Nitrogen management in the southern region practices facts and gaps

Background

While nitrogen (N) fertiliser is consistently the largest and most costly input for grain crops nationally it represents one of the highest returns on investment when well managed. Accurate predictions of N fertiliser requirements are therefore very important to growers and advisers.

There is a widespread concern that these N decisions sometimes miss the mark despite the vast range of N management decision tools available to growers and advisers. Potential contributing factors to this uncertainty include changing rainfall patterns, declining soil organic matter, a lower legume frequency in farming systems and increasing adoption of zero tillage/stubble retention systems and changes in how and when N fertiliser is applied. 

A national project (UQ00079), within the More Profit for Crop Nutrition II (MCPNII) initiative, aims to assist grain growers and their advisers to reduce the uncertainty and financial risk associated with N management. The initial part of this research involves understanding the assumptions and perceptions of growers and advisers that influence N decision making and comparing that to the recent and past science. The aim is to identify gaps in knowledge and design future research and extension to cover those gaps.

This paper reports on the assumptions and perceptions of southern region advisers when assisting clients to manage nitrogen inputs. It provides learnings for both the advisor and research community that if adopted could markedly improve N decision making, particularly in difficult seasons.

Methodology

This survey assessed current knowledge and practices used by advisers when making decisions concerning N management. A written questionnaire was undertaken with 32 advisers in the southern region across a range of soil types and rainfall zones. Most respondents (48%) were from medium rainfall areas, but there was sufficient exposure to low rainfall and high rainfall situations to gain an understanding of advisor perceptions in the various agro ecological zones within the GRDC southern region. Focus interviews of 25 per cent of the questionnaire respondents were subsequently undertaken. The survey data set contained highly knowledgeable and well informed advisers. Key areas of interest identified by the investigating officer were the response to N losses, the use of soil tests and how N mineralisation was accounted for. The level of understanding of mechanisms and assumptions behind rules of thumb was also discussed at length.

Results and discussion

Use of decision-support tools 

A range of tools were being used and it often depended on the individuals learning and operation style. The general consensus was a combination of tools gives the best outcome with 75 per cent using a combination of tools.

There was a high emphasis on pre-sowing tools, particularly soil testing (deep N) however, other questions identified that the uptake of deep N testing was much lower despite the acknowledgement of its usefulness. 

‘Experience’ and ‘gut feel’ had a major influence on N fertiliser recommendations and the varying approaches used by advisers. 

Less than half (44%) indicated use of computer aided tools such as Yield Prophet or personalised Excel models that allow greater flexibility to account for the multitude of variables (but are often built on significant assumptions and simplifications). Confidence in Yield Prophet was good for the majority surveyed but there were still some major concerns that needed to be addressed (e.g. correctly identifying the soil type and the high level technical understanding of soil science needed to achieve acceptable precision).

There was a clear desire for tools and guidelines to be simple with a focus on problem solving. Half (50%) of respondents used ‘rules of thumb’ that dated back to the early 1990s and there was the feeling that many advisers were not aware of the source of the rules of thumb or the limitations around using them. 

There was support for simple and practical extension guidelines to remind advisers of the principles behind existing ‘rules of thumb’ and to update these for aspects of N management to account for significant changes that have since occurred to farming system and climatic patterns- specifically:

  • N Mineralisation: estimates of quantity based on soil type, temperature, rainfall pattern, season length and organic carbon
  • Clarification of what situations (soil types, seasonal conditions) and scale of different forms of N loss may occur, particularly denitrification and volatilisation.
  • Estimates of crop uptake efficiencies based on time, method of N application and seasonal conditions.

The respondents thought that much of the information needed for updated ‘N management guidelines’ exists from current and previous research (e.g. recent NANORP work; Armstrong/Harris/Wallace) but needs packaging up in a consumer friendly manner.

There was a perception that ‘a lot of the research has been done’ but needs to be adapted and validated to current farming systems and management practices. This information must be then packaged into a form that is of use to advisers (e.g. as ‘rules of thumb’) with accompanying background information the basis for the rules. The question which remains however is: how well do the advisers perceptions of knowledge generated by researchers (past and present) match that of what has been published (in both the peer-reviewed and grey literature)? For example, many advisers currently assume a value of 50 per cent recovery of fertiliser N by crops. However recent evidence (AWallace et al., pers. comm. 2015) suggests that crop utilisation of fertiliser N can be less than 10 per cent.

The focus group was requesting effective science communication and extension. Advisers emphasised that the science is important but is only part of the information an advisor weighs up. The challenge for research bodies is to communicate the key science findings in a manner that supports the other decision making tools that advisers use, which are primarily adviser experience and intuition. Guidelines must recognise that on farm logistics and scale require estimations not precision, hence the need for simple ‘rules of thumb’. However these ‘rules of thumb’ must also indicate their relative accuracy and strongly this leverages the overall result.

Table 1: Percentage of respondents utilising various tools.
Yield Prophet® or another computer based tool
44%
'Rules of thumb' 50%
Paddock history (including previous grain protein levels) 50%
N balance 3%
Soil testing 81%
Seasonal outlook 44%
In-season crop assessment (e.g. tiller number) 22%
In crop sensing (e.g. Green Seeker®) 13%
Combination 75%

Understanding of the mechanisms assumed in the approach/tools used to determine N fertiliser requirements, particularly mineralisation, N loss and uptake efficiency

The sample population used for the questionnaire/survey and subsequent focus group interviewees predominantly contained highly knowledgeable and well informed advisers, some with post graduate qualifications in crop nutrition. As such, we consider that their level of scientific understanding to be far greater than the ‘industry average’. However, only 12 per cent indicated a very high level of understanding of N mechanisms while 44 per cent responded ‘well understood’, and 44 per cent indicated neutral or not well understood.

In-crop N mineralisation

The main area that advisers considered not well understood was in-crop mineralisation, particularly in relation to variable seasonal conditions. Two thirds of respondents accounted for possible contributions from N mineralisation but many considered it a difficult aspect of determining N supply and making N fertiliser recommendations. Although outdated, the equation for estimating soil N mineralisation of ‘organic carbon percentage x 0.15 x growing season rainfall (mm)’ was still the dominant ‘rule of thumb’ for estimating mineralisation. The advisers interviewed had modified that ‘rule of thumb’ based on experience and on the limited amount of information that had been released from time to time.

Both experienced and less experienced advisers strongly supported the production of updated and simple ‘rules of thumb’ to account for in-crop N mineralisation based on the accumulated understanding from research and commercial experience over the past decade. Recent data (Dunsford et al., 2015) suggest that whereas these ‘rules of thumb’ can sometimes be reliable, in other circumstances they perform poorly (Table 2).

Table 2: Correlations (R2) between in-crop N mineralisation (ICM) and the two calculators.
 Data source Ridge (LTA) Ridge (Season)
All treatments 0.10 0.07
Wheat data only 0.27 0.23
Wheat data excluding pasture legume rotation 0.50 0.56
SCRIME + MC14 (excluding pasture legume rotations) 0.63 0.63
SCRIME + NANORP wheat (excl. pasture legume rotations) 0.68 0.74

LTA = long term average rainfall; Season = GSR (mm). (Source: Dunsford et al. 2015)

N losses and crop uptake efficiency

When asked ‘Do your N fertiliser recommendations take into account N losses such as leaching, ammonia volatilisation or other gaseous losses?’, 38 per cent responded ‘Yes’, 41 per cent ‘Sometimes’ and 28 per cent ‘No’. The telephone interviews suggested that advisers generally assume that the 50 per cent N efficiency ‘rule of thumb’ allows for N losses as well as crop uptake efficiency and that advisers don’t seem to separate the two when calculating N requirements.

In response to the question ‘What do you regard as the major source of N losses in your region?, 25 per cent indicated ‘Leaching’ , 50 per cent ‘Ammonia volatilisation’, 38 per cent ‘Denitrification’ and 16 per cent ‘Nitrous oxide emissions’. The responses depended on the soil type and recent rainfall/seasonal conditions experienced by the advisers. Those who noted leaching as the main source of N loss is their region tended to service clients who farmed a range of soil types including siliceous white sand that was prone to leaching beyond the roots zone after rainfall. The advisers that listed volatilisation as their major source of N loss were working on medium textured alkaline soils with clay subsoils, hence leaching being less of a concern. When asked why volatilisation was deemed a higher risk than denitrification, these advisers noted that it has been too dry for denitrification to occur except in 2010 and perhaps 2011. The importance of denitrification was a common response by advisers working in medium to high rainfall zones with duplex soils. Nitrous oxide was listed by a small number of participants as a major source of N loss in higher rainfall zones. In lower and medium rainfall zones, the likelihood of NO2 losses was considered low.

Most advisers in the phone interviews were unsure of ‘What proportion of soil/fertiliser N do you think is lost in an average season in your region?’ but their answer was based on their recollection of publicised data. In the questionnaire, 22 per cent of respondents indicated that 0-10 per cent of soil/fertiliser N was lost in an ‘average season’, 50 per cent indicated ‘10-20 per cent ’ and 22 per cent indicated ‘between 20-40 per cent of N’. Others felt the amount of N lost varied too much from season to season to put a number on it.

Recent research would suggest otherwise (Figure 1).

A bar chart showing amount of soil/fertiliser N recovered.

Figure 1: Recovery of 15N labelled fertiliser (50 kg N/ha) by wheat at Taylors Lake (2013). Vertical lines represent l.s.d. (5%). 50N - urea incorporated at sowing; 50EN - urea + entec applied at sowing; 0:50 - urea top dressed at GS30; 50GU - Green urea applied at GS30. (Source: A Wallace pers. comm.).

The large range of responses recorded and a tendency for respondents to combine ‘losses’ and ‘crop uptake efficiency’ into the same category , highlights the need for a refresher on terminology and scenarios where various sources of loss are most likely to occur (e.g. Table 3).

Table 3: Sources of N loss and high risk scenarios (Source: Grace et al., 2015).
Term Simple Definition Risk increased Risk unlikely
Volatilisation Loss of Ammonium (NH4) from soil surface / urea fertiliser as ammonia (NH3) gas. Alkaline soils Topdressing in dry conditions?? Acid soils
Denitrification Conversion of nitrate by microorganisms to the gases nitric oxide (NO), nitrous oxide (N2O) and di-nitrogen (N2) under anaerobic conditions Source of labile C. Increases as water content (or WFPS) increases above field capacity. Heavy textured or poorly structured soils, greater in neutral alkaline soils. Well-structured and coarse textured soils Dry conditions; low soil nitrate; acid reaction.
Leaching The movement of nitrate (NO3) down (and beyond) the root zone Coarse textured soils such as non-wetting sands Medium to heavy textured soils; low rainfall
Nitrous Oxide Gaseous emission of N2O High soil N and C; wet and warm soil; acid soils Low background soil N and C; dry conditions

Most respondents (75%) were aware of the need account for crop uptake efficiency in N fertiliser recommendations. The focus interviews suggest that most advisers used the grain N demand ‘rule of thumb’ of 50 per cent of ‘N’ (both soil and fertiliser) being utilised by the plant. For example if 20kg/N per tonne of wheat is removed, then the N demand is 40kg N/t. Some advisers with more advanced levels of knowledge modified the crop uptake efficiency factor depending on the situation they were advising to. There was concern expressed that there may be a lack of understanding in the industry, particularly with less experienced advisers about the basis for N demand rules of thumb and what they account for.

Factors influencing the ability to predict crop N response

Most respondents (94%) were very comfortable with their ability to predict crop response to N. However, this judgement of success was largely subjective based on client feedback during the planning process for the following season. Very few advisers undertake a validation of actual versus expected grain yield and protein. Very few advisers also calculated N recovery or N use efficiency. The uncertainty around N decision making could be lessened by more formal validation of previous decision making. Advisers that undertake this practice find it a very useful tool.

Fertiliser N recommendations within 10 to 15 per cent of potential yield were considered sufficiently accurate by 59 per cent of respondents. Rainfall zone and yield potential influenced the response. Only 49 per cent of respondents servicing lower rainfall zones (<375 mm annual rainfall) desired accuracy within 15 per cent of yield potential. In medium to high rainfall (>375mm or irrigation), 90 per cent of respondents expected accuracy within 15 per cent of yield potential. In the <375mm rainfall regions, final N management decisions are made in June and July when the seasonal indicators are far less reliable so it is unrealistic to expect too greater accuracy.

Potential yield accuracy at the time of N application within 10 to 15 per cent of final yield was considered adequate by 56 per cent of respondents and a further 28 per cent of respondents felt yield estimations within 15 to 25 per cent was sufficient. The number of contributing variables to yield potential (other soil fertility issues, weed control, root and foliar disease, variety potential, timing of rainfall, timing of heat or frost stress) were considered factors that make more accurate predictions difficult to achieve.

When asked ‘What do you believe are the main factor(s) that affect fertiliser N requirements (soil water at sowing; seasonal conditions; rotation; soil test; other limitation e.g. soil constraint; financial risk)?’. All factors listed were considered to affect N fertiliser requirements to varying degrees (Table 4).

Table 4: Assessment of factors affecting fertiliser N requirements (per cent of respondents that consider the factor important).
Factor Soil water at sowing Seasonal conditions Rotation Soil test Other limitations (e.g. soil constraints, disease) Financial risk
% of respondents 63 88 69 47 50 38

Seasonal conditions 

Seasonal conditions were considered a significant factor by 88 per cent of respondents and soil water by 63 per cent. Despite the perceived importance of seasonal conditions, few advisers are using all the available resources to factor in the possibilities of certain seasonal outcomes based on the information available (e.g. climate models). There is still a tendency to use long-term average rainfall for predicting the likely amount for the remaining of the season (or making only a small deviation from this). 

Advisers may not be fully utilising all resources as human factors associated with living and working among communities under stress due to adverse seasonal conditions can influence the ability to objectively analyse climate scenarios and often the grower audience is not overly receptive. 

There is great potential to improve practices for revising potential yield based on soil water and a range of probable seasonal outcomes using tools such as Yield Prophet® or more simple approaches such as that proposed by Burke (2016) (Figure 2). Tools such as CliMate, and DEDJTR’s The Break, Fast Break and Very Fast Break could be more widely utilised as should the BOM’s summaries.

A flow chart showing a process for improving potential yield forecasting.

Figure 2: Suggested process for improving potential yield forecasting and N decision making. (Source: K Burke)

Deep N soil testing

There were mixed messages delivered by advisers on the value of soil testing.

Eighty one per cent of respondents listed soil testing as a tool they used to guide decision making. However, soil test was cited by only 47 per cent of respondents as an important factor affecting fertiliser N requirements. Seasonal conditions, crop rotation and soil water were all ranked higher. When asked ‘How important soil testing was for determining soil fertiliser requirements’, it was considered very important or important by 82 per cent. Critically, three per cent considered it not important and 25 per cent considered it important but did not test.

Soil sampling to a depth of 60cm (commonly termed ‘deep soil testing’) was considered labour intensive, time consuming and logistically difficult. Sampling and the associated costs were a barrier to effective utilisation of deep N soil tests. Most advisers encouraged their clients to test a representative sample of paddocks each season and extrapolate results to the remaining fields. Using previous experience to estimate likely soil N scenarios for various situations was the most common method employed in lieu of soil testing.

Those who habitually used soil testing as a major part of their decision making process were strongly supportive of industry adopting more soil testing and felt that the enormous financial benefits could be realised across the southern region in terms of improved fertiliser decision-making leading to better risk management in dry seasons and tapping into unrealised yield potential in favourable seasons.

Conclusions

Learnings raised from the survey and opportunities to improve N decision making include:

  • The opportunity to increase industry knowledge of how to interpret seasonal outlook information to improve estimation of potential yield and N demand.
  • The clear discrepancy between desire and practice regarding soil sampling. Soil tests were considered ‘very important’ or ‘moderately important’ by the vast majority of consultants. However anecdotal evidence (and other limited survey work) indicates that few paddocks are soil sampled prior to sowing (especially to depth). This highlights the need for either a more cost effective and/or logistically and timely procedure to estimate soil mineral N reserves prior to sowing as this appears to be the main barrier to adoption.
  • Advisers’ perceptions of key soil N processes such as mineralisation and loss will be influenced by several biases including recent experiences, particularly in adverse seasonal conditions. There is a strong demand to clearly communicate the science and mechanisms involved in soil N dynamics and soil N supply in the context of modern farming systems.
  • The Australian grains’ R&D community is relatively small and dispersed by international standards. As a result, research undertaken in one particular environment and farming system is extrapolated to many other scenarios. Whether this is always appropriate is often not considered, let alone formally assessed. This study showed that advisers are often ‘aware of previous research’ (‘exposure’) but there is often little understanding of the relative strengths/weaknesses of this work and in what context it is most applicable (which is necessary for practice change). Researchers and science communicators must recognise this when providing information to advisers that is then used to assist growers in their decision making. Especially when this advice has considerable financial risk attached to it. 

Acknowledgments

We are extremely grateful to the many consultants who answered the written questionnaire and participated in the follow up telephone interviews.

Funding for this work was jointly provided through GRDC Project UQ00079 and Agriculture Victoria (DEDJTR) and their support is gratefully acknowledged.

Useful resources

Dunsford  K, Armstrong R, Tang C, Sale P (2015). Estimating in-crop nitrogen mineralisation in dryland cropping systems of southern Australia. Proceedings 17th Australian Society of Agronomy Conference. (Hobart 19-24 September).

Grace P, Armstrong R, Harris R,Wallace A,Schwenke G, Li G (2015) ‘Where does the nitrogen finish up?’ GRDC Grains Research Adviser Update. Adelaide, January 2015.

Contact details

Roger Armstrong
Agriculture Victoria, Department of Economic Development, Jobs, Transport and Resources
Grains Innovation Park. PB 260, Horsham Victoria 3401
03 5362 2336
roger.armstrong@ecodev.vic.gov.au

GRDC Project Code: UQ00079,