the previous crop's residues.
NBudget also provides an estimate of end-of-fallow plant-available soil water, determined using either fallow rainfall records and fallow efficiency values, depth of wet soil and conversion factors related to texture or by other means (e.g. 'HowWet?', http://www.apsim.info/How/HowWet/how%20wet. htm, accessed 9 May 2017; Freebairn et al. 1994).
Assessment of crown rot risk
The expected level of crown rot is selected in NBudget from a drop-down menu. The yield loss for bread wheat, durum and barley is then calculated using data (ranges of grain yield loss are 0-50% for wheat, 0-80% for durum and 0-40% for barley) from the NSW DPI Grain Pathology research program, Tamworth (S. Simpfendorfer, pcrs. comm.).
Estimating grain yields and proteins, fertiliser N requirements and legume N2 fixation
Expected grain yields for the coming season are calculated automatically using a transpiration efficiency value of 12.5 kg grain mm 1 available water for bread wheat after subtracting 100 mm for evaporation (French and Schultz 1984; Felton et al. 1995, 1998; Dalai etal. 1998; Herridge et al. 1998; Marcellos et al. 1998; Edwards 2000; Sadras and Angus 2006; Hochman et al. 2009). Yields for the other winter crops are also automatically calculated by multiplying the estimated wheat yields by constants (1.05 for durum, 1.33 for barley, 0.50 for canola, 0.64 for chickpea and 0.80 for faba bean) derived from data from experiments in which all or some of these crops were grown (Felton et al. 1995, 1998; Herridge et al. 1995, 1998; Dalai et al. 1997, 1998; Marcellos et al. 1998; Edwards 2000; Thomas et al. 2007) and from the National Variety Trial (N VT) databases for 1998-2004 (http://www.nvtonline.com.au/nvt-results-reports/?nocollapseomatic= 1 #filterYear=2012&postcode=Postcode&filterCrop=&filterSubcrop=, accessed 1 December 2016). A similar approach is taken to calculate the values for the summer crops.
The user inserts the target grain proteins for the average season for bread wheat, durum, barley and sorghum. The remaining crops are set at default values (i.e. 21% for canola, 22% for chickpea and 24% for faba bean). Soil nitrate-N and fertiliser N requirements for the cereals, canola and sunflower are then automatically estimated using N-use efficiency (NUE) values, adjusted for grain protein in the case of the cereals (Herridge 2013). For example, NUE values for wheat vary from 0.58 kg grain N [kg.sup.-1] sowing nitratc-N at 10% grain protein to 0.39 kg grain N [kg.sup.-1] sowing nitrate-N at 13% protein (all grain proteins at 12% moisture). Amounts of N fixed by the legumes are also calculated, together with residual (post-fallow) nitrate levels for all crops. Fertiliser N is assumed to be converted to soil plant-available N with an efficiency of 0.8 (Strong and Cooper 1992; Pilbeam 1995; Strong et al. 1996a; Cox and Strong 2008).
Validating NBudget soil nitrate-N estimates against independent datasets
Warra (1988-96), Nindigully (1996-2001) and Cryon (1996-99)
Three independent sets of data, from the 1988-96 Warra (Dalai et al. 1998) and 1996-2001 Nindigully experiments in southern Queensland (Thomas et al. 2007) and the 1996-99 Cryon experiments in northern NSW (Edwards 2000), were used to test the accuracy of NBudget in predicting soil nitrate levels at sowing. For the Warra simulations, the clay soil was classed as very low fertility (0.65% C) after 50 years of essentially continuous wheat cropping (Dalai et al. 1995). No-tillage fallow management was assumed for the simulations. In the NBudget model, 'cultivated' means reasonably aggressive cultivation (i.e. three or more cultivations) rather than minimal cultivation (i.e. one to two cultivations). Strong et al. (1996b) had previously reported no differences between no-tillage and cultivated treatments for either sowing soil nitrate or sowing soil water in experiments at the Warra site. The crop sequences involve wheat and chickpea (n = 16).
The Nindigully clay soil was similarly classed as very low fertility (0.65% C) after 40 years cultivation and cropping. The first season of the experiment was in 1996 and the site was no-tilled. The crop sequences involved wheat and barley (n = 21).
The Cryon clay soil had an organic C level of 0.76% and had been cropped for 20 years with a mixture of cultivation and conservation tillage (Edwards 2000). For the simulations, the soil was classed as high fertility following chickpea in 1993, a drought-enforced fallow in 1994 and low-yielding wheat in 1995. Sowing soil nitrate levels in the initial experimental season of 1996 and the following 1997 season were, on average, 140 and 150 kg N [ha.sup.-1] (1.2 m depth) respectively. There were no differences between soil nitrate values for no-tilled and cultivated plots. Thus, the same logic was applied here that was applied to the Warra data and no-tillage fallow management was assumed for the simulations. The crop sequences involved wheat, chickpea, canola and long-fallow after sorghum (n = 28).
Northern NSW (2011-12)
There is no provision in NBudget for potential denitrification losses of N from flooded and waterlogged soils (Schwenke et al. 2015). NBudget has a scalar that reduces the N applied as fertiliser by 20% to account for gaseous loss (15%) and soil N immobilisation (5%), but there are no denitrification-associated losses of nitrate N (Freney and Humphreys 1987; Strong and Cooper 1992; Strong et al. 1996a; Cox and Strong 2008).
In a second validation exercise, NBudget was used to simulate sowing soil nitrates for seven N nutrition experiments across six sites in northern NSW (Herridge and Gardner 2013).
Warra (1988-96), Nindigully (1996-2001) and Cryon (1996-99)
Measured and NBudget simulated values for sowing soil nitrate-N levels at the three sites are shown in Fig. 6. Average measured and simulated soil nitrate levels for the Warra site were 67 and 74 kg N [ha.sup.-1] respectively. For Nindigully, average measured and simulated soil nitrates were 61 and 68 kg N [ha.sup.-1] respectively. For Cryon, both measured and simulated average soil nitrates were 150 kg N [ha.sup.-1]. Across the three sets of experimental data, there was good agreement between measured and simulated soil nitrate values, with the slope of the line of best fit reasonably close to unity O = 0.91x + 16.8; [r.sup.2] = 0.78).
Northern NSW (2011-12)
There was good agreement between measured and simulated values at Blackville and Walgett 1 (the difference between the two was in the range 3-15 kg N [ha.sup.-1]), only moderate agreement at Bithramere and Gurley (21-28 kg N [ha.sup.-1] differences) and very poor agreement at Moree, Tamworth and Walgett 2 (65-92 kg N [ha.sup.-1] differences; Table 3). All sites received well above average November-April rainfall during the 2011-12 summer-autumn period (i.e. +20-100%), and all may have experienced denitrification losses. There was no evidence of leaching at these sites. The sites with the greatest differences between measured and simulated soil nitrates (i.e. possibly the highest N losses) either recorded very high rainfall totals and/or were flooded for a period during the fallow.
NBudget was designed to step the user through to the end-point estimates of grain yields and fertiliser N requirements in the case of cereal and oilseed crops, and grain yields and N2 fixation inputs with legumes. As part of the process, soil nitrate-N at the time of sowing is also estimated. Within the context of grain cropping in northern NSW--southern Queensland, the sowing soil nitrate values appear to be reasonably accurate (Fig. 6), but not when denitrification losses are likely to be significant (Table 3). Accounting for denitrification losses in wet years remains challenging, a point made by northern grain advisors in a recent N management survey (G. D. Schwenke, pers. comm.).
In terms of accuracy, Armstrong et al. (2016) reported that advisors are not necessarily looking for precision in DS tools; rather, they are looking for estimations and simple rules-of-thumb (see also Lawrence et al. 2000). However, at the same time they are looking for the science behind the estimations and guidance as to the level of confidence that should be afforded those estimations. The science behind NBudget is a mixture of empirical and theoretical relationships derived from scientifically robust field experiments under a range of environmental and cdaphic conditions at several sites across the region (see Herridge 2013). In the development of the logic, values and functions in NBudget, data and knowledge from the 235 publications referenced in Herridge (2013) plus unpublished data from the NSW DPI northern NSW farming systems sites were used. The variation around any of the values and functions in NBudget is readily acknowledged (see Herridge 2013), but the reasonable agreement between observed and predicted values for sowing soil nitrate-N from the three independent datasets (Fig. 6) provides justification for the approach.
Although NBudget was designed to provide estimates of fertiliser N requirements (and N2 fixation) for the six species in the winter crop version and the four species in the summer crop version, there may be additional value in the tool's predictions of sowing soil nitrate-N levels. Soil testing for nitrate-N is recommended and surveys of farmers and advisors about N management inform us that it is considered important. However, the surveys also inform us that farmers do not routinely soil test each paddock each year, can have a low confidence in soil test results and can have a low use of soil test data, either on its own or within a DS tool framework, when making decisions about fertiliser N inputs (Long and Parton 2012; Armstrong et al. 2016; Lobry de Bruyn and Andrews 2016; G. D. Schwenke, pers. comm.).
The ultimate decision by a farmer or advisor about fertiliser N inputs is a complex one, with the amount applied determined by a range of factors, including attitude to risk, the ratio of grain price to fertiliser cost, history of fertiliser inputs for the particular paddock and the general financial status of the farmer (see also Lawrence et al. 2000; McCown et al. 2009; Long and Parton 2012). It may be that the fertiliser N rate is the decision of the farmer and/or advisor, and they alone will make it. However, there could be benefits in factoring relevant information like current soil N supply into that decision.
A second consideration for DS tools is that the tool may have a limited time frame of use until the particular relationship or determination is understood (Long and Parton 2012). Once understood in its entirety, the tool may be discarded as the users develop their own rules-of-thumb. For more complex determinations, the longevity of the DS tool is likely to be longer. Estimating soil nitrate-N without support from a valid soil test could be regarded as one of those complex determinations. Clearly, ongoing support for users is critical for continuation of use.
As stated earlier, the winter and summer versions of NBudget were key components of the NSW DPI CropMate website during 2012-13 (Mcintosh and Herridge 2012). The CropMate website was closed down in late 2013 following extensive restructuring within the agency. NBudget currently exists as Excel files only. Although there appears to be ongoing use of NBudget in NSW and Queensland (G. D. Schwenke, pers. comm.), its usefulness is limited if it remains as Excel files. The files are currently available on request from the author. NBudget would have a greater chance of longevity and more widespread use as a stand-alone app or web-based tool.
The motivation for developing NBudget was to assist northern grains farmers make more informed decisions about fertiliser N rates for cereal and oilseed crops. The fact that many paddocks are not deep cored for soil nitrate testing before sowing means that farmers need to estimate soil nitrate levels at sowing to determine the likely fertiliser requirement for target grain yields and proteins. Published and unpublished data from more than 35 years of research trials in the region were used in the development of NBudget, and considerable effort went into structuring the calculator to be logical and simple to use. Comparing N Budget-derived estimates of sowing soil nitrate levels with measured values indicated generally good agreement, although NBudget does not account for losses from waterlogged or flooded soils.
The development of NBudget drew, to a large extent, on the work of scientists in the NSW and Queensland departments of primary industries, notably Warwick Felton, David Doyle, Ian Holford and the late Harry Marcellos in NSW and Wayne Strong, Ram Dalai, Greg Thomas and David Freebaim in Queensland. The late Chris Cole of the NSW DPI provided much needed impetus for this project at critical times. Ms Caihua (Wei) Chen is acknowledged for her skilful programming of the Excel and web-based versions of NBudget. Financial support from the Grains Research & Development Corporation (GRDC) is also gratefully acknowledged.
Australian Bureau of Agricultural and Resource Economics (ABARES) (2016) Australian crop report: September 2016. No. 179. Available at http://www.agriculture.gov.au/abares/display?url=http://l 220.127.116.11/ anrdl/DAFFService/display.php?fid=pb_aucrpd9aba_20160913_Nmx JS.xml [verified I December 2016].
Angus JF, Fischer RA (1991) Grain and protein responses to nitrogen applied to wheat growing on a red earth. Australian Journal of Agricultural Research 42, 735-746. doi: 10.1071/AR9910735
Armstrong R, Wallace A, Dunsford K, Barton L, Bell M, Burke K (2016) Nitrogen management in the southern region practices facts and gaps. Grains Research & Development Corporation, Canberra, ACT. Available at https://grdc.com.au/Research-and-DevelopmentyGRDC-Update-Papers/ 2016/02/Nitrogen-management-in-the-southern-region-practices-facts-and-gaps [verified 15 November 2016],
Bell M J, Strong W, Elliot D, Walker C (2013) Soil nitrogen-crop response calibration relationships and criteria for winter cereal crops grown in Australia. Crop and Pasture Science 64, 442-160. doi:10.1071/CP12431
Buyanovsky GA, Wagner GH (1986) Post-harvest residue input to cropland. Plant and Soil 93, 57-65. doi:10.1007/BF02377145
Carberry PS, Hochman Z, Hunt JR, Dalgliesh NP, McCown RL, Whish JPM, Robertson MJ, Foale MA, Poulton PL, van Rees H (2009) Re-inventing model-based decision support with Australian dryland fanners. 3. Relevance of APSIM to commercial crops. Crop and Pasture Science 60, 1044-1056. doi:10.1071/CP09052
Cox H, Strong W (2008) 'The nitrogen book: principles of soil nitrogen fertility management in southern Queensland and northern New South Wales farming systems.' (Department of Employment, Economic Development and Innovation: Brisbane, Qld)
Dalal RC, Probert ME (1997) Soil nutrient depletion. In 'Sustainable crop production in the sub-tropics'. (Eds AL Clarke, PB Wylie) pp. 42-63. (Department of Primary Industries: Brisbane, Qld)
Dalal RC, Strong WM, Weston EJ, Cooper JE, Lehane KH, King AJ, Chicken CJ (1995) Sustaining productivity of a Vertisol at Warra, Queensland, with fertilisers, no-tillage or legumes 1. Organic matter status. Australian Journal of Experimental Agriculture 35, 903-913. doi: 10.1071 /EA9950903
Dalal RC, Strong WM, Doughton JA, Weston EJ, McNamara GT, CooperJE (1997) Sustaining productivity of a Vertisol at Warra, Queensland, with fertilisers, no-tillage or legumes 4. Nitrogen fixation, water use and yield of chickpea. Australian Journal of Experimental Agriculture 37, 667-676. doi: 10.1071/EA97018
Dalal RC, Strong WM, Doughton JA, Weston EJ, Cooper JE, Wildermuth GB, Lehane KJ, King AJ, Holmes CJ (1998) Sustaining productivity of a Vertisol at Warra, Queensland, with fertilisers, no-tillage or legumes 5. Wheat yields, nitrogen benefits and water-use efficiency of chickpea-wheat rotation. Australian Journal of Experimental Agriculture 38, 489 501. doi:10.1071/EA98027
Dalgliesh N, Foale M (1998) Soil matters. Monitoring soil water and nutrients in dryland farming. CSIRO: Toowoomba, Australia. Available at https://www.apsim.info/Portals/0/APSoil/Soil%20matters. pdf [verified 20 November 2016]
Edwards J (Ed.) (2000) 'Western farming systems project trial results 1996-1999.' (NSW Agriculture: Orange, NSW)
Elias NV, Herridge DF (2014) Crop-available water and agronomic management, rather than nitrogen supply, primarily determine grain yield of commercial chickpea in northern New South Wales. Crop and Pasture Science 65, 442-152. doi: 10.1071/CP13397
Felton WL, Marcellos H, Martin RJ (1995) A comparison of three fallow management strategies for the long-term productivity of wheat in northern New South Wales. Australian Journal of Experimental Agriculture 35, 915-921. doi : 10.1071 /EA9950915
Felton WL, Marcellos H, Alston C, Martin RJ, Backhouse D, Burgess LW, Herridge DF (1998) Chickpea in wheat-based cropping systems of northern New South Wales. II. Influence on biomass, grain yield, and crown rot in the following wheat crop. Australian Journal of Agricultural Research 49, 401-107. doi:10.1071/A97067
Freebairn DM, Hamilton AH, Cox PG, Holzworth D (1994) HowWet? Agricultural Production Systems Research Unit, Toowoomba, Queensland.
French RJ, Schultz JE (1984) Water use efficiency of wheat in a Mediterranean-type environment. I. The relationship between yield, water use and climate. Australian Journal of Agricultural Research 35, 765-775.
Freney JR. Humphreys E (1987) Gaseous losses of nitrogen from agricultural systems in temperate Australia. In 'Nitrogen cycling in temperate agricultural systems'. (Eds PE Bacon, J Evans, RR Storrier, AC Taylor) pp. 297-315. (Australian Society of Soil Science: Wagga Wagga, NSW)
Herridge DF (2013) Managing legume and fertiliser N for northern grains cropping. Revised 2013 version. Grains Research & Development Corporation. Canberra, ACT. Available at https://grdc.com.au/~/ media/6E5659619C7C4063AB3C8E58A4DE39E7.pdf [verified 9 May 2017],
Herridge D, Gardner M (2013) Denitrification contributing to crop N deficiencies in 2012: analysis using 'NBudget' and soil test data. In 'Northern grains region trial results Autumn 2013'. (Eds L Serafin, S Simpfendorfer, M Gardner, G McMullen) pp. 187-189. (NSW Department of Primary Industries: Orange, NSW). Available at http:// www.dpi.nsw.gov.au/_data/assets/pdf_file/0004/468328/Northem-gra ins-region-trial-results-autumn-2013.pdf [verified 5 November 2016].
Herridge DF, Marcellos H, Felton WL, Turner GL, Peoples MB (1995) Chickpea increases soil-N fertility in cereal systems through nitrate sparing and N2 fixation. Soil Biology & Biochemistry 27, 545-551. doi: 10.1016/0038-0717(95)98630-7
Herridge DF, Marcellos H, Felton WL, Turner GL, Peoples MB (1998) Chickpea in wheat-based cropping systems of northern New South Wales. III. Prediction of N2 fixation and N balance using soil nitrate at sowing and chickpea yield. Australian Journal of Agricultural Research 49, 409-418. doi:10.1071/A97068
Herridge D, Belfield S, Serafin L (2010) 'NBudget'--a nitrogen management tool for cropping systems. In 'Food security from sustainable agriculture. Proceedings of the 15th Australian Agronomy Conference 2010', 14-18 November 2010, Lincoln, New Zealand. (Eds H Dove, R Culvenor) Available at http://www.regional.org.au/au/asa/20 10/crop-production/nutrients/7174_herridge.htm#TopOfPage [verified 15 November 2016],
Hochman Z, Holzworth D, Hunt JR (2009) Potential to improve on-farm wheat yield and WUE in Australia. Crop and Pasture Science 60, 708-716. doi:10.1071/CP09064
Jensen ES (1997) Nitrogen immobilization and mineralization during initial decomposition of [sup.13]N-labclled pea and barley residues. Biology and Fertility of Soils 24, 39-44. doi:10.1007/BF01420218
Khan DF, Peoples MB, Schwenke GD, Felton WL, Chen D, Herridge DF (2003) Effects of below-ground nitrogen on N balances of field-grown fababean, chickpea and barley. Australian Journal of Agricultural Research 54, 333-340. doi: 10.1071/AR02105
Ladd JN (1987) Mineralization and immobilization of nitrogen. In 'Nitrogen cycling in temperate agricultural systems'. (Eds PE Bacon, J Evans, RR Storrier, AC Taylor) pp. 198-207. (Australian Society of Soil Science: Wagga Wagga, NSW)
Lawrence DN, Cawley ST, Hayman PT (2000) Developing answers and learning in extension for dryland nitrogen management. Australian Journal of Experimental Agriculture 40, 527-539. doi: 10.1071/ EA99147
Lobry de Bruyn L, Andrews S (2016) Are Australian and United States farmers using soil information for soil health management? Sustainability 8, 304-336. doi:10.3390/su8040304
Long B, Parton K (2012) Decision support systems (DSS) - where success is failure of continued use. In 'Capturing opportunities and overcoming obstacles in Australian agronomy. Proceedings of the 16th Australian Agronomy Conference', 14-18 October 2012, Armidale, NSW. Available at http://www.regional.org.au/au/asa/2012/crop-production/81 53Jongw.htm#TopOfPage [verified 15 November 2016],
Marcellos H, Felton WL (1993) Wheat yield targets, and water and nitrogen use efficiency in northern New South Wales. In 'Farming from paddock to plate. Proceedings of the 7th Australian Agronomy Conference', 19-24 September 1993, Adelaide. Available at http:// www.regional.org.au/au/asa/1993/concurrent/crop-nutrition/p-02.htm# TopOfPage [verified 15 November 2016],
Marcellos H, Felton WL, Herridge DF (1998) Chickpea in wheat-based cropping systems of northern New South Wales. I. N2 fixation and influence on soil water and nitrate. Australian Journal of Agricultural Research 49, 391-400. doi:10.1071/A97066
McCown RL, Carberry PS, Hochman Z, Dalgliesh NP, Foale MA (2009) Re-inventing model-based decision support with Australian dryland farmers. 1. Changing intervention concepts during 17 years of action research. Crop and Pasture Science 60, 1017-1030. doi:10.1071/ CP08455
Mcintosh G, Herridge D (2012). CropMate[TM] a web based decision support tool helping farmers make agronomic decisions using historic and forecast weather and climate data. In 'Capturing opportunities and overcoming obstacles in Australian agronomy. Proceedings of 16th Australian Agronomy Conference', 14-18 October 2012, Armidale, NSW. Available at http://www.regional.org.au/au/asa/2012/precisionagriculture/8235_mcintosh.htm#TopOfPage [verified 10 May 2017].
Myers RJK (1987) Modelling the behaviour of nitrogen in soil-plant systems. In 'Nitrogen cycling in temperate agricultural systems'. (Eds PE Bacon, J Evans, RR Storrier, AC Taylor) pp. 397-425. (Australian Society of Soil Science: Wagga Wagga, NSW)
Passioura JB (1996) Simulation models: science, snake oil, education or engineering? Agronomy Journal 88, 690-694. doi: 10.2134/agronj1996. 00021962008800050002X
Pilbeam CJ (1995) Effect of climate on the recovery in crop and soil of l;,N-labelled fertilizer applied to wheat. Fertilizer Research 45, 209-215. doi: 10.1007/BF00748591
Sadras VO, Angus JF (2006) Benchmarking water-use efficiency of rainfed wheat in dry environments. Australian Journal of Agricultural Research 57, 847-856. doi:10.1071/AR05359
Schwenke GD, Peoples MB, Turner GL, Herridge DF (1998) Does nitrogen fixation of commercial, dryland chickpea and faba bean crops in north-west New South Wales maintain or enhance soil nitrogen? Australian Journal of Experimental Agriculture 38, 61-70. doi: 10.1071/EA97078
Schwenke GD, Herridge DF, Scheer C, Rowlings DW, Haigh BM, McMullen KG (2015) Soil N20 emissions under N2-fixing legumes and N-fertilised canola: a reappraisal of emissions factor calculations. Agriculture, Ecosystems & Environment 202, 232-242. doi: 10.1016/ j.agee.2015.01.017
Schwenke GD, Herridge DF, Scheer C, Rowlings DW, Haigh BM, McMullen KG (2016) Greenhouse gas (N20 and CH4) fluxes under nitrogen-fertilised dryland wheat and barley on sub-tropical Vertosols: risk, rainfall and alternatives. Soil Research 54, 634-650. doi: 10.1071/ SR15338
Strong WM, Cooper JE (1992) Application of anhydrous ammonia or urea during fallow period for winter cereals on the Darling Downs, Queensland. I. Effect of time of application on soil mineral N at sowing. Australian Journal of Soil Research 30, 695-709. doi: 10.1071/SR9920695
Strong WM, Dalai RC, Weston EJ, Cooper JE, Lehane KJ, King AJ (1996a) Nitrogen fertiliser residues for wheat cropping in subtropical Australia. Australian Journal of Agricultural Research 47, 695-703. doi: 10.1071/AR9960695
Strong WM, Dalai RC, Weston EJ, Cooper JE, Lehane KJ, King AJ, Chicken CJ (19966) Sustaining productivity of a Vertisol at Warra, Queensland, with fertilisers, no-tillage or legumes 2. Long-term fertiliser nitrogen needs to enhance wheat yields and grain protein. Australian Journal of Experimental Agriculture 36, 665-674. doi:10.1071/EA9960665
Thomas GA, Dalai RC, Weston EJ, Holmes CJ, King AJ, Orange DN, Lehane KJ (2007) Zero tillage and nitrogen fertiliser application in wheat and barley on a Vertosol in a marginal cropping area in southwest Queensland. Australian Journal of Experimental Agriculture 47. 965-975. doi : 10.1071 /EA05253
Turpin JE, Herridge DF, Robertson MJ (2002) Nitrogen fixation and soil nitrate interactions in field-grown chickpea (Cicer arietinum) and fababean (Vicia faba). Australian Journal of Agricultural Research 53, 599-608. doi:10.1071/AR01136
Unkovich M, Baldock J, Forbes M (2010) Variability in harvest index of grain crops and potential significance for carbon accounting: examples from Australian agriculture. Advances in Agronomy 105, 173-219. doi: 10.1016/S0065-2113(10)05005-4
David F. Herridge
School of Environmental and Rural Science, University of New England, Armidale, NSW 2351, Australia. Email: email@example.com
Received 20 January 2016, accepted 2 May 2017, published online 24 May 2017
Caption: Fig. 1. An example of the output worksheet of the winter crop NBudget showing estimated levels of sowing soil nitrate (87 kg N [ha.sup.-1]) and water (124 mm) and predicted yields, sowing nitrate-N and fertiliser N requirements for bread wheat, durum, barley, canola. Predicted yields and N2 fixation inputs are shown for chickpea and faba bean.
Caption: Fig. 2. Published data (Felton et al. (1995, 1998), Herridge et al. (1995, 1998) and Marcellos et al. (1998) and unpublished data from the New South Wales (NSW) Department of Primary Industries long-term farming systems experiments in northern NSW showing effects of tillage on the accumulation of soil nitrate-N during the summer fallow (1.2m depth; n = 208).
Caption: Fig. 3. Published data (Felton et al. (1995, 1998), Herridge et al. (1995, 1998) and Marcellos et al. (1998) and unpublished data from the New South Wales (NSW) Department of Primary Industries long-term farming systems experiments in northern NSW showing the effects of tillage and previous paddock history on soil nitrate levels at sowing (1.2m depth; n = 460). Median soil nitrate levels at sowing for Paddock Histories I 5 (no-till) were 56, 64, 76, 100 and 136 kg N [ha.sup.-1] respectively; those for Paddock Histories 6-10 (cultivated) were 77,63, 105, 108 and 194 kg N [ha.sup.-1] respectively. Previous crops and paddock histories are as follows: 1 and 6, cereal with 0-40kg fertiliser N [ha.sup.-1]; 2 and 7, cereal with 50-100 kg fertiliser N [ha.sup.-1]; 3 and 8, cereal with >100kg fertiliser N [ha.sup.-1]; 4 and 9, pulse; 5 and 10, no crop or long fallow.
Caption: Fig. 4. Data used from Schwenke et at. (1998) of on-farm surveys of chickpea and faba bean cropping during the 1994 and 1995 seasons in northern New South Wales show no relationship between soil organic C (0-15 cm depth) and soil nitrate-N (0-90 cm depth) at sowing (n = 50).
Caption: Fig. 5. APSIM-simulated values for mineralisation of background (native) soil organic N during (a) the post-harvest summer fallow and (h) in-crop for the three soil types based on texture (clay, red brown earth and sand or sandy loam) and for the four classes of soil fertility. Data shown are for no-tilled soils and winter cropping.
Caption: Fig. 6. Measured and N Budget-simulated sowing soil nitrate (0-1.2 m depth) levels from the Warra (n = 16) and Cryon (n = 28) farming systems and Nindigully (n = 21) agronomy experiments. Crops in the various sequences included wheat, barley, chickpea, canola and sorghum. The 1:1 line is shown. The line of best fit (not shown) is described by the equation y = 0.91x+ 16.8 ([r.sup.2] = 0.78).
Table 1. Paddock soil descriptions of fertility status used in NBudget The approximate years of cropping equivalence for the four levels of soil fertility were derived from studies of long-term cropping and cultivation effects of soil properties in southern Queensland (Dalai and Probert 1997). The final three columns show the approximate range of values for soil organic C according to rainfall (e.g. low rainfall at Walgett. Dubbo, St George; medium rainfall at Gunnedah. Moree, Goondiwindi; and high rainfall at Tamworth, Croppa Creek, Dalby) Soil Description Approximate fertility equivalence in status cropping (years) Very low Long cropping history; 40 negligible inputs of N Low-medium Long cropping history; low 30 moderate inputs of N Medium Short cropping history; 20 moderate-high inputs of N High High inputs of N via legume 10 pasture leys, pulses and fertiliser; high-level management Soil Approximate equivalence fertility n soil organic C levels status (0-10 cm; %) Low Medium High rainfall rainfall rainfall Very low 0.4 0.7 0.8 Low-medium 0.55 0.8 1.0 Medium 0.7 1.0 1.3 High 0.9 1.3 1.8 Table 2. Simulated net mineralisation or immobilisation of crop residues during the post-harvest summer fallow Relativities of the grain yields of the different crops are taken from NBudget. Transpiration efficiencies relative to wheat, before adjusting barley and durum yields for losses from crown rot, in NBudget are 1.05 for durum, 1.33 for barley, 0.50 for canola, 0.64 for chickpea and 0.80 for faba bean Crop species Grain yield Grain protein N released or (t [ha.sup.-1]) (%) immobilised (kg [ha.sup.-1]) Wheat 3.0 11.5 -19 Barley 4.0 10.5 26 Durum 3.0 13.0 -12 Canola 1.5 23.0 +8 Chickpea 1.9 21.8 +17 Faba bean 2.4 23.9 +18 Table 3. Measured and N Budget-simulated sowing soil nitrate levels from N nutrition experiments in northern New South Wales during 2012 Soil nitrate-N values are to 1.2 m depth Site Soil fertility Sowing soil nitrate (kg N [ha.sup.-1]) status Measured Simulated Difference (A) Walgett 1 Medium 116 119 -3 Blackville Medium 43 58 -15 Gurley Low-medium 81 102 -21 Bithramere Medium 55 83 -28 Moree Low -medium 51 116 -65 Tamworth Medium 52 118 -66 Walgett 2 Medium 56 148 -92 Site Presowing November 2011-April 2012 rainfall (mm) Walgett 1 562 Blackville 466 Gurley 776 Bithramere 482 Moree 742 Tamworth 690 Walgett 2 562 (A) The difference in sowing soil nitrate was calculated as measure simulated.
|Printer friendly Cite/link Email Feedback|
|Author:||Herridge, David F.|
|Date:||Aug 1, 2017|
|Previous Article:||Experimental validation of a new approach for rice fertiliser recommendations across smallholder farms in China.|
|Next Article:||Soil mineral nitrogen benefits derived from legumes and comparisons of the apparent recovery of legume or fertiliser nitrogen by wheat.|