Correction

Conversion of forest to pasture

The most dramatic transformation of the landscape occurred in the latter half of the nineteenth century when over half the remaining forest cover was removed (Guthrie-Smith 1969; Lancashire 1990; Arnold 1994; Knight 2009; Beattie and Star 2010). Forest removal continued at reduced pace until the end of the twentieth century (Ewers et al. 2006). Lowland forest that had not been felled or burnt was harvested to fulfil local and global demand for timber. Species such as cocksfoot (Dactylis glomerata) were introduced in the 1850s as a preferred grass to be sown into bush burn areas, through being hardier and more drought tolerant than ryegrass (Lolium perenne) (Rowarth et al. 1998). Most grass seed was imported from Europe until about 1880 when locally harvested seed reached sufficient quantities to supply domestic demand (Stewart 2006). The provision of cocksfoot seed was the beginning of what is now a successful forage seed production industry in New Zealand supporting domestic and international markets.

The clearance of bush over the last 200 years to establish an agricultural ‘industry’ follows the international pattern whereby forested areas are cleared to support an expanding population. This occurred in Europe between 500 and 900 AD and continues today throughout Asia and South America. The result is many formerly forested zones in the temperate and subtropical (and tropical) world have been cleared and ‘pastured’ with grasses adapted to the local conditions (Curtis et al. 2018). These pastures do not have a particular terminology other than ‘high rainfall pastures’ but could be included in the definition of mesic grassland (Ratajczak et al. 2014; Kalusová et al. 2017). In contrast, natural grasslands dominated in regions with rainfall below 600-700 mm. A meta-analysis combining a comprehensive list of European habitats and their species composition with a database of plant naturalisation records worldwide has shown that a broad habitat range, together with human-induced disturbance experienced in native-range habitats, can increase a species’ chance of becoming naturalised in other parts of the world (Kalusová et al. 2017), including New Zealand.

The loss of indigenous forest cover in New Zealand has parallels in other regions. In Australia, Victoria had 88% forest cover in 1869 but this is now at 34% (Australian Bureau of Rural Sciences 2010). Similar statistics are available for all continents subjected to European settlement. Indeed, Britain after the last ice age gradually expanded its forest cover to what would be a maximum at 6000 BC. From 4000 to 2000 BC the onset of agricultural development coincided with large fluctuations in woodland composition and taxa (Whitehouse and Smith 2010). By 1000 AD England had only 15% forest cover which declined further to only 5% by 1900. Of note, Britain went through its expansion of pastures over the course of 6000 years, while in New Zealand it occurred in only 200 years. Further, the transplanting of European species, including livestock, to New Zealand was reliant on only a few adapted species (Goldson et al. 2020). As part of the transplant, they had to cope with the existing native flora and fauna, particularly in the soil microbiome (Attwood et al. 2019). Significantly the native flora of New Zealand has not contributed to the pastures, except for residual natural grasslands in drier regions. Attempts to develop a native grass seed industry has resulted in only one species, Poa imbecilla used for revegetation purposes (Stewart 2005).

Successful land development was based on adhering to four major principles: (a) using the correct fertiliser treatment to overcome soil fertility deficiencies and lime to optimise pH; (b) using perennial ryegrass and white clover (Trifolium repens); (c) preparation of suitable seedbeds to ensure establishment; and (d) controlled grazing management (Smallfield 1956). Re-establishment of pastures which had become unproductive, or sowing after cropping, also highlighted the importance of seedbed preparation, sowing depth, method of sowing, time and rate of sowing, soil fertility and grazing management on cultivable land (Brougham 1969).

There are many ways to partition New Zealand into vegetation zones from a grassland perspective, but three clear zones have emerged since the advent of pastoral agriculture:

1. Maritime and temperate rainforest: ryegrass/clover-based pasture has predominated

2. Semi-arid tussock grassland: dryland species, cocksfoot, white clover, sub clover (T. subterraneum), lucerne (Medicago sativa) are common; and

3. Subtropical: kikuyu (Pennisetum clandestinum) and paspalum (Paspalum dilatatum) are significant.

At the first New Zealand Grassland Association conference held in 1933, Alfred Cockayne stated that

This reinforced the fact that New Zealand’s soils and climate were well suited to grassland farming.

It was recognised early (Connell 1933) that not all farms are equal. Farms that are essentially the same with respect to soil and climate can differ in carrying capacity by up to 80%. Hill country sheep farms of similar aspect and soil conditions can differ significantly in pasture quality and weediness. This difference was attributed to the gap between knowledge of how to advance grass farming and its general application in practice to achieve improved productivity. This realisation influenced the establishment of government sponsored technology transfer and advice through different consulting services within the Ministry of Agriculture after World War II (Te Ara 2008b). The government service was effectively abolished in 1990 and their role has been only partially replaced by private agribusinesses consultants. The latter tend to focus on financial advice and not all consultants are driven to find relevant and credible science. Analysis of the impact of Dairy Board consulting officers suggested that by all farmers using extension services, national dairy production would increase by 10% over 10 years, simply by improving methods of pasture utilisation (McKenzie 1980).

A key factor that improved grassland productivity was the phosphate revolution and use of lime (Langer 1977; Morton et al. 2005). The additions overcame nutrient deficiencies in native soils by increasing phosphate, sulphur and potassium fertility, which in turn enabled greater use of nitrogen-fixing clovers. Consequent improvements in soil nutrition have increased organic matter (Schipper et al. 2017) leading to improvements in the soil microbiome that contribute to more vigorous plant growth (Dignam et al. 2018, 2019). European earthworms (species of Lumbricidae and Megascolecidae) were also introduced (Lee 1961), although in many soils the species diversity of these is much lower than in European soils (Springett 1992). Native earthworms, of which there are 171 species, have a limited ecological range and live mostly in forests (Lee 1961). The lift in soil nutrition facilitated the shift from low fertility adapted species such as browntop (Agrostis capillaris syn. A. tenuis), Danthonia decumbens, crested dogstail (Cynosurus cristatus) and sweet vernal to more productive cocksfoot and then to ryegrass and white clover (Williams and Haynes 1990; Kemp and Lopez 2016).

Developments in seed and breeding

The importance of plant breeding to develop grasses and clovers for pastoral agriculture in New Zealand was recognised by the 1920s (Gorman 1934; Saxby 1934). Prior to that seed germination testing laboratories were established which lifted the quality of seed being sold to farmers (Cockayne 1913). New Zealand was one of the first countries to develop a seed certification scheme (Crump 1985). Beginning in 1929, the aim was to provide confidence in germination rates and line superiority (Broadfoot 1990). In this way, inferior types were eliminated as uncertified. In this process, both grass and clover ecotypes were collected from many regions of New Zealand. The best of these local ecotypes was superior in production and persistence compared with overseas germplasm. A superior ecotype of perennial ryegrass was identified from Hawkes Bay as early as the 1930s (Saxby 1934) and in 1955 a derivative of this became known as ‘Grasslands Ruanui’ (Stewart 2006). The first white clover cultivar followed in 1964, known as Grasslands Huia (Caradus et al. 1995a). White clover had been categorised into six types based on leaf size, stolon density and production (Gorman 1934). Type 1 ecotypes, collected from Hawke’s Bay and North Canterbury, were deemed superior. These had large leaves, widely spreading stout stolons and were highly productive and persistent.

The cost of pests and weeds

The introduction and success of largely European plant species, adapted to grazing in conditions predisposed to rainforest, occurred in effectively all these regions of the world. However, and importantly, the local insect fauna and soil microbiomes differed across regions. These differences, and the introduction of various cohorts of insects (sometimes deliberately, but more often inadvertently), has had enormously damaging consequences for New Zealand’s introduced pastoral systems (e.g. Goldson et al. 2020). Here it is estimated that the annual economic impact from the introduction of non-indigenous invertebrate pests ranges from $800 million to $2 billion (Goldson et al. 2005). The greatest impacts are caused by pasture nematodes (species of Paratylenchus, Heterodera, Meloidogyne, Paratrichodorus, and Pratylenchus) (up to $600 m/year) (Watson and Mercer 2000), Argentine stem weevil (Listronotus bonariensis) ($200 m/year), African black beetle (Heteronychus arator) ($242 m/year), and clover root weevil (Sitona lepidus) ($235 m/year) (Zydenbos et al. 2011; Ferguson et al. 2019). In addition, New Zealand pastures are badly impacted upon by the native scarab grass grub (Costelytra zealandica) causing losses of between $215 and $585 m/year, and a group of similar-looking porina moths (Wiseana spp.) that causes losses up to $170 m/year (Pottinger 1968; Ferguson et al. 2019). Lesser pests are also known to contribute to lost production. The transplantation of an incomplete ecosystem from Europe to New Zealand has led to disparity in many pests’ natural enemies in terms of both numbers and ecosystem distribution. Related to this, there is a lack of biotic resistance to invasive species. In the evolved and complex ecosystems of Europe, such problems associated with ecosystem imbalance occur far less often with less intensity (e.g. Goldson et al. 2017, 2020). The regulated introduction of biocontrol options, both parasitoids and biopesticides, has been a strategy used over many decades with mixed success (Fernández-Arhex and Corley 2003; Hawkins and Cornell 2008; Tomasetto et al. 2017, 2018a, 2018b).

The introduction of European grasses and clovers, and in some instances forage brassicas (mustard, rape and turnips), also resulted in the arrival of a range of undesirable and weedy species. Plant species which can invade and form self-sustaining populations impacting on pasture production and quality total about 187, of which 180 are introduced (Bourdôt et al. 2007; Ghanizadeh and Harrington 2019). Deliberate introductions of plants such as gorse (Ulex europaeus), to mimic its use in the UK as fencing hedges and windbreaks, simply resulted in the release of noxious weeds (Worsley 1999). Other plants species introduced for horticultural and amenity purposes or use in home gardens have also become weeds after escape from cultivation (Healy 1952). Whether deliberately or unintentionally introduced the number of non-native plant species in New Zealand is now greater than native plant species (Hulme 2020). Based on the estimated annual cost of pastoral weeds, California thistle (Circium arvense) is considered to be the most important followed by meadow buttercup (Ranunculus acris), gorse (Ulex europaeus), nassella tussock (Nassella trichotoma), Chilean needle grass (Nassella neesiana), blackberry (Rubus fruticosus), broom (Cytisus scoparius), barley grasses (Critesion spp.), ragwort (Senecio jacobaea); nodding thistle (Carduus nutans); brier rose (Rosa rubiginosa), and hawkweed (Pilosella spp.) (Bascand and Jowett 1982; Bourdôt et al. 2007; Saunders et al. 2017). These species have been primarily managed using selective herbicides (Ghanizadeh and Harrington 2019), and grazing management (Hartley et al. 1978; Hartley and Thai 1979; Betteridge et al. 1994; Espie 1994; Popay and Field 1996; Tozer et al. 2011b), with biocontrol options used having limited success (Bourdôt et al. 2007; Fowler et al. 2010; Suckling 2013) (Table 1 Supplementary Information). The estimated costs of both weed-covered land reducing production and weed control is about $1.2 billion per year (Bourdôt et al. 2007).

Seed technologies

The Plant Variety Rights law initially passed in 1973, and enacted in 1975 (Whitmore 1979), was a stimulus for commercial plant breeding and the gradual move away from government-funded and sponsored plant breeding activities (Paterson 1979). The history of forage plant breeding in New Zealand as it moved from introduced germplasm, to government sponsored breeding and then on to the private sector has been well described by Smith and Mather (1985). Commercial seed production processes and technologies to deliver a high quality and cost-effective product to farmers have been developed and refined over many years for all of the important, and of some more minor pastoral species (Rolston and Clifford 1989; Rolleston 2016) (Table 2 – Supplementary Information) but remains a challenge for more niche species (Monk et al. 2016). Improvements in seed production leading to higher yields per ha and improved seed quality, outlined in detail with appropriate references in Table 2 of Supplementary Information, include for:

• Annual and perennial ryegrass – appropriate plant density, use of nitrogen, weed control and use of growth regulators and fungicides.

• Tall fescue – appropriate plant density and use of nitrogen.

• Epichloë containing grasses – careful use of fungicides; low seed moisture content at harvest; treat seed infected with selected endophyte as a high-value perishable product.

• Cocksfoot – use of phosphorus and potassium.

• Prairie grass – use fungicide treated seed, use appropriate plant density and use of nitrogen and avoid grazing.

• Phalaris – no defoliation between sowing and harvesting; correct harvest time to avoid seed shattering.

• White clover – appropriate row spacings and sowing rate, avoiding overgrazing, close mid-November, harvest 1 month from main flowering, provision of bee hives, and appropriate irrigation.

• Red clover – reduce paddock size to improve pollination, close after early December, harvest March or early April.

• Lotus pedunculatus – spring sowing; herbicides to control clover; harvest when 70-80% of pods turn brown.

• Lucerne – low seeding rates, early sowing, good early weed control; use of alkali and leaf cutting bees.

• Chicory – time of closing, seed development, response to nitrogen, herbicide tolerance and methods of harvesting.

• Brassica – isolation distances within pollination groups for both insect and wind pollinated species.

Pasture renewal and renovation

Regular pasture renovation is an effective on-farm strategy to increase pasture yields. Compared with old pasture, this can amount to over 40% in the first autumn-to-spring period and then by 24% in subsequent years (Chestnut 1986). Some of this is due to the benefit of plant improvement which has been estimated to have added $12–$18 per ha per year (depending on region) operating profit on dairy farms since the mid-1960s (Chapman et al. 2017). It has been estimated that on dairy farms a 10% increase in dry matter yield due to pasture renewal has the potential to increase annual on-farm profitability from $271/ha to $478/ha (Brazendale et al. 2011). Glassey et al. (2010) demonstrated a yield advantage of 4% and feed quality advantage of 7%, contributing to the additional profit of more than $900/ha/yr. However, despite gains in pasture yield and quality with establishment of new pastures on dairy farms, pasture renewal rates have remained low – at about 6–8% p.a. (Sanderson and Webster 2009) and only 2–3% p.a. for sheep and beef (Caradus 2006). A 2010 survey of Waikato and Bay of Plenty dairy farmers has shown that they were more certain about pasture renewal decisions than about which cultivars or endophytes to use (Kelly et al. 2011). Various models have shown the payback period for pasture renewal is between 2 and 3 years on dairy farms (Stevens and Knowles 2011; Fraser et al. 2016). Sowing practices to ensure successful pasture establishment have been reviewed for both dairy (Thom et al. 2011) and non-cultivatable sheep and beef grazed hill country (Tozer and Douglas 2016). Decision-making about when and how to renew pasture can often be subjective but needs to be more informed via data to achieve the optimal gains through cost:benefit analysis (Kerr 2020).

Concerns by farmers that pasture renewal benefits are at times not realised (Kelly et al. 2011; Tozer et al. 2011a) have been examined by comparison of renewed and unrenewed pastures over a 5-year period (Tozer et al. 2015). Renewed pastures produced on average an additional 1.73 T DM per year over the first 3 years with a greater contribution of clover, sown grasses and unsown grasses and a smaller contribution of broadleaf weeds. In addition, the abundance of invertebrates was lower in renewed than resident pastures in some years. With the estimated cost of pasture renewal being $1300 per ha, an additional 6.5 T DM per ha would be profitable on most farms.

Success of pasture renovation can be improved using herbicides, sub-division, grazing management, appropriate fertiliser (Macfarlane and Bonish 1986), and use of salt in some circumstances (Gillespie et al. 2006), as well as using certified seed (Charlton 1991). On unploughable land the use of stock treading, while not as effective as herbicide, has been beneficial in promoting seedling establishment (Sithamparanathan et al. 1986). The development and use of direct drilling/seeding technologies practised as no-tillage and reduced tillage offer an additional dimension for sustainable agriculture globally in both seed industries and general food production (Stevens et al. 2000).

Managing feed quality

Pasture quality parameters that improve feed intake and nutritive quality are major determinants for improved liveweight gain, milk production as well as animal health and reproductive performance (Lambert and Litherland 2000). The two most commonly used measures of nutritive value are digestibility and metabolisable energy levels (Ulyatt 1970). Nutritive value is influenced by pasture species composition, and the age of pasture and the amount of dead matter present, plus the negative impact of some environmental factors such as high temperature, fertiliser imbalances, and extremes of soil moisture levels (Waghorn and Clark 2004). The inclusion of forage legumes and herbs is known to improve feed quality compared with grass monocultures (Golding et al. 2011). Forages which contain compounds such as condensed tannin can improve animal growth rates and health (Min et al. 2003; Waghorn and McNabb 2003; Woodward et al. 2004; Waghorn 2008), while other compounds such as ergovaline (Caradus et al. 2020) and lolitrem B (Lane 1999) from some Epichloëendophyte strains found in grasses can be detrimental. Effective novel endophytes which do not produce concerning levels of mammalian toxic secondary metabolites are now available (Caradus et al. 2021).

Grazing management through subdivision and electric fencing

The grazing management strategies available to New Zealand farmers have developed through extensive subdivision, resulting in improved feed budgeting and pasture use as well as reduced labour costs (Squire 1986; Jones 1988). Improved catchment management, water quality and reduced soil erosion has also been achieved (King 1969). In addition, subdivision has allowed the use of specialised pastures (Brown and Green 2003), effective management of spring feed surpluses (Lambert et al. 2000), and accommodation of differences between aspects/topography in hill country (Grant and Brock 1974; Lambert 1976). The ability of hill and high-country farmers to match pasture supply and quality to appropriate classes of livestock has enhanced on-farm productivity despite declining ewe numbers (Cocks et al. 2002; King et al. 2016; Stevens et al. 2016; Moot and Davison 2021). For dairy, grazing management principles for ensuring production and pasture persistence are well known (MacDonald et al. 2011). The developments have also coincided with improved weed control management (Rolston et al. 1982), and pasture establishment (Gillespie et al. 2006). Paddock subdivision per se is not, however, a panacea for poor on-farm management (Parker and McCall 1986). The use of electric fencing began over 80 years ago; the concept started in the USA but was quickly adopted, developed and manufactured in New Zealand, revolutionising management flexibility (Jones 1988). The advent of virtual fencing will further minimise fencing costs and increase flexibility of animal management (Brier et al. 2020).

Feed budgeting and pasture growth rate models

The development of models for understanding farm systems and to assist with decision making on-farm began in the 1970s and 1980s. Models were based on pasture growth rate data routinely collected by government agencies (e.g. Radcliffe 1974a, 1974b; Radcliffe 1975; Rickard and Radcliffe 1976; Morton and Paterson 1982; Roberts and Thomson 1984a, 1984b). The trends derived from these data continue to be used by farmers and consultants to develop feed budgets and for feed provision planning, particularly in times of drought or adverse weather events. Computer modelling has improved the prediction of pasture production with relatively few environmental inputs such as on-farm rainfall and temperature (Mills et al. 2021). These have resulted in decision-support tools such as the Pasture Forage Forecaster (DairyNZ 2021b), and Farmax which provides a planning and budgeting tool to test the commercial and biological feasibility of different land-use and management scenarios on farm (Farmax). The models have been largely based on ryegrass and clover mixed swards, with new equations provided for lucerne (Moot et al. 2021b). There remains a need to develop further components for other species, such as red clover and plantain. A central repository to consolidate the years of formal and informal data collection and allow the development of model components has recently been developed (Lincoln University; Moot et al. 2021a).

Pasture measurement techniques

The provision of information for feed budgeting requires reliable measurement of variations in pasture production, both annual and seasonal, across a range of scales from paddock, to farm and region (McNeur 1953; Cochrane 1976). This process can be laborious and so a variety of methods have been developed to expedite this task which are outlined in detail with appropriate references in Table 3 of Supplementary Information. These employ both direct and indirect measures and vary in terms of reliability, accuracy and ease of use and affordability. Direct measures include cage and trim technique, possibly using a mower or hand clippers to a constant height, by taking random quadrats. Indirect measures include:

• Visual assessments.

• Pasture height.

• Capacitance metre.

• Weighted disc/rising plate metre.

• C-DAX rapid pasture metre.

• LIDR – Light Detection and Ranging.

• Aircraft photography using colour and panchromatic photography or multispectral cameras

• Satellite imagery using multispectral scanners

• Hyperspectral imaging

• Fixed invisible near-infrared light sensors

• Photogrammetry

Technologies for measuring and managing soil and water conservation

The instability of up to 40% of New Zealand landscapes (Gibbs 1963), particularly steep hill country with shallow soils, has been long recognised (Cumberland 1947; Wilkie 1954; King 1969; Tran et al. 2020). In 2008, this was estimated to cost the New Zealand economy $157 million per year (Jones et al. 2008). The list of methods and technologies to manage both soil and water conservation is extensive. It includes planting trees (McGregor et al. 1999; Wilkinson 1999; Douglas et al. 2013), grazing management to stop fragile soil over-grazing., inappropriate stock class use (e.g. heavy animals on wet steep hill country) (Hicks et al. 2011), cultivation confined to easy contours (Basher 2013), fencing and riparian planting along waterways (to trap sediments and nutrients) (Bewsell et al. 2007; Daigneault et al. 2017), water management through using artificial channels to direct water flow and the use of dams to manage debris and water flow (Gregg 2008). Additionally, dense vigorous pasture cover will reduce sheet, rill and wind erosion but, largely because of the shallow root systems, the pastures will not prevent slips and slumping (Blaschke et al. 1992; Basher 2013).

Protecting soils during wet winters, particularly on dairy farms, has led to several off-paddock facility developments (Longhurst et al. 2006a, 2006b), thus allowing the removal of animals from pasture when it is prone to treading damage. These innovations can lead to reduced overland flow of water and nutrients, less nitrate leaching and lower greenhouse gas emissions on poorly drained soils (van der Weerden et al. 2017). Cost-effectiveness continues to be under debate (Laurenson et al. 2017), but depends on type of facility, climate, soil type and, of course, what factors are modelled.

A range of soil and water conservation managed Environmental Farm Plans have continued to be applied voluntarily (Manderson et al. 2007; Cameron 2016), but they have recently become a regulatory requirement (Ministry for the Environment 2020) managed by Regional Councils. Their success in achieving world excellence in environmental management is contingent upon plans being farmer-centric, farmer-owned and adding value to their business (Stokes et al. 2021) but is likely to require further scientific research.

Irrigation

From 2002 through to 2019 the area of irrigated agricultural land in New Zealand increased from 384,000 ha to 735,000 ha (Stats NZ 2021a), which is 2.75% of New Zealand’s total land area of 26.7 million ha (of which about half is farmed) (Stats NZ 2021b). Irrigated pastureland is now 16% dairy compared with 2% each for sheep and beef. The largest irrigation schemes were developed in Canterbury in the late 20th and early 21st centuries (Young et al. 2004), although the first recorded irrigation system in Canterbury was constructed by a Wakanui farmer, Joseph Hunt, in 1878 (Lobb 1968). Initially, the ‘wild-flood’ irrigation schemes of Central Otago were derived from previous water rights obtained for gold sluicing in the 1880s (Heiler 2008). While in Canterbury irrigation trials started in the 1880s, large scale irrigation border dyke systems did not start until the 1930s (Taylor 1974; Taylor et al. 1985). These have been now largely replaced by overhead pivot or lateral irrigation technology (Evans 2004; Pangborn and Woodford 2011; Saunders and Saunders 2012). In 2007, irrigated land nationally (about 4%) contributed about $1 billion per year (Le Prou 2007), and by 2012 this was estimated to be $2.17 billion (NZIER 2014).

The often-inefficient designs of early irrigation schemes have been resolved with improved designs and increasing water use efficiency (Taylor et al. 1985). However, to manage irrigation water efficiently farmers need to measure and monitor irrigation application depths and uniformity (Thomas et al. 2006). Advances in precision and variable rate irrigation have reduced water use, reduced nitrate loss and decreased energy requirements (Hedley et al. 2009; McDowell 2017).

Fertiliser – superphosphate, mineral nitrogen and lime

On most New Zealand pastoral lands, soil science research has guided the use of both phosphorus and sulphur to improve pasture quantity and quality (Daly et al. 1999; Roberts and White 2016). These nutrients are applied to encourage the legume component which also requires lime to ensure the pH is >5.5 for micronutrient availability for plant uptake (Widdowson and Walker 1971). Specifically, molybdenum is essential for effective N-fixation of legumes and reduction of nitrate as a step towards protein synthesis (Sherrell and Metherell 1986). In New Zealand, deficient levels of molybdenum were first discovered through investigation into a disease of cauliflowers termed ‘whiptail’ (Davies 1952), managed by the application of small quantities of ammonium molybdate (Mitchell 1945; Sherrell and Metherell 1986). On cultivatable land it is recommended that lime be used after correcting for other major nutrient deficiencies (Edmeades et al. 2016).

Aerial top dressing, first trialled in 1949 (James 1984), has been viewed as synonymous with successful hill country development (Tebb 1959). Improved returns can still be achieved through new technologies that allow the application of fertiliser at variable rates to match the growth potential of contrasting hill country zones (Gillingham et al. 1984, 1999; Roberts and White 2016). Further work is required to improve the accuracy and precision of differential rates of application (Chok et al. 2016) and is the subject of considerable research (e.g. Yule et al. 2015).

Large amounts of seed and fertiliser were applied to unploughable steep hill country from the 1930s (Smallfield 1938) with the best establishment of introduced clovers occurring in early autumn (Suckling 1951). Applications of phosphate were most effective when applied after the seedlings were well-established.

Nitrogen is the main limitation to pasture production (Carran 1978; Mills et al. 2006) and has therefore become the main fertiliser used to drive pasture production in dairy systems (Moot et al. 2020). This is particularly the case for irrigated areas, while in rain-fed regions nitrogen is used more as a tactical tool to meet short term deficits in the shoulders of the season (early spring and late autumn). Prior to the 1990s, the predominant source of nitrogen in pastoral ecosystems was through companion clovers (Ball 1969; Brock et al. 1989). Analysis in the 1980s concluded that use of high rates of nitrogen fertiliser (> 80 kg N/ha) was unlikely to be profitable on dairy farms (Buxton 1981; Bryant et al. 1982; Holmes 1982). However, with the advent of cheap synthetic mineral nitrogen fertiliser and the increase in the value of milk solids there was a major shift to its use in the 1990s. This coincided with the arrival of clover root weevil (Sitona obsoletus) that led to an even greater reliance on synthetic nitrogen (Mills et al. 2006). Rates of up to 400 kg/N/ha and irrigation have enabled stocking rates to increase to 3.47 cows/ha resulting in an increase of average herd sizes to 810 cows, compared to the New Zealand average of 2.84 cows/ha and herds of 415 cows (DairyNZ 2021a). Nitrogen application resulted in predictable pasture growth over short time periods which, for a dairy farmer, was able to be captured as milk ‘in the vat’ and therefore improved on-farm profitability (Barr 1995). Combining the advantages of clover in terms of per cow performance (Harris et al. 1997, 1998) and nitrogen for increased dry matter production was considered possible at up to 200 kg N applied/ha/year if the additional pasture was utilised in spring to maintain clover content (Harris and Clark 1996) and stocking rates or conservation polices were modified to use the additional feed (Harris et al. 1994).

Similarly, for sheep farming, the use of synthetic nitrogen in late autumn was found to provide a 14% increase in weaned lamb liveweight (Lambert and Clark 1986). On dry hill country, over a 7-year period, the use of nitrogen at 30 and 50 kg N/ha/year applied in early to mid-winter resulted in increases of $30/ha and $50/ha compared with no nitrogen applied (Gillingham et al. 2004). The key is to maximise pasture use through increased stocking rates (Luscombe and Fletcher 1982) during the period of reliable soil moisture and pasture growth conditions in the winter/spring. On dryland, yields of cocksfoot increased from 6.5–16 t DM/ha/year with the application of synthetic nitrogen (Mills et al. 2006). Response rates of over 20 kg DM/kg of applied nitrogen have been recorded in hill country pastures (Luscombe 1979; Lambert and Clark 1986; Gillingham et al. 1999; Fasi et al. 2008), which points to severe nitrogen deficiency. On higher fertility dairy pastures, a response of less than 20 kg DM/ha per kg N applied can be expected (Glassey et al. 2013). However, with nitrogen fertiliser application a significant reduction in clover content can occur (Luscombe and Fletcher 1982; Harris et al. 1994; Harris and Clark 1996).

Impacts of mineral deficiencies in NZ pasture on animal health and production – Co, I, Se, Cu and Na

Many New Zealand soils have low natural levels of essential trace-elements, e.g. selenium, cobalt, copper and iodine, all of which are required for healthy animal production (Grace and Clark 1991). This can be remedied by topdressing the pasture, oral dosing and/or injections (Grace 1992). Low levels of selenium affect up to 30% of New Zealand’s agricultural area (Grace 1994; Grace et al. 2011). This is not required by plants, but selenium blood concentrations and production responses by grazing animals have been positively correlated (Wichtel 1994; Wichtel et al. 1994). Supplementation options include drenches, injectable products, slow release intra-ruminal boluses or selenised feed additives designed for farms with extensive grazing and/or low stocking rates (Metherell et al. 1996). However, it has been demonstrated that the direct application of selenium (usually with fertiliser) to intensively grazed pastures is a highly effective long-term strategy (Watkinson 1989; Moorhouse et al. 1999; Grace and West 2006).

Cobalt is important in the synthesis of vitamin B12 in the rumen. Cobalt deficiency in animals is usually associated with specific soil types, including the pumice soils of the North Island and granite soils of Nelson (Andrews 1955). Cobalt topdressing of affected land has become an accepted practice (Metherell 1989; Hawke et al. 1994).

Iodine deficiency has been a feature of New Zealand soils and can result in impaired reproduction and perinatal mortality of both sheep (Sargison et al. 1998) and cattle (Mee et al. 1995). Even recent surveys have shown that iodine concentrations in 26% of pastures were insufficient for sheep nutrition, and 87% of pastures for cattle nutrition (Jensen et al. 2019). Management of iodine deficiency has been through direct animal treatment (Grace 1992) but the use of iodine fertiliser to pastures can lift herbage iodine concentrations to acceptable levels (Smith et al. 1999), although this has not been applied in practice.

Copper deficiency can cause nerve and bone disorders in lambs, and poor growth and reproductive problems in cattle and can be exacerbated by high molybdenum levels in pasture (Grace 1969). The application of copper through fertiliser and by licks can eliminate these issues for both sheep and cattle (Cunningham 1944), but copper boluses are more commonly used in dairy animals and deer, where required.

Sodium, an essential element for both animal health and production, can be at low levels in some regions of New Zealand (Edmeades and O'Connor 2003) particularly those regions inland from the coast (Suttie 2010). Treatment of suspected sodium deficiency can be through NaCl oral drenching, water trough treatment, salt licks or fertiliser Na applications (Edmeades and O'Connor 2003).

Consumer and market demands

On-farm productivity and profitability

On dairy farms, operating profit from productivity (PFP) is the difference between actual operating profit and the operating profit that would have occurred with no productivity changes since a base year (). The main contributory components of PFP are milk solids production increases since the base year, operating expenses savings in inflation adjusted terms since the base year, and end-of-year operating profit per kilogram milk solids (influenced by milk prices). DairyNZ and dairy farmers must increasingly focus on achieving cost-efficient milk solids production and PFP provides a means for monitoring the value of productivity gains over time. However, the volatile nature of PFP () is caused by fluctuations in both milk prices and seasonal rainfall, both of which are largely outside the control of dairy farmers.

Similarly, for sheep and beef farmers, profitability is at the centre of decision-making when considering changes in management and uptake of new technologies. Farmers need to understand and have the confidence to make changes which in the future will include supporting a community with a rewarding working environment and lifestyle. This will involve improvements in productivity (i.e. more output with the same, or lower physical inputs), managing external pressures especially with the environment (e.g. riparian management and water quality) which will increase compliance costs, and satisfying new markets with a greater spread of product supply off-farms (Fennessey et al. 2016).

Future research and development requirements

A review of scientific and technological advances in pastoral agriculture based on publications in the New Zealand Grassland Association conference proceedings over 75 years concluded that ‘New Zealand’s international competitive advantage in producing agricultural products will be maintained through the ongoing application of innovative technologies and smart business practices leading to an increase in on-farm efficiencies, productivity, and added value’ (Caradus 2006). The same still applies today, where research and development investment is a prerequisite for continued improvements in pastoral agricultural productivity (Caradus 2007) and environmental sustainability. This will require:

1. Development of new technologies and their uptake by farmers to improve productivity, biosecurity, feed supply options, environmental integrity, energy use efficiency, and the opportunity to create new added-value products

2. Maintaining high objective standards of scientific integrity and scrutiny

3. Providing opportunities for debate and interaction within the interested community, viz. farmers, extension specialists, agribusiness, policy advisors and makers, regulatory managers, and scientists.

Plant-based and insect-based proteins, and in vitro meats

Plant-based diets have been heralded as the means to a sustainable global food system resulting in reduced greenhouse gas emissions, improved animal welfare, and enhanced human health (Ferdowsian and Barnard 2009; Faber et al. 2020; Morris and Livesey 2020; United Nations 2020). However, analysis has revealed that alternative protein companies are reporting only their own impact on the environment, not that of the supply chain (Ceres 2021), suggesting that the impact of the plant and energy sources required for product development are ignored.

Willingness to accept a plant-based diet can vary depending on the consumers’ cultural origins, their diet preferences, awareness of animal production systems (Wang and Scrimgeour 2021), or in the case of insect-based protein, the fear of the new and unfamiliar (de Koning et al. 2020). It is acknowledged that economics, health/nutrition and aesthetics/taste are important factors in determining consumer food choices rather than explicitly environmental benefit and sustainability (Tucker 2018). A survey of over 1000 consumers from Germany and New Zealand balanced across age, gender and income showed a general preference for meat based rather than in vitro meat diets (Lemken et al. 2019). Modelling has indicated that manufacture of cultured meat is not necessarily environmentally superior to cattle (Lynch and Pierrehumbert 2019). Additionally, it is proposed that plant-based diets will incur higher cost to households (Kidd et al. 2021) and on their own may not provide the full nutrition needed for a global population: balanced diet is plant based and animal optimised (Smith et al. 2021). For New Zealand’s introduced grassland, the prospect of using grass protein concentrate has shown that refining the extraction method was crucial for achieving optimum protein functionality during its use for food applications (Kaur et al. 2021). In vitro meats are being produced using animal cells under laboratory conditions (Post 2012). In New Zealand there is little understanding of this potential food source, and consumers are hesitant to engage with in vitro meats due to lack of familiarity (Malavalli et al. 2021). In Europe and USA, many consumers appear willing to eat in vitro meats, although in the case of the USA it is deemed unlikely to replace farmed meat in their diet (Hocquette et al. 2015; Wilks and Phillips 2017; Bryant and Dillard 2019). The debate between those promoting alternative protein sources and those supporting the value of conventional livestock production looks set to continue (Sexton et al. 2019).

These latest technological developments could be seen to threaten New Zealand’s pastoral sector and indeed, it is possible that market share may decline in some areas, but this will coincide with rapidly growing middle-class populations, particularly in Asia. Based on its in situ grassland farming, New Zealand can rightly point to the natural quality of its products and this has appeal compared with vat-based microbial culture production systems. Additionally, it has been estimated that if 10% of the world’s meat consumption (i.e. 40 million tonnes) were to be supplied as cultivated meat this would require 4,000 ‘factories’, each with 130 bioreactor lines, and each of these having 10,000 L bioreactor tanks (Food Navigator 2021). It is therefore quite probable that the demand for absolute volumes of premium grass-fed protein will remain or even grow with world protein supplies under threat and demand increasing (McLeod 2011). The importance of protein in reducing malnutrition (Adesogan et al. 2020) ensures that there is a global market sufficient to support both animal and plant-based protein production. The challenge for New Zealand is to ensure that animal protein production is clearly and demonstrably efficient in the use of resources, including land, water, nutrients and energy. This will require new research leading to new technologies providing greater efficiencies of production with reduced and more sustainable inputs. Every effort must be directed to showing that animal-based protein production is not unsustainable, and that consumption of animal protein is a healthy option (Alparslan and Demirbaş 2020).

New Zealand could also become the producer of premium feedstocks that can be used in some of these vat-based foods. Arguably New Zealand’s impressive productive base could become more mixed although any transition will be gradual. However, it remains to be seen what the effect of the COVID pandemic will have on patterns of international food production and trade (Aday and Aday 2020; Kaiser et al. 2021).