Previous PageTable Of Contents

Precision Management Of Fertiliser Application To Pasture

Allan G Gillingham

AgResearch, Private Bag 11008, Palmerston North, New Zealand
Telephone +64 6 351 8064; Fax +64 6 351 8032


On most pastoral farms in New Zealand, fertiliser is the single largest expenditure item in the annual budget. Means of improving the efficiency of return from fertiliser use is of constant concern, however on most farms common fertiliser forms and rates are generally applied over the whole property or large blocks. Variability in pasture productivity and associated soil and plant characteristics occurs on all farms. To date this has usually been recognised as being present but ignored as being relatively unimportant, or unable to be differentially managed.

The advent of precision agriculture technology has produced two advances. The first is the ability to precisely identify and map small-scale variability, and the second is the development of variable rate fertiliser application technology. The challenge now is how to use these advances to improve on-farm economics.

The optimum differential application of fertiliser to pasture will require a more sophisticated basis for recommendation than has sufficed to date. Most models use generalised response curves derived over a number of contrasting seasons, and or from a number of sites with contrasting growth potentials. Such response curves are inadequate to fully exploit the potential from differential fertiliser application to contrasting growth zones. New models should be developed which incorporate the growth potential, together with the associated economic analysis, into the fertiliser recommendation for contrasting sites.

From the results to date there appear to be worthwhile opportunities for adoption of PA approaches into pastoral farming.


Pastoral productivity from New Zealand farmland is largely dependant on the application of super-phosphate fertiliser which provides predominantly phosphate (P) and sulphate (S) nutrients. In many cases this is the single largest item in the annual farm budget. Many years of government and industry funded research has produced models such as Outlook (Metherell, 1999) subsequently incorporated into the OVERSEERTM model developed by AgResearch, which can provide recommendations for fertiliser application to pastoral farms ranging from steep, dry hill country, which may carry less than one stock unit per hectare, to intensive lowland, irrigated dairy farms carrying 5 cows per ha. The level and rate of improvement in stocking capacity of a property is widely accepted as depending on soil fertility and pasture improvement, which in turn is driven by the level and regularity of fertiliser application. Because pastoral products such as milk or meat do not earn high prices compared with other enterprises such as horticulture or viticulture, there is a constant striving for improved efficiency in fertiliser use so that a profit margin can be satisfactorily maintained. Consequently the OVERSEERTM model incorporates several features other than soil nutrient status, in deriving fertiliser recommendations, including the stocking rate, topography and soil type relevant to the farm.

In general fertiliser recommendations are derived on the basis of soil type or obviously differing topographic units, such as flat or undulating land compared with steep slopes that have differing pasture growth potentials. This may result in two or three different fertiliser rate or type recommendations to relatively large blocks of land per farm. In hill country the definition of these blocks is primarily influenced by the method of application of fertiliser. As a result most hill blocks have pastures which are fertilised from fixed wing aircraft with a common rate of fertiliser despite the knowledge that the area contains a multitude of differing slopes and aspect sites with a similarly complex pattern of pasture productivity and species composition. To a lesser extent this also happens on flatter dairy land where between-paddock productivity may be recognised as variable, but for the purposes of fertiliser application is assumed to be relatively uniform, and although the opportunity is available for differential application of fertiliser by ground spread equipment, this is done to only a very limited extent.

Precision agriculture (PA) technology and approaches set out to identify and map the spatial variability within land based production systems, understand the drivers causing that variation, and consider the opportunities for differential management. The application of fertiliser is probably the input that most easily lends itself to differential management. The development of recommendations for differential fertiliser management may seem relatively straight-forward considering the considerable background of research information collected to date on fertiliser responses. However there are a number of issues that have to be examined before such information can be utilised in this way.

The components of differential management

There are a number of conditions that are required before it will be worthwhile attempting to differentially manage pasture at scales smaller than conventionally used (Gillingham and Betteridge, 2001). Perhaps the three most important of these are:

  • The variability between or within paddocks must be spatially stable throughout the season, or persist long enough for differential management inputs to be effective.
  • Specific management recommendations for each production zone must be developed either by field response trials or by a modelling approach, incorporating a climate component to express the yield potential, in contrasting seasons. For example the best yielding zone in a dry year, because it remains moist, may give the lowest yield in a wet year because of poor drainage, and so needs less fertiliser inputs.
  • The output management recommendations must be in quantitative, rather than in relative terms, if the information is to be used to provide guidelines for optimum economic decisions.

During the last 10 years, using sophisticated GPS and GIS technology, there has been a large amount of on-farm information gathered relating to detailed soil parameters and crop yield variability. This is especially so in the USA. The challenge is how to use this information for the financial betterment of the land-owner, or for the improvement in associated environmental conditions. The general limitations to PA management are first the availability of management advice that is specific to each differentiated zone, second the scale at which the management inputs can be differentially applied, and thirdly the economics of applying the selected scale of management.

This paper discusses the first two of these issues with some reference to the economics of differential management where this has been examined.

The scale of differential fertiliser management

Dairy farms

Many dairy farms are managed as uniform areas, ignoring the possibility of variation from paddock to paddock in productivity and therefore in nutrient requirements. There is scope for improved efficiency on each unit, by acknowledging that such variability exists and varying annual fertiliser applications accordingly. For example paddock records from the No 1 Dairy farm at Massey University showed a greater than 50% range in pasture production from lowest to highest producing of 40 paddocks. A similar range was evident in a dairy research farm in the Taranaki region of the western North Island. With the current rapid expansion of dairying onto previous sheep and beef farm-land there a strong probability that the new properties will also show at least as much variability in production, even under irrigation. So simply addressing the paddock variability in productivity could lead to improved nutrient use efficiency. This can be done by direct measurement of pasture cover prior to and after grazing and summing the differences to obtain annual pasture utilisation, or alternatively by relating variations in milk yield to the previous paddocks grazed. This scale of management will not require new fertiliser application equipment or technology.

Hill country farms

Hill country is typically comprised of complexes of slopes and aspects, each with differing pasture growth potentials, according to the levels of soil moisture and temperature prevailing during each season. The general classes of land, such as flat from steep or easy slopes, or north from south aspects, are usually recognised and fenced into separate paddocks where practicable, but each still contains a mix of slopes and sub-aspects.

The land management unit on such properties for stock management is the paddock, but for fertiliser application is usually much larger areas or blocks of generally similar, but individually complex, paddocks. This approach has been necessary to date because of the limitations in aerial fertiliser application technology, which basically involves dropping fertiliser through a hole in the bottom of an aircraft at a common rate over large areas. Current developments in the use of GPS in topdressing aircraft will enable it to fly a pre-programmed GIS prescription map and allow the fertiliser rate to be adjusted according to the topographic unit and associated growth requirements of the pasture below. This essentially does the same job as is currently able to be done by variable rate ground spreaders, except of course the speed of travel is about 10 times as fast. As well as more appropriate application of fertiliser according to pasture requirements this development will allow environmentally sensitive areas within hill country farms to be excluded more precisely than is able to be done at present.

The size of the management units which can be differentiated by variable rate topdressing aircraft has yet to be defined by field tests, but will depend on such factors as aircraft speed, hopper reaction times and the granulation characteristics of the fertiliser material. Inevitably the prescription maps for hill farms will still contain some compromises in terms of treating contrasting land units with the same fertiliser rate, but the net effect should be for much improved efficiency of use of fertiliser to such land units, compared with current practice.

The availability of variable-rate fertiliser advice

The current situation

In New Zealand the AgResearch OVERSEERTM model provides recommendations for fertiliser requirements of pasture. These are based on the results of a large number of field trials over many years and in many locations. For P, these trials have contributed to the definition of an average, standard, relative response curve, derived from a range of growth conditions, which in conjunction with Olsen soil P, stocking rate and other site related variables, is used to define fertiliser P requirements. The site growth potential is factored in through stock carrying capacity, but otherwise not included in the recommendation. Annual rainfall is required to assess the potential sulphate S leaching, but for no other purpose.

Is such a standard curve adequate for the optimum definition of fertiliser nutrient requirements for sites with differing growth potential, or in other terms, sites with differing degrees of limiting factors?

In assessing the fertiliser P requirements of pastoral zones with contrasting growth potential, the essential question is “ Do these zones have relative nutrient response curves which differ in their shape ?” If they do not, then the OVERSEERTM average curve will suffice for prescriptions to management units with widely differing growth potentials. If the curves do differ significantly in shape then the result will be a less than optimum prescription; lower in some cases and higher in others.

New requirements

Pastoral growth potential may be defined as the pasture production which will occur in the prevailing seasonal rainfall/ irrigation (available soil moisture) conditions where unlimited soil P and N (and other nutrients) are available. In reality all factors are limiting to some degree but for New Zealand pastures P and N are generally the greatest deficiencies. In general the response of pasture to a limiting nutrient is greater, and less curvilinear (ie more like a 1:1 relationship), when there are no other limiting factors. As these others become more apparent the pasture response to the limiting nutrient is less, and the shape of the response curve flatter (Gillingham and Power, 1984). This is evident on easy slopes of hill country, which have higher soil moisture, clover populations (N) and production than steeper slopes. The associated P response curve is flatter for steep than easy slopes because growth is limited more by lower soil moisture and available soil N levels (Gillingham et al, 1984).

It follows then that the fewer limiting factors present, the less curvilinear will be the response curve to an applied nutrient. So if management can address both N and P deficiencies, rather than P deficiency alone, then differences in actual growth on contrasting sites, due to differences in the remaining limiting factor (eg soil moisture), will be even greater.

A further refinement, which is worth pursuing, is the prediction of seasonal, rather than just annual responses to added nutrients. This is more relevant to N fertiliser use than for P, but because of the interacting effect of fertiliser N on pasture clover content, which can then have a modifying effect on the subsequent response of pasture to P, is relevant in a seasonal context also.

The final feature of any new or modified model developed to produce more refined recommendations for P and or N fertiliser use, is that it should allow an economic evaluation that will estimate the marginal returns from alternative options.

The conclusion from the above discussion is that a nutrient response curve derived as an average from a range of contrasting growth potential situations, eg the OVERSEER TM model, will not adequately represent all of these individual conditions, and will therefore not provide optimum recommendations, especially for those growth conditions on the margins of the response envelope.

It is not practicable to evaluate the responses of a range of nutrients over the full range of growth conditions that exist in New Zealand dairy and sheep and beef farms. It will be necessary therefore to develop prediction model(s) that can adequately represent these response conditions.

The new requirement is for a model that can predict actual pasture growth for a given set of soil fertility, soil moisture and temperature conditions. Such a model will contain a water balance, such as produced by Woodward et al.( 2001), modified to allow for seasonal, slope and aspect differences in solar radiation incidence ( Mc Aneney and Noble, 1976), and incorporate nutrient response functions which represent a range of conditions, such as developed by Moir et al.(2000) for P. The information to do this is probably already within our research data base.

Results of PA approaches to differential fertiliser application.

Recent studies on dairy farms (Gillingham and Betteridge, 2001), have illustrated the variability in production that can exist within dairy farm paddocks. They showed that within a 2ha paddock that three production zones could be mapped in mid winter, and that each of these responded differently to subsequently applied N fertiliser (Figure 1). The most economic, modified application recommendation would have saved about one third of the usual fertiliser applied without reducing pasture production. The differential application of fertiliser at this scale would require GPS guided machinery, or for the farmer to have marked the zones prior to applying N, maybe from a 4WD motorbike.

For a range of hill country farming conditions, a desk-top study by Gillingham et al. (1999) concluded that a differential fertiliser application approach, using the same total amount of P fertiliser, would improve net economic returns by 7.5-9.6% on low and moderate fertility farms respectively. In summer-dry hill country the differentiation in pasture species composition and growth on contrasting slopes and aspects can be greater than occurs on hill farms with more reliable and higher summer rainfall. On summer-dry farms there may be little or no legume growth on steep, north facing slopes, which contrast with adequate clover on south aspects. Consequently the need for a differential fertiliser application approach is even greater than on summer-moist hill farms. The results from recent farmlet trials (Gillingham et al.1998), showed that there was little point in continuing to apply P fertiliser to areas that had little or no clover and better responses were obtained by using urea (N) fertiliser. A subsequent desk-top study (Gillingham et al. 1999), showed that the economic returns from a differential fertiliser policy where capital P fertiliser was applied to easy slopes, and N fertiliser applied to steep slopes on both north and south aspects, were about 43% better than from a low P situation overall, and about 8.1% better than a policy of high overall capital P fertiliser application, at a lower total financial outlay than required for the overall capital fertiliser expenditure (Table 1). Where the hill land contains a higher proportion of steep slopes than the examples used in these studies, as has subsequently been shown to be more representative, then these estimated economic benefits of differential fertiliser application may be 2-3% higher.


The recognition of variability in productivity within pastoral landscapes is not new, but both because it has been largely un-quantified, and the technology has not been available to permit differential management, it has been largely ignored. Interest in mapping and researching production variability at scales smaller than previously able to be managed, together with the advent of GPS and GIS technology has now stimulated an examination of such management opportunities. The limited research to date, on a few examples, indicates that variable rate fertiliser application has significant potential benefits to pastoral farmers, and warrants the development of new fertiliser recommendation models, specific to contrasting land zones and pastoral growth potentials, in order to fully exploit unutilised potential and to allow the technology to be more widely applied.


Gillingham, A.G. and Betteridge, K.R. (2001). Opportunities for more precise pastoral management in New Zealand. . In Eds. L.D. Currie and P Logonathan. Occasional Report No 14. Precision Tools for Improving Land Management. Proceedings of the Massey University Fertiliser and Lime Research Centre Workshop,14-15th February. In Press.

Gillingham, A.G., Gray, M.H. and Smith, D. (1998). Pasture responses to phosphorus and nitrogen fertilisers on dry hill country. Proceedings of the New Zealand Grassland Association Conference, 60, 135-140.

Gillingham, A.G., Maber, J., Morton, J.D. and Tuohy, M. (1999). Precise aerial fertiliser application on hill country. Proceedings of the New Zealand Grassland Association, 61, 221-226.

Gillingham A.G. and Power, I.L. (1984) Influence of moisture and other factors on plant growth in hill soils. I. Soil P status, New Zealand Soil News, 32 (5), 200.

Gillingham A.G., Richardson S. and Riley J. (1984). Rationalising topdressing of hill country Proceedings of the New Zealand Grassland Association Conference, 45, 92-197

McAneney, K.J., and Noble, P.F. (1976). Estimating solar radiation on sloping surfaces. New Zealand Journal of Agricultural Research, 26, 7-13.

Metherell, A. (!999). Economic assessment of fertiliser strategies for pastoral farms. . In (Eds. L.D. Currie and P Logonathan) Occasional Report No 12 Best Soil Management practices for Production. Proceedings of the Massey University Fertiliser and Lime Research Centre Workshop, February, 127-142.

Moir, J.L., Scotter, D.R., Hedley, M.J., Mackay, A.D., Tillman, R.W., Turner, M.A. and Horne, D.J. (2000). Adding value to soil fertility- pasture growth relationships with a water balance. . In (Eds. L.D. Currie and P Logonathan) Occasional Report No 13. Long Term Nutrient Needs for New Zealand’s Primary Industries. Proceedings of the Massey University Fertiliser and Lime Research Centre Workshop, February, 225-239.

Woodward, S.J.R., Barker, D.J., and Zyskowski, R.F. (2001). A practical model for predicting soil water deficit in New Zealand pastures. New Zealand Journal of Agricultural Research, 44, 91-109.

Table 1: Comparative economic returns from uniform and differential fertiliser application to contrasting hill farm types (modified from Gillingham et al. 1999).

Hill farm type

Uniform fertiliser

Differential fertiliser

% Difference

Net Margin ($/ha)

Net Margin ($/ha)


Moist-Low fertility




Moist-Mod fertility




Summer dry- Low P+N




Summer dry-High P




Figure 1: Differential pasture growth zones A, B and C in Paddock 43 Massey University No 1 Dairy; June 2000., and associated responses to N fertiliser in each zone (from Gillingham and Betteridge, 2001).

Previous PageTop Of Page