Phosphorus nutrition in the Birchip Cropping Group Farming Systems Trial
Ben Jones, Birchip Cropping Group and Mallee Focus
Introduction
Input management is challenging for farmers growing crops on clay textured soils in a low-rainfall environment. Clay soils in low-rainfall environments typically have high inherent nitrogen fertility, and the potential to grow high yielding crops in wetter years. In drier years, high evaporation may result in low yield, excess fertility lowering yields further through excess transpiration when water is available during times of the year unfavourable for yield (eg. winter). Phosphorus nutrition interacts with nitrogen nutrition, and can limit the potential ‘up-side’ in wet years, but in a relatively dry environment farmers are reluctant to invest in phosphorus because of the uncertain payoff. The association of nitrogen with phosphorus in the cheapest form of P fertiliser, MAP, means that maintaining soil P status may introduce unwanted extra nitrogen.
There are a number of schemes for managing phosphorus, including Colwell P soil tests, Colwell P with the aid of the Phosphorus Buffer Index, and phosphorus budgeting techniques. There are also a range of farming systems that might be used in these areas. The Birchip Cropping Group farming systems trial has been running since 1999, comparing four ‘farming systems’ under the management of farmer champions, with a standard rotation on a 32 hectare site. There have been some measurements of soil phosphorus status initially and in recent years, along with some plant measurements and a good record of seed and fertiliser inputs, and grain and lifestock outputs. The aim of this work was to demonstrate differences in phosphorus balance between systems in the BCG farming systems trial, to look for signs of phosphorus deficiency or surplus, and to consider whether any of the possible phosphorus management schemes would give better results.
Methods
Site history
The BCG farming systems trial began with a setup year in 1999, being previously sown to triazine tolerant canola (1998, 0.2 t/ha yield), barley (1997, 0.8 t/ha, and 1995, 3 t/ha), and medic pasture (1996 and 1994). The barley crops were sown with 80 kg/ha DAP (16P, 14.4N), medic with 6P as single superphosphate, and canola with 80 kg/ha grain legume superphosphate + zinc (13.4P), 100 kg/ha urea (43N), and 1 t/ha gypsum. One gypsum dump is on the fence between plots 15 and 16 (check). Barley crops were sown with pre-emergent trifluralin, and post-emergent MCPA and metsulfuron methyl (Ally). The canola crop was sown with simazine (0.9l/ha), atrazine (0.6l/ha), the grass herbicide butroxydim + fluazifop-p (Fusion, 200g/ha) and endosulphan.
Experiment
The experiment is laid out in plots of approximately 1 hectare each, in 4 rows of 8 columns of plots. Each ‘system’ has 5 plots, with a 4-year standard rotation that has 3 replicates interspersed regularly throughout (http://www.malleefocus.com.au/data/systemstrial.aspx). A conspicuous feature of the site is Gilgai (known locally as ‘crab-holes’), bowl-shaped depressions about 10-20m in diameter that are characteristic of the shrinking/swelling calcarosol soil (soils characterised in plot 8: http://www.dpi.vic.gov.au/dpi/vro/malregn.nsf/pages/mallee_soil_birchip_lwa20 and in plot 27: http://www.dpi.vic.gov.au/dpi/vro/malregn.nsf/pages/mallee_soil_birchip_lwa21).
Systems
Since 1999, crops have been sown according to the rules established by the respective system ‘champions’, apart from the ‘standard’ rotation, which follows a regular wheat – pea – canola – fallow sequence (Table 1). The main features of the systems are:
Fuel burner – mainly cereals, regular use of tilled fallow (1-2/5 plots), commenced prior to harvest, low intensity livestock mainly for fat lambs, full-disturbance tillage at sowing.
Hungry sheep – intensive cropping (mainly cereals) and intensive grazing, winter lambing with stocking rate decided in May and feeding to fill feed gap, sheep trading over summer to take advantage of stubbles and control weeds, early sown cereal/pasture forage for feed (1/5 plots), generally full-disturbance tillage at sowing.
No till – minimum soil disturbance seeding with knife points and press-wheels on 30cm spacing (has varied prior to 2007), no livestock, initially high use of break crops, now mainly cereals and some chemical fallow (commenced prior to harvest).
Reduced till – flexible approach, can use tillage/full disturbance sowing but has mainly been chemical weed control and same seeding system as No till, mix of cereals, canola and lower-value break crops, some livestock on agistment over summer.
Livestock grazing has also been a feature of the trial. All systems except no till have been grazed up to 2006, since when most sheep have been in the hungry sheep system (pattern of grazing in Figure 1, overall rates in Figure 2).
Table 1. Crops sown and yields in the BCG farming systems trial, 1999-2008. The yields for 2006-8 are for the south half of each plot. {would be better to give original system yields}
Figure 1. Diagram summary of livestock movements in the BCG farming systems trial, to end 2008. Plot zero is a feedlot off-site.
Figure 2. Level of stocking across all paddocks in each system during the trial. Note that the y-axis is logarithmic. The standard plots had 127.5 DSE for 11 days in July 2000 (not on the chart).
Treatments
In addition to the systems, plots were split in half and additional treatments imposed on one half of each plot in 2006 to examine particular issues in the trial. These were:
No-till versions of ‘till’ systems – identical to the fuel burner and hungry sheep systems, except weed control is exclusively chemical, crops are sown with knife-points, press-wheels and incorporated-by-sowing herbicides, and in-crop herbicides targeted to particular problems on each plot half.
Tillage in no-till systems – a selection of plots from the no till and reduced till systems had single, dry cultivations before sowing in 2006 or 2007, but have otherwise been managed identically. These were intended to demonstrate any immediate or on-going negative (or positive) consequences of tactical tillage in an otherwise no till system (for example to remove woody weeds).
Straw in no-till systems – the balance of no-till and reduced-till plots had a single 5 t/ha application of wheat straw in January 2007, initially in windrows, spread before the next major rain in March 2007. These were intended to demonstrate the degree to which lack of biomass production affects the benefits of no-till systems.
Where relevant, measurements from the treatments are reported on individually, otherwise results refer to the half-plot that continues the original system.
Measurements
Soil sample locations have been standardised since November 2005, a transect of four soil core locations 15m apart from corner to corner of each plot. Six topsoil (0-10 cm) samples were taken randomly around each deep core location, bulked and sub-sampled for oven-drying at 40ºC. The sampling pattern was not recorded for earlier samples. Plant harvests were made on the same transects, 10 random 0.5m lengths of row at GS30 for each crop. Soil samples were analysed for Colwell P and PBI by CSBP, Bibra Lake, Western Australia. Plant samples were oven-dried at 60ºC, then analysed for total P using ICP-AES after extraction with nitric acid and hydrogen peroxide (also CSBP).
Phosphorus budget
A phosphorus budget for each half-plot was constructed using actual fertiliser inputs, and estimates of phosphorus in seed, grain, hay, sheep feed and meat (Table 2), according to actual inputs and yields from each plot.
Table 2. Assumptions of phosphorus content for budget calculations.
Phosphorus in sheep meat was estimated at 7.5 kg/kg empty live weight (Bellof and Pallauf 2007), vetch hay phosphorus from 2007 and 2008 measurements of vetch hay on focus paddocks, remaining grain largely derived from Bolland et al. (1999).
Plant phosphorus status
Plant phosphorus status was calculated from plant nitrogen and phosphorus uptake at GS30, and the Sadras (2006) relationship between N and P for crops at maximum yield. The equation (N uptake as a function of P uptake at maximum yield) was transposed to get phosphorus uptake that matched nitrogen uptake at maximum yield. Phosphorus status was then expressed as actual plant phosphorus uptake as a proportion of what would be required to match nitrogen uptake, at the optimum N:P ratio.
Results
Phosphorus balance in farming systems
Individual plots
Phosphorus balance calculated in the BCG systems trial reflects how actual crop performance, in the long term, has lived up to champion expectations. For all the systems apart from the Fuel burner (Figure 3a, rest of systems b-d), balance was only maintained in the initial, wet years of the trial (1999-2001), and since then has trended up. On the whole the Fuel burner system maintained balance to approximately 2004, since when it has also begun trending up.
Quite marked differences have developed between the plots within each system over time, and these could be related to quirks of crop selection x year, or spatial variation. Particular plots stand out: balance has been best in fuel burner plot 18, and worst in no till paddocks 6 and 16, and reduced till 3 and 14. Paddocks 5 and 13 have been most balanced of the hungry sheep system.
Figure 3. Phosphorus balance in individual paddocks of the fuel burner (a) and hungry sheep (b) systems at the BCG farming systems trial.
Figure 3 continued. Phosphorus balance individual paddocks of the no till (c) and reduced till (d) systems at the BCG farming systems trial.
The standard rotation has three replicates of each phase of a fixed, 4 year rotation, and demonstrated the relative importance of spatial variation and the quirks of year x crop selection. There has been very little variation between replicates (Figure 4), but one phase has been separated from the rest (plots 7, 17 and 28, green line) by coinciding wheat crops with drought years of 2002 and 2006, and canola with 2004 and 2008.
Figure 4. Phosphorus balance in plots of the standard rotation in the BCG farming systems trial.
Whole systems
When phosphorus balance was averaged across all plots within a system, the distinction between systems was clear (Figure 5a). Most of the separation between systems occurred between 2000-2003 (Figure 5b), when the no till and standard systems had much higher net phosphorus balance than the fuel burner system, and greater phosphorus use in 2003 separated the hungry sheep and reduced till systems from the standards. Since 2003, all systems have been tracking approximately parallel (Figure 5a).
Figure 5. Average phosphorus balance in each system, cumulative at harvest between 1999 and 2008 (a) and on a yearly basis (b). The error bars are standard error of the mean (n=5 plots for systems except standard, n=12).
The main flows of phosphorus have been inwards in fertiliser, and outwards in hay and grain, both decreasing over time (Figure 6). Seed phosphorus input has been relatively small, as has export in lambs from the hungry sheep and fuel burner livestock, and liveweight gains of trade sheep. Feed for sheep has made a small but significant phosphorus input into the hungry sheep system in 2004-6.
Figure 6. Components of the yearly phosphorus balance of each system.
Soil phosphorus in different systems
There is a range in phosphorus balance both between systems on average, and between plots within systems. Changes in Colwell P measurements since the experiment began would in theory be related to changes in phosphorus balance, to the extent that changes in phosphorus balance led to changes in available phosphorus, and to the extent that Colwell P could measure some proxy of phosphorus available to plants. An initial question is whether there may be some variation in the phosphorus buffer index (PBI) related to soil variation within the site, and whether that ought to be considered along with Colwell P measurements.
Relationship between Colwell P and PBI
Colwell P and PBI were measured together twice at the site, on 22 of the plots in March 2004, and on all 52 subplots in March 2008. Using the Moody (2007) estimate of a critical Colwell P for 90% maximum wheat yield (a function of PBI), Colwell phosphorus was insufficient for 90% maximum wheat yield in about half of plots measured, and not correlated to PBI measurements (Figure 7a, b). In 2004, two hungry sheep plots had very high Colwell P but there were otherwise few differences (not all plots were measured). In 2008 all systems had a similar range of Colwell P, apart from fuel burner which had more low Colwell P plots. Many of the standard plots had low Colwell P. There was also no correlation between 2004 and 2008 measurements (Figure 7c), which may reflect both spatial variation within plots, and the coarse nature of the 2004 measurements (rounded to nearest 10, versus 1 decimal place in 2008).
Figure 7. Critical Colwell P for 90% maximum yield calculated from PBI measurements in 2004 (a) and 2008 (b), in relation to simultaneous Colwell P measurements, and compared between 2004 and 2008 (c).
Colwell P time trend
Colwell P was only measured on a few plots before the experiment commenced in 1999. Measurements at that time show considerable variation across the site, from 7.6 in one of the standards, up to 20 in a no till plot (Figure 8). None of the plots in 1999 had Colwell P that would be considered sufficient for 90% maximum wheat yield according to the criteria of Moody (2007). Colwell P increased between 1999 and 2004, particularly in the fuel burner and hungry sheep plots. There were sharp falls between then and 2005 in the fuel burner and hungry sheep plots, after which measurements were similar in March 2008, and decreased in the fuel burner plots.
Figure 8. Trend in Colwell P for plots measured in May 1999, before treatments commenced.
Colwell P vs P balance – absolute
On later measurements, it is possible to compare Colwell P with P balance on individual plots, although not all plots were measured at a single sampling until March 2008. At each sampling, nearly half (or more) plots had Colwell P below the critical level for wheat yield at the site (Figure 9a-c). There was a greater spread of Colwell P values in March 2004 and 2008; in 2005 Colwell Ps were within a narrower band. There was little overall relationship between Colwell P and phosphorus balance in 2004 and 2005, but in 2008 there was an approximate relationship across all plots. A few plots have high ‘leverage’ in the relationship: the very high phosphorus balance plots in the no till and standard systems, and the low phosphorus balance plot 18 in the fuel burner system. Without these there would be little suggestion of an increasing relationship.
Considering particular systems, the no till system was the only system to have an increasing relationship between phosphorus balance and Colwell P in all three years (more so in 2004 and 2008 than 2005). Apart from that, there were no consistent Colwell P – phosphorus balance relationships between plots of a particular system, at any sampling.
Figure 9. Relationship between Colwell P measured in March 2004 (a), 2005 (b) and 2008 (c) and phosphorus balance at the same time in the plots of the BCG farming systems trial. The dashed line is the average critical P for 90% maximum yield calculated according to Moody (2007).
Colwell P vs P balance – changes
Spatial variation in initial phosphorus is one possible explanation for the lack of relationship between Colwell P and phosphorus balance, apart from within-plot spatial variance and other causes of sampling error. There was a general relationship between increase in Colwell P and increase in phosphorus balance, when all changes for consecutive sampling were considered (Figure 10b), and also for changes between non-consecutive samplings (Figure 10a). A linear regression of change in Colwell P modelled by change in phosphorus balance is Change in Colwell P (mg/kg) = -11.21±3.65 + change in P (kg/ha) x 0.97±0.23 (n=48, p<0.001, R2=0.25). Thus a change in Colwell P of one unit reflects a change in P balance of approximately 1 kg P/ha (not far from the 1.3-1.5 kg/ha that would be predicted from bulk density x Colwell P). The model also implies that Colwell P will decrease unless the P balance is positive; the model estimates the necessary ‘base’ P balance as +11.5 kg P/ha. This suggests some sort of fixation process, although combining three quite different periods (1999-2004, 2004-5 and 2005-8) should give an erroneous estimate of a fixation process that occurs in response to time, temperature, rainfall, or a combination of all.
Figure 10. Change in Colwell P between samplings related to change in phosphorus balance. The graph contains (a) changes between non-consecutive (1999-2005, 1999-2008, 2004-2008) samplings or (b) changes between consecutive samplings (1999-2004, 2004-2005, 2005-2008).
When the component periods are examined, the points from the first period (Figure 11a) contain too little range compared to the variation in data to show any useful trend, having both a big increase in Colwell P and phosphorus balance. The second period appears to have a lower x-intercept (7-10 kg P cf. 15-20 kg P, by eye) and possibly higher slope than the third. Unfortunately there is too much variation in the data to model this (no significant difference in slope or offset between periods).
Figure 11. Change in Colwell P between samplings related to change in phosphorus balance, for the period 1999-2004 (a), 2004-2005 (b) and 2005-2008 (c).
Effect of tillage and straw
Between 2006 and 2007, a number of new treatments were imposed on plot halves: no-till version of the fuel burner and hungry sheep systems, and tillage and additional straw in the no till and reduced till systems. At the time Colwell P was measured (March 2008), most fuel burner and hungry sheep plots had been sown to two years of no-till crops, apart from plot 13 which was sown to oats in both years on both halves with the no-till seeder, but was cultivated once in December 2006 to control erosion (same as 26). Reduced till and no till plots had been tilled once, either in autumn 2006 (3, 11 and 14) or autumn 2007 (22 and 30). Plots with straw treatments had been under straw for a year.
Tilled fuel burner plots, the two deep tilled hungry sheep plots, and reduced till plot 30 all had higher Colwell P on the no till plot half (Figure 12). Tillage in the no till system, in contrast, led to higher Colwell P on the till half. One of the fuel burner plots, 10, and the remaining hungry sheep and reduced till plots, all had similar Colwell P regardless of tillage.
Figure 12. Effect of tillage treatments on Colwell P measured in March 2008.
Straw treatments had less consistent effects with regard to systems. In no till plot 6, in which no crop was sown in 2007, Colwell P was unchanged for the straw treatment (Figure 13). Otherwise straw plots tended to have lower Colwell P, apart from 27, which had higher Colwell P.
Figure 13. Effect of straw treatments on Colwell P measured in March 2008.
Plant uptake of phosphorus and relationship with nitrogen
Cereal crops in the systems trial (except standard wheat) were sampled in 2007 at GS30 for whole-plant phosphorus and nitrogen concentration. The season had begun well in 2007 and it was hoped that system-related nutrition differences might be apparent at GS30. In 2008 the season was relatively poor, but differences between straw and till treatments and no-till controls were of interest and these plots were sampled. In both years, crops were relatively phosphorus-limited (and in luxury nitrogen supply) at GS30 (Figure 14a, b), compared with the N:P relationship at maximum yield estimated by Sadras (2006). Some patterns were evident between NT/RT treatments in 2007; +straw plots tended to have less nitrogen, whereas +till plots tended to have nitrogen at the upper extreme. In 2008 it was more difficult to discern treatment-related patterns. Even in 2007 the relative amounts of P uptake at GS30 were small; 3 kg P/ha and less than 1.5 kg/ha in 2008.
Figure 14. Cereal crop nitrogen and phosphorus uptake at GS30 in 2007 (a) and 2008 (b). The curve is the N:P uptake relationship of high yielding crops in field experiments, at maturity in Sadras (2006).
Relationship to Colwell P and P balance
Phosphorus fertiliser management is undertaken with the intention that techniques such as Colwell P/PBI soil tests, or phosphorus budgets, will identify paddocks that are likely to suffer phosphorus deficiency in the coming crop. In these crops, plants tended to accumulate an excess of nitrogen, with respect to phosphorus. Phosphorus is the limiting nutrient – which should in theory have some relationship to indices used in phosphorus management. Here we have characterised the degree of phosphorus limitation by expressing actual phosphorus uptake as a percentage of the theoretical uptake that should accompany nitrogen taken up by the plant.
There is the semblance of a relationship between plant phosphorus status (with respect to nitrogen uptake and the Sadras N:P for maximum yield relationship), and Colwell P with respect to the PBI-determined Colwell P required for 90% maximum yield (Figure 15a, b). Unfortunately there was no Colwell P to pair with 2007 data, and the March 2008 data in (a) would only approximately reflect March 2007 measurements. The outliers to the expected increasing relationship in 2007 were a HS no-till plot (32) with very high P status for Colwell P, and the till side of NT plot 11, with very low P status for Colwell P. Plot 32 was grazed for two weeks before sowing (after soil sampling), sown late and the till half harrowed post-sowing. Plot 11 till was cultivated once pre-sowing in 2006; the crop on the no till half had a similar P status (58.4%) but Colwell P just higher than required for maximum yield. It is quite surprising that a single event could have a residual effect when measured more than 2 years later.
In 2008 crops in RT and till plots had higher phosphorus status if they had higher Colwell P, but the same trends weren’t as obvious for the NT and straw plots (Figure 15b). In particular the NT no till crop in plot 16, which had been re-sown, had low P status for the Colwell P, whereas the straw half had similar P status but Colwell P near critical. Plots with high P status for Colwell P were 22 no till, and the straw half of 24.
Figure 15. Relationship for (a) 2007 and (b) 2008 between phosphorus status of cereal crops at GS30 (with respect to N uptake, and Sadras N:P for maximum yield), and difference between Colwell P measured in March 2008 and critical P for 90% maximum yield.
There was some overall relationship between P balance and crop P status in 2007, but little evidence for useful relationships within particular systems (Figure 16a). In 2008, there was quite a good relationship between P balance and crop P status for reduced till plots, and all till treatment plots (Figure 16b). The main outliers to the relationship were from NT plots 16 and 6, for both straw and no till halves. These plots have had consistently high P balance since 2004 (Figure 3c).
Figure 16. Relationship for (a) 2007 and (b) 2008 phosphorus status of cereal crops at GS30 (with respect to N uptake, and Sadras N:P for maximum yield), and plot phosphorus balance.
Discussion
Phosphorus balance of different farming systems
The results from the BCG Farming Systems Trial show that different systems, and paddocks within systems, have quite different phosphorus balance. The standard rotation (Figure 4) shows that these are more likely to result from inadvertent association of particular crops (both inputs and yields) with particular years, rather than spatial variation. Apart from the standard rotation, many of the differences between systems arose in the early years of the experiment (1999-2003), which is when the champions were adapting to the environment. There was also more rainfall, which led both to more nutrient removal with successful systems, and encouragement to apply more nutrients in following years. In later years, the systems have followed similar trajectories (Figure 5a). The exception is the standard rotation, which continues to increase at a faster rate and can not easily be adjusted to suit the drier trend in season. All systems are, on average, currently applying about 5 kg/ha more P than is removed.
An excess of nutrient application should normally lead to increasing soil test levels, yet this has not been observed in the trial. All Colwell P levels increased between the setup year of the trial and early 2004 (Figure 8), but have since decreased and/or stabilised. This is the reverse of what might be expected. Only with the most recent Colwell P measurements (March 2008) is there some sort of relationship between phosphorus balance and Colwell P.
A likely explanation for the poor relationship comes from looking at the relationship between change in Colwell P and phosphorus balance (Figure 10b). The relationship crosses zero change in Colwell P at a positive balance, which implies that a fixation process is continually removing P from the pool measured by the Colwell test. Possibly some of the processes that might return P from the fixed state are rainfall (or plant root) driven, hence the early increases in Colwell P in wetter years, and recent stability despite high net P input in drier years {need to check literature here}.
A caveat to the phosphorus balance results presented here is that the P concentration of grain from the site has not been verified. In literature grain P for wheat has varied in the range between 0.17 and 0.45% (Kim et al. 2002). Consistent errors from the assumed value (0.33%, Table 2), particularly between systems, could mean that the net positive phosphorus balance across 9 years of crops has been greatly over- or under-estimated.
The implication from these results for practice is that the current positive P balance of some paddocks does not necessarily mean they are building a bank of available P (not one that we can measure, anyway). It is possible, however, that some fixed P may become available in a wetter year. The main concern with low or no phosphorus application (desirable to manage cash flow risk after dry seasons) is missing yield in wetter seasons because of deficiency. Phosphorus deficiency can also reduce the rate of crop response to nitrogen (Holford et al. 1992). Some economic analyses {need a ref here} suggest that taking advantage of good years can be a key to increasing net wealth {although probably conditional on particular environments!}; others that minimising losses in poor years is key!
Soil tests and phosphorus budgets for phosphorus management
A firm relationship between the Colwell P soil test or phosphorus budget and the degree of crop phosphorus requirement (deficiency) would give confidence in the use of either for phosphorus management of crops. The ultimate indicator for phosphorus deficiency (apart from a P response experiment in the crop) is plant nutrient status. Where measured in these crops, plants were always under-supplied with phosphorus, relative to nitrogen. As suggested in the analysis of focus paddock data, this probably relates to dry seasons and the concentration of phosphorus in topsoil, although 2007 was relatively wet up to GS30, when measurements were made.
The degree of phosphorus deficiency was related to both Colwell P and the phosphorus balance (Figure 15a, b, Figure 16a, b). When crops had Colwell P equal to the ‘critical’ Colwell P for 90% maximum wheat yield, according to Moody (2007), they had about 70% of the P necessary to match nitrogen uptake, according to the N:P at maximum yield relationship of Sadras (2006). Considering that 2008 Colwell P measurements were used with 2007 crops, and that there was no consideration of fertiliser in the year of the crop, the relationship with Colwell P (with PBI) was reasonably useful in predicting degree of P deficiency. Similar crops had phosphorus balance (at July – same time as the plant harvests were taken) of about +50kg (1999-2007) or +60kg (1999-2008). Although the phosphorus balance was useful in predicting P deficiency, careful thought seems necessary in the timeframe used (and possibly soil type) if different paddocks with different soil types are being compared. It is likely that the ‘phosphorus fixation’ implied by these positive P balances would differ between soils and cropping histories.
System specific considerations
There were system-specific differences in measurements of phosphorus status. Tillage in the fuel burner and no-till systems had quite different effects on Colwell P measured in March 2007 (Figure 12), even when the tillage had taken place nearly a year ago (no-till plots). There was less effect in the reduced till and hungry sheep systems. These different responses to tillage probably reflect differences in nutrient stratification, which would be greatest in no-till and least in the fuel burner systems, probably intermediate in reduced till and hungry sheep. It would be good to have a measure of stratification differences between systems at this point, particularly because the soil cracks when dry and it is likely that considerable surface soil P has shifted to depth in fuel burner plots. Similarly in 2008, the reduced till plots had quite consistent relationships between Colwell P and crop P status (Figure 15b), whereas no till plots did not. This may have had something to do with the different P balances between systems (Figure 16b). Straw treatments (5 t/ha in Jan) also led to reductions in Colwell P in some plots, as might be expected from immobilisation (Figure 13), but this did not happen in all plots.
It could be that there are season x system interactions in the relationship between P balance and crop P status. It may be necessary to know these to use techniques such as P replacement budgets effectively. Knowing the P balance of different paddocks may help to understand differences in soil tests and crop response to P between paddocks (or simply where to expect some).
Conclusions
Phosphorus measurements from the BCG Farming Systems Trial demonstrate some of the challenges for farmers in managing phosphorus. Even on the same soil type, with the same rainfall, soil and plant phosphorus measurements are not necessarily easily interpreted in terms of phosphorus fertiliser and export history. A useful initial recommendation is to include a fixation component in phosphorus replacement budgets, which may need to be bigger in drier years. Further analysis is required to show whether this ‘fixed’ fertiliser is returned in later years. It may be necessary to consider deeper soil P for systems that include tillage on cracking soil types. The limited set of measurements on the trial suggest that Colwell P (with reference to PBI) is a useful indicator of soil P supply, but this may not apply in all systems. P replacement budgets could also be useful but might require careful interpretation, particularly with soil and cropping history differences.
Acknowledgements
Thank you to the Grains Research and Development Corporation for funding the project (BWD00008), to the farming systems champions for their consistent input, to Harm van Rees for being the early ‘champion’ of the trial, and to the many BCG staff involved in data collection, processing and collation.
References
Bellof G, Pallauf J (2007) Deposition of major elements in the body of growing lambs of the German Merino Landsheep breed. Small Ruminant Research 73, 186-193.
Bolland MDA, Siddique KHM, Loss SP, Baker MJ (1999) Comparing responses of grain legumes, wheat and canola to applications of superphosphate. Nutrient Cycling in Agroecosystems 53, 157-175.
Holford ICR, Doyle AD, Leckie CC (1992) Nitrogen response characteristics of wheat protein in relation to yield responses and their interactions with phosphorus. Australian Journal of Agricultural Research 43, 969-986.
Kim JC, Mullan BP, Selle PH, Pluske JR (2002) Levels of total phosphorus, phytate-phosphorus, and phytase activity in three varieties of Western Australian wheats in response to growing region, growing season, and storage. Australian Journal of Agricultural Research 53, 1361-1366.
Moody PW (2007) Interpretation of a single-point P buffering index for adjusting critical levels of the Colwell soil P test. Australian Journal of Soil Research 45, 55-62.
Sadras VO (2006) The N:P stoichiometry of cereal, grain legume and oilseed crops. Field Crops Research 95, 13-29.