Download as PDF

Return to NMI Home page

Long term experiments – nutrient balances and lessons in the Wimmera & Mallee

Presented at Long Term Fertilizer Workshop June 4-5, 2007, Adelaide. Organised by Nutrient Management Systems and Peter Wylie.

 Rob Norton¹ , Roger Armstrong², Roy Latta³, Lauren Dart^1,, Vu Dang⁴, Caixian Tang⁴, Charlie Walker⁵ and Rob Christie⁵.

1   School of Agriculture and Food Systems, The University of Melbourne, Private Bag 260, Horsham, Victoria. 3401.

2   Victorian Department of Primary Industries, Grains Innovation Park, Private Bag 260, Horsham, Victoria, 3401.

3   Victorian Department of Primary Industries, Mallee Research Station, Meridian St, Walpeup 3507.

4  Department of Agricultural Sciences, La Trobe Univerisity, Bundoora, Victoria 3086

5  Incitec Pivot Limited, PO Box 260, North Geelong, Victoria 3212.

Abstract

This paper summarises the apparent nutrient balance from three long-term cropping systems experiments in the Wimmera and Mallee regions of north-western Victoria.  The longest running cropping systems experiment at Horsham on a vertosol (Longerenong Rotation 1 – LR1) and those rotations where a grazed oats treatment is included show a small P deficit (< 0.5 kg P/ha/y), while all the other rotations have a positive P balance and high levels of total P in the soil.  However, there are differences among the P fractions indicating low P availability to plants.  The two other rotations studies, one at Horsham on a vertosol (Sustainable Cropping Rotations for Mediterranean Environments - SCRIME) and the other on a Calcarosol at Walpeup (MC14), show positive P balances across all crop rotations and tillage practices.

The inclusion of pasture legumes or pulses into rotations results in a positive apparent N balance in all the rotations at all three experiments.  In general, rotations with net negative N balances show lower total soil N levels than those with legumes, and C contents tend to follow the N levels.  The fractionation of soil C is incomplete but there is a tendency for lower particulate organic carbon levels where cultivation and/or fallowing was included in the system.

Key Words

Nitrogen, phosphorus, crop rotations, nutrient balance

Introduction

The use of long-term agronomic experiments in Victoria was reviewed by Centre for Land Protection Research (2003).  Those authors identified that these experiments were able to be used as benchmarks for agroecological systems where the effect of various management treatments can be assessed over time on soil conditions such as organic matter levels, soil structural stability, changing soil nutrient levels, root disease build up and deep drainage.  Such changes occur relatively slowly, so that the identification of trends in these features requires decades rather than years or months to assess.  The data generated can be used to validate cropping systems models over time, which can then be used to assess multi-factor effects through simulation modeling.  As well, these sites also act as a platform for other research and in effect become the field laboratory for agronomists.

Within western Victoria, there are five long term experimental sites and as part of the GRDC project UM00023, we are assessing the impact of different management strategies on long term soil nutrient levels.  The National Land and Water Resources Audit (2001) proposed that the Victorian grains industry showed a net negative N greater than 50 kg N/ha/y and a net negative P balance of more than 10 kg P/ha/y.  The objective of this short paper is to present a summary of three of these long-term experiments and the effect of rotation on apparent nutrient balances.  We have not yet completed analyses of the Incitec Pivot Long Term Fertilizer site, but the effect of long term P applications on a Calcarosol (MM1) in the Mallee have been reported elsewhere (Vu et al. 2006).

Experiments reviewed

Longerenong Rotation 1 (LR1)

This is Australia’s longest running annual cropping system experiment, established in 1916 and so has its own historical significance.  It is located on a vertosol (cracking clay soil) at Longerenong (near Horsham) with an average rainfall of 420 mm.  The experiment compares seven cropping rotations and although not spatially replicated, it is phase replicated and the trends under the management regimes were analysed by Hannah and O’Leary (1995).  The rotations are continuous wheat (WWW), fallow/wheat (FW), fallow/wheat/oats grazed (FWOg), wheat/barley/peas (WBP), wheat/oats/peas (WOP), wheat/oats grazed/fallow (WOF) and wheat/oats/fallow/oats grazed (WOFOg).  The crops receive no fertilizer nitrogen, 10 kg P/ha on wheat and 5 kg P/ha on other harvested crops.  Crop establishment, weed control and crop protection activities follow district practice.

Grain yield has been recorded each year and grain protein (N) in recent years, although seed P content has not been measured.  To develop a nutrient balance for this experiment, the apparent balance of N or P was calculated on an annual basis as:

                        N_bal = N_fer + N_fix – N_grn

or                     P_bal = P_fer – P_grn

No estimates were made for free living N fixation, non-biological N inputs, N leaching, N volatilization or N lost in soil erosion.  The N₂ fixation for the pea phases were estimated using the peak biomass for peas from the pea grain yield assuming a harvest index of 0.3, then converting this peak biomass to N fixed by using the conversion of 25 kg N per t of biomass (Peoples et al. 2001).  Grain N removal was estimated by the grain N content multiplied by the yield of peas, barley or wheat.  Both the grazed oats and the crop stubbles were retained within the plots.  Grain P content was estimated from grain P contents taken in 2005, but the actual grain P contents may be quite different due to different soil P levels.

Table 1 shows the average N balance for this rotation for the period 1980 until 2005.  These data indicate an average decline in soil N of 12 kg N/ha/y where no pulse was included and a slightly positive N balance where peas were grown. 

Table 1.  Apparent mass balance of N and P for rotations of LR1 (1980-2004).

Rotation & Mean wheat yield(t/ha)		P Balance kg/ha/y			N Balance kg/ha/y
		P input  fertilizer	Estimated P removal	Net balance	N input fertilizer	N fix	Grain N removal	Net balance
WWW	0.75	10.0	2.7	+7.3	0	0	7.6	-7.6
WF	1.57	5.0	4.1	+0.9	0	0	12.3	-12.3
WOgF	2.29	3.3	3.7	-0.3	0	0	11.6	-11.6
WBP	1.75	8.4	5.1	+3.2	0	37.9	33.0	+4.8
WOP	1.65	6.7	5.5	+1.2	0	35.6	32.8	+2.8
WOF	2.05	7.5	4.6	+3.0	0	0	14.8	-14.8
WOOgF	2.34	3.8	3.9	-0.1	0	0	13.1	-13.1

There was no baseline soil archived when the experiment was established 90 years ago, and so a “fenceline” sample was taken in an uncultivated area adjacent to the site.  The soil N and C values (top 10 cm) largely reflect the average estimated N decline.  While it is not possible to fully analyse these data due to the nature of the experimental design, the C:N ratios for the rotations suggest that there have been large changes in the nature of the organic matter under the WF rotation compared to the other rotations.

Table 2.  Soil N, C, Colwell P, total P and selected P fractions as a percentage of total P for rotations of LR1 2005.

	WWW	WF	WOgF	WBP	WOP	WOF	WOOgF	Fence Line
Total Soil N %	0.070	0.056	0.063	0.085	0.087	0.061	0.066	0.162
Total Soil C %	0.93	0.91	0.93	1.10	1.11	0.85	0.92	2.12
C: N Ratio	13.3	16.2	14.9	13.0	12.9	13.9	13.8	13.1
Total P mg/kg	486	367	307	341	329	330	322	295
Colwell P mg/kg	69	52	40	40	47	66	50	18
% HCl P	39	25	18	25	22	23	19	7
%P Bicarb. extract (i + o)	8.2	7.8	8.8	8.2	9.2	10.2	9.0	5.6
% Residual P	35	43	47	49	52	50	61	75

Table 1 also shows the P balance for the various rotations at LR1.  Dart (2005) and Tang et al. (2006) undertook P fractionation of the soils from this experiment using a modified method of Hedley et al. (1982) and a summary of some of these results is given in Table 2.  All rotations are in a positive P balance except for the two grazed oat rotations.  The total amount of P and the acid-soluble P fraction increased in all rotations, especially in the continuous wheat which also had the highest P balance.  These results suggest that long-term cropping practices used in this experiment increase total P content of the soil, with the amount of P in the less available pools increasing whereas the relative proportion of P in the “plant available” pools decreases.

Sustainable Cropping Rotations in Mediterranean Environments (SCRIME)

This experiment was establish in 1998 and is located near the LR1 experiment and on the same soil type.  It includes newer and more diverse cropping systems including lucerne and a range of pulse crops and other than for one rotation, is managed using reduce tillage practices. There are 10 phase and spatially replicated rotations and the data included here are taken from six rotations, viz. continuous wheat (WWW), wheat/barley/peas (WBP), wheat/lentil/peas (WLP), wheat/peas/canola (WPC), wheat/pea/canola under conventional tillage (WPC CT) and fallow/wheat/chickpea (WCpF). 

This experiment is managed in essentially the same way as LR1, except that 35 kg N/ha is applied to the wheat and canola crops and 12 kg P/ha to most other crops and the lucerne.  Crop agronomy is based on district practice.  Crop residues are retained.  Estimates of N₂ fixation were made with ¹⁵N natural abundance techniques (Unkovich et al. 1997) in a range of the legume phases, and these values for N derived from the atmosphere were similar to the estimates for other crops reported by McCormick (2004) from the same environment.  Mass balances were estimated using the same techniques as those used for LR1, and the data presented in Table 3 are the mean of all three phases/spatial replicates.

Table 3  Apparent mass balance of N and P for rotations of SCRIME (1998-2005)

Rotation & Mean wheat yield (t/ha)		P Balance kg/ha/y			N Balance - kg/ha/y
		P input  fertilizer	Estimated P removal	Net balance	N input fertilizer	N fix	Grain N removal	Net balance
WWW	1.95	11.2	5.07	+6.16	35.1	0	41.4	-6.3
WBP	2.27	10.3	6.91	+3.39	16.8	45.0	56.1	+5.7
WLP	1.87	9.4	4.21	+5.19	13.1	57.9	39.5	+31.5
WPC CT	1.96	10.3	5.67	+4.63	24.2	36.0	41.4	+18.8
WPC	2.37	10.3	5.95	+4.35	24.2	33.0	42.7	+14.5
WCpF	3.07	6.5	4.14	+2.36	12.4	40.0	34.3	+18.1

The only rotation currently in negative N balance is the WWW, while all the others, which include one or more pulse phases, have apparent positive balances.  There were small differences in N balance between the tillage treatments, a consequence of somewhat lower legume fixation and higher N removal in grain where reduced tillage was used compared to conventional tillage.

All rotations have positive P balances, and as yet there have been no P fractionations completed as yet on soils from this experiment.  Table 4 shows the soil C and N contents for this experiments and the site means in 1998 were 1.13% (±0.10%) for total C and 0.091% (±0.008%) for total soil N.

Table 4.  Soil N and C and their standard errors for six rotations of SCRIME 2005.

	WWW	WBP	WLP	WPC CT	WPC	WCpF
Total Soil N %	0.081±0.002	0.077±0.002	0.083±0.002	0.083±0.002	0.081±0.002	0.074±0.002
Total Soil C %	1.09±0.024	1.06±0.024	1.08±0.024	1.09±0.024	1.07±0.024	1.01±0.024
C:N Ratio	13.5±0.2	13.8 0.2	13.1±0.2	13.1±0.2	13.1±0.2	13.6±0.2

Across the experiment, 8 years of cropping showed significantly lower soil C and N in the WCpF rotation, which has a mechanical fallow phase.  Care should be taken when comparing the 1998 values to the 2005 values for N and C because the samples were not analysed at the same time.  The C:N ratio of the organic fractions provides a coarse measure of the nature of the soil organic matter pools, and these data from SCRIME (Table 4) shows the WBP to have significantly higher C:N than WLP or the WPC rotations.  Preliminary estimates of soil C pools using MIR (Janik et al. 2007) under these rotations indicate that the WCpF rotation with the fallow phase had the lowest level of particulate organic carbon (POC).   However, until calibrations are completed for C fractions, these data should be viewed with caution.

Mallee farming system experiment ( MC14)

This experiment was established in 1985 to investigate the effects of tillage and crop rotation on wheat production in the Victorian Mallee.  It is located on a Calcarosol in a region where average annual rainfall is 335 mm.  The experiment is a split-plot design with three rotations – wheat/pasture/fallow (WPaF), wheat/fallow (WF) and wheat/pasture (WPa) combined with either conventional cultivation (CC) and zero tillage (DD).  The experiment is both phase and spatially replicated.  The pasture phase is a self-regenerating medic pasture and data on crop yields (all years), pasture production (some years) and grain protein (most years) have been collected and reported by Latta and O’Leary (2003).

Using the same methods are described for LR1 and SCRIME, apparent N and P balances have been constructed to assess the impact of management on nutrient balances.   No N fertilizers have been used and the amount of pasture production was estimated using the technique described by Robertson (2006) and that biomass estimate derived was then used to estimate N₂ fixation.  Where no actual grain protein content was measured, a value of 12% was used.  Each rotation has 12 kg P applied in the crop phase only.  The pastures are not grazed or cut for hay.

Table 5 gives a summary of the apparent N and P balance for this rotation since it was established.  Similar to the other rotations, MC14 shows an apparent P surplus of around 3 kg P/ha per year across all rotations and cultivation treatments.  There was very little difference between the cultivation treatments for either N or P balance.  Apparent N balance was negative for all rotations except the Wpa which appears to be near neutral.

Table 5  Apparent mass balance of N and P for rotations of  MC14 (1985-2005)

Rotation & Mean wheat yield (t/ha)		P Balance kg/ha/y			N Balance – kg/ha/y
		P input  fertilizer	Estimated P removal	Net balance	N input fertilizer	N fix	Grain N removal	Net balance
WF CC	2.08	6.0	2.7	+3.3	0	0	22	-22
WPaF CC	2.38	4.0	2.1	+1.9	0	17	25	-8
WPa CC	2.18	6.0	2.8	+3.2	0	25	23	+2
WF DD	1.91	6.0	2.5	+3.5	0	0	20	-20
WPaF DD	2.25	4.0	2.0	+2.0	0	17	24	-7
WPa DD	1.91	6.0	2.5	+3.5	0	25	20	+5

Soil samples were taken in 2005 and total soil N and C were assessed, and these results are given in Table 6.  Both soil N and soil C are significantly different among the various rotations.  Direct drilled systems had higher soil N and C levels than cultivated systems, and the inclusion of pastures and exclusion of fallowing from the rotations both tended to give higher soil N and C values.

Table 6.  Soil N and C and their standard errors for six rotations of MC14 2005.

	WF CC	WPaF CC	WPa CC	WF DD	WPaF DD	WPa DD
Total Soil N %	0.055±0.003	0.069±0.002	0.077±0.003	0.060±0.003	0.074±0.002	0.079±0.003
Total Soil C %	0.658±0.041	0.825±0.034	0.945±0.041	0.711±0.041	0.844±0.034	0.962±0.041
C:N Ratio	12.8±0.4	11.9±0.4	11.9±0.3	11.5±0.3	12.0±0.4	12.0±0.4

The C:N ratio did not show any significant differences due to either rotation or tillage treatment.  Similar to the soil from SCRIME, some C fractionation was undertaken using MIR.  Again these results are preliminary until proper calibrations can be developed for this soil type.  The total C was about 75% of the C content of the vertosols from SCRIME, and the Calcarosol at the MC14 showed less charcoal than at the SCRIME site.  The direct drilled treatments had higher apparent POC levels than similar rotations with cultivation.  The nature of these differences will be investigated more thoroughly during 2007, with actual fractionation of the C planned for MC14, LR1 and SCRIME soils.

Phosphorus stratification under different tillage systems in three contrasting Victorian soils

A study was undertaken to investigate the effects of tillage and crop rotations on the vertical stratification in the soil profile and the distribution of P in different soil pools in three different DPI long-term trials representing the major agroecological grain production zones of Victoria.  The results from MC14 and SCRIME are presented here.

Labile (resin-Pi) and less labile Pi (NaOH- and H₂SO₄- extractable P) soil P pools were concentrated in the top soil (0-10 cm) of both soils (Figure 1). The distribution of P in the different soil fractions varied markedly with agro-ecological zone (soil type and rainfall).  A greater proportion of P was tied up in the less labile pools in the Vertosol (39% sand, 24% silt & 37% clay) compared to the lighter textured Calcarosol (91% sand, 1% silt & 8% clay). 

Tillage systems significantly affected the proportion of soil P contained in both the NaOH-Po and H₂SO₄-Po fractions. This is probably due to the changes on organic C with tillage practice, suggesting that soil C can influence the dynamics of organic P in the soil. However, stubble management/ tillage practise did not significantly affect the vertical distribution of inorganic phosphorus in the profile.

Figure 1: Effects of tillage on inorganic (Pi) fractions in Calcarosol (I) and Vertosol (II).  The rotations are WPaF DD, WPa DD, WPaF CC and WPa CC from MC14 and the WPC (reduced) and WPC RT rotations from SCRIME.  Error bars represent LSD_0.05. 

Incitec Pivot Long Term Fertilizer Experiment Dahlen

This experiment, established in 1996, has five N (0, 20, 40, 80 and 100 kg N/ha) and four P (0, 9, 18. 36 kg P/ha) rates applied in factorial combination on a Vertosol just to the west of Horsham.  In addition, the two N rates are applied either as “at-sowing” or “split” although no N is applied on the pulses, sown at 40 treatments in total across 3 replicates.  The experiment is sown to a different crop each year (1996 Barley, 1997 Chickpea, 1998 Canola, 1999 Wheat, 2000 Barley, 2001 Lentil, 2002 Wheat, 2003 Wheat, 2004 Barley, 2005 Lentil, 2006 Canola, 2007 Wheat) although 2002 and 2006 were complete crop failures.

Soil samples were taken in 1996 before sowing the first crop and the whole site has been sampled again this year, although the results are yet to be collated for presentation.  We are in the process of completing the collation of yield and grain protein contents to develop a mass balance for this site for N and P.  As well, P fractionation of some treatments has been commenced.  We are also undertaking a preliminary economic analysis of the results of these long term fertilizer strategies on this site. 

Conclusion

The data presented here indicate that current cropping systems are not reducing the apparent levels of total soil P.  Most of the rotations studied are managed using current commercial fertiliser P application rates and are in positive P balance. Concentrations of soil total P are steadily rising, but most of this P is in pools that are sparingly solubility to crops in the short term. It has been estimated (B Holloway using data from Bertrand et al 2003) that on some highly calcareous soils in South Australia contain up to the equivalent of 312 kg P/ha in these soil P pools. This figure highlights the economic need to maximise the proportion of fertiliser P that is directly utilised by the crop through the use of appropriate application strategies.

Our data suggested that rotations with low frequencies of legumes (pulse or pastures), or with frequent cultivations, tend to have a net negative N balance.  This manifests itself as a run down in soil N and C.  Our work is continuing to investigate the relative size of the C pools within the soil and its role in controlling N supply to crops.

These preliminary data do not support the proposition put by the National Land and Water Resources Audit (2001) that cropping systems are highly exploitative of N and P, with positive apparent nutrient balances for systems that received modest levels of fertilizer and incorporate pasture legumes or pulses into the rotations at least once each three years.  Measurements of soil nutrient balances are subject to a comparative high error (reflecting inherent soil spatial variability, analytical error and changes in soil bulk density), so there has to be a relative large change in the net nutrient balance between treatments to detect a ‘significant’ change. This situation is made more difficult by the two complete crop failures experienced at the SCRIME site (in 2002 and 2006) as well as several poor seasons (eg 2004) that limited primary productivity. The use of soil fractionation schemes that measure changes in relatively labile soil nutrient pools will greatly assist our ability to detect these trends. However the data presented can also be interpreted that soil fertility is relatively resilient to the type of rotations employed in both the SCRIME and MC14 experiments.

For more information on the data presented here see http://www.jcci.unimelb.edu.au/NMIUM00023.htm.

Acknowledgements

This work has been supported by the Grains Research and Development Corporation through the project UM00023.  The authors would like to acknowledge the foresight of those who established these experiments, as well as the staff involved in managing them since their inception.

References

Centre for Land Protection Research (2003).  Review of long term agronomic experiments in Victoria.  Centre for Land Protection Research (Vic.).  CLPR Research Report No 35. http://www.dpi.vic.gov.au/dpi/vro/vrosite.nsf/pages/lwm_ltae  Accessed May 15, 2007.

Dart L (2005)  Impact of long-term crop rotation on phosphorus fractions in a Vertosol soil.  Honours Thesis (BAgSc) Latrobe University.  49 pp.

Hannah MC and O’Leary GJ (1995) Wheat yield response to rainfall in a long-term multi-rotation experiment in the Victorian Wimmera. Australian Journal of Experimental Agriculture 35, 951–960

Hedley MJ, White RE, Nye PH (1982) Plant-induced changes in the rhizosphere of rape (Brassica napus var. Emerald) seedlings. III Changes in L value, soil phosphate fractions and phosphotase activity. New Phytologist 91, 45-56.

Janik LJ, Skjemstad JO, Shepherd KD and Spouncer LR (2007).  The prediction of soil carbon fractions using mid-infrared-partial least square analysis.  Australian Journal of Soil Research 45, 73-81.

Latta, J and O’Leary, G.J. (2003). Long-term comparison of rotation and fallow tillage systems of wheat in Australia. Field Crops Research 83, 173-190.

McCormick KM (2004).  Fenugreek for SE Australian farming systems.  PhD Thesis, The University of Melbourne 133pp.

National Land and Water Resources Audit (2001) Nutrient management in Australian agriculture, Australian Agricultural Assessment 2001, Natural Heritage Trust. http://audit.deh.gov.au/ANRA/agriculture/docs/national/Agriculture_Nutrient.html Accessed May 15, 2007.

Peoples MB, Bowman AM, Gault RR, Herridge DF, McCallum MH, McCormick KM, Norton RM, Rochester IJ, Scammell GJ, Schwenke GD (2001) Factors regulating the contribution of fixed nitrogen by pasture and crop legumes to different farming systems of eastern Australia. Plant and Soil 228, 29-41.

Robertson SM (2006) Predicting pasture and sheep production in the Victorian Mallee with the decision support tool, GrassGro.  Australian Journal of Experimental Agriculture 46, 1005-1014.

Tang C, L Dart, C Rogers, DT Vu, R Armstrong (2006). Phosphorus fractions in a Vertosol after 88-year crop rotations,  The 3rd International Symposium on Phosphorus Dynamics in the Soil-Plant Continuum,May 14-19, 2006, Uberlandia, Brazil.

Unkovich MJ, Pate JS, Sanford P (1997) Nitrogen fixation by annual legumes in Australian Mediterranean agriculture. Australian Journal of Agricultural Research 48, 267-293.

Vu DT, C Tang, RD Armstrong (2006). Changes in phosphorus fractions of a Mediterranean calcareous sandy soil following long-term application of P fertilizers, Proceedings of the 18^th World Congress of Soil Science, July 9-15, 2006,  Philadelphia, PA.  http://crops.confex.com/crops/wc2006/techprogram/P17300.HTM