5 Grazing land management

5.1 Specific activities considered, and effects not covered

Grazing land is the internationally accepted term to cover what in New Zealand is essentially pasture land, and the international term is used here for comparability with other countries. Grazing land management encompasses all practices aimed at manipulating the amount and type of forage and livestock produced, including regulation of animal stocking rates, forage species selection, fertilization, liming and irrigation.

Grasslands (pasture lands) of New Zealand are, for the most part, recently created plant communities. Their composition, distribution and productivity vary with change in landform, climate, soil, and human management of animals (Daly 1990). New Zealand is a mountainous country and much of our grasslands occur on hills. Tall tussock (Chionochloa) species are the natural vegetation on subalpine zones above the treeline, but at lower altitudes have been induced by natural and man-made fires. Grazing, combined with regular burning of grasslands to produce new shoots palatable to stock, as well as, periodic rabbit infestation have reduced both tussock biomass and coverage.

Of the 12.01 Mha of pasture, arable, horticultural and grazed tussock land (MAF, 2001), the 0.3 Mha component of arable and horticultural land is covered under Cropping Land above. There are about 5 Mha of surface sown pastures on unploughable hill country out of podocarp rain forest, fern or scrub, largely in the North Island (White 1990). In addition, there are over 2 Mha of hill and high country in eastern districts of South Island partly covered in tussock grasslands above 1000m and under the Conservation Estate. The total grassland area, according to 1996/97 Land Cover Data Base (LCDB), is 13.73 Mha [12.01-0.3+2.02]. Consistent with this, a 1990 value of 13.4 Mha has been used in Table 5.5 from the estimates based on Vegetative Cover Map (VCM) for 1992 (Tate et al 1997).

Pastoral farming is the backbone of New Zealand agriculture. The key to grassland improvement in hill country is the introduction of high producing legumes after the correction of soil nutrient deficiencies. Improved pastures, producing 10-20 Mg DM ha-1, consist of perennial ryegrass and white clover, and require regular fertiliser application. Superphosphate has historically played a vital role in improving and sustaining productivity of legume-based pastures, because fertiliser phosphorus improves legume growth, resulting in increased soil nitrogen status. Unimproved or reverted hill pastures are brown-top dominant, with little or no clover, and produce 3-5 Mg DM ha-1. In between are lowland farms or well-developed hill farms producing 5-10 Mg DM ha-1.

Grassland is managed with the dual goals of matching animal feed demand with supply while ensuring the longer term viability of the plant production system. This is achieved in practice in a variety of ways but varying the intensity of grazing (that is the proportion of the standing crop that is utilised) is the key step. Highly productive systems use far more of the NPP for animal production than do less productive systems. More or less fixed carbon is therefore routed through the productive pathway than through the decomposer network (Brock et al 1990). An example of this is the carbon loss experienced when a more intensive system such as dairying replaces a less intensive system such as sheep farming (Ghani et al 1996). Given the complexity of plant responses to management and its interactions with weather, a range of positive, negative or neutral feedbacks to soil carbon content should be expected and these may vary with time.

Grazing land management includes: fertilisation and irrigation to increase biomass, manipulation of the intensity and duration of grazing to match feed supply; selection of stock class to match seasonal patterns of biomass supply; planting on pasture and hayland to produce high quality forage, improving efficiency of ruminant livestock; mechanical treatment to increase vegetation production capacities; and biological and chemical management to maintain or improve plant community sustainability and productivity (e.g., prescribed fire, herbicide applications, and introduction of organisms to control invasive plants and plant community dynamics).

Many established pastures can be considered to have attained steady-state soil carbon levels (Tate et al 1997). Shifts in grazing land management may or may not change soil carbon stocks. This will depend on the length of pastoral farming, soil type (Saggar et al 1996, 1999b), level of net primary productivity (NPP) and the degree of utilisation all of which affect the time the pasture ecosystem will take to attain a new steady-state condition.

Soil texture and mineralogy play a major role in stabilising soil carbon (Saggar et al 1994, 1996, 1999b). Thus soil carbon contents generally increase with an increase in clay content, and soils containing allophane can hold even higher amounts of carbon. A combination of pyrophosphate extractable-Al, Fe oxide, allophane and clay concentration explains the greatest amount of variation in carbon concentration in New Zealand soils (Percival et al 2000).

The implications of erosion are discussed in section 7.4. The effects of fire, herbicide applications, and introduction of organisms to control invasive plants are not covered but are also not thought to be large. As in other sections of this report the focus here is on carbon stocks and the implications of land-use change for fuel and energy related emissions have not been considered as we expect those to be treated under Article 3.1 of the Protocol.

5.2 Effects on non-CO2 greenhouse gases

New Zealand's grazing lands are used primarily to support ruminant livestock. Recent ruminant numbers are shown in Table 5.1. Currently we have 45.2 million sheep, including 30.3 million breeding ewes, 4.2 million dairy cattle, 4.6 million beef cattle, 1.6 million deer, and 0.2 million goats, which collectively account for about 88% of all New Zealand's anthropogenic methane sources. Livestock methane is produced largely in the rumen as a result of microbial (methanogenic) activity and depends to some extent on the composition of pasture grasses (Lassey and Ulyatt, 1998).

Despite an 11 million decrease in sheep numbers and 0.8 million decrease in goats, there have been only small decreases in nitrous oxide and methane emission from animals between 1990 and 1999 (Table 5.1). This is due to an increase of 0.9 million dairy cattle and 0.7 million deer with higher per animal emission factors. Future changes to dairying from sheep in the Southland may result in further increased emissions of non-CO2 gases.

The majority of nitrous oxide in New Zealand is also emitted from grazed grasslands due to the large nitrogen input from fertiliser, fixation through legumes in ryegrass/clover based pastures, and the rapid cycling of nitrogen via grazing animal excreta. Nitrous oxide emissions also depend on pasture management (nitrogen fertilization, legume component), stock type and soil type.

Table 5.1. Changes in New Zealand livestock population, and livestock emissions of nitrous oxide and methane calculated using IPCC method based on NZ excretion and emission rates for each category of animal.

Livestock type Livestock number (Millions)a Nitrous oxide (kt y-1)b Methane (Tg y-1)b
1990 1999 1990 1999 1990 1999
Sheep 56.80 45.20 6.71 5.33 0.87 0.69
Dairy Cattle 3.40 4.24 2.94 3.68 0.26 0.33
Beef Cattle 4.50 4.58 2.85 2.89 0.31 0.31
Deer 0.95 1.64 0.19 0.32 0.03 0.05
Pigs 0.38 0.37 0.06 0.06 0.00 0.00
Goats 1.00 0.18 0.09 0.02 0.02 0.003
Total 67.08 56.21 12.84 12.30 1.49 1.39

a Animal numbers are taken from Ministry of Agriculture and Forestry, New Zealand website http://www.maf.govt.nz/statistics/primaryindustries/
b Nitrogen excretion and methane emission for each type of livestock were those used by Ministry for the Environment. Nitrous oxide emission calculated as 1% of excretal N. (Ministry for the Environment 2000).

As shown in Table 5.2, the relative importance of methane and nitrous oxide can vary substantially among farming systems depending largely on soil type - particularly drainage characteristics - and nitrogen inputs.

Table 5.2. A comparison of methane and nitrous oxide emissions (direct and indirect) from two sites with different drainage characteristics (Carran and Theobald 2000a).

Site Tokomaru Manawatu Manawatu + 400 kg N
  Kg C or
N ha-1 y-1		 T CO2
ha y-1		 Kg C or
N ha-1 y-1		 T CO2|
ha y-1		 Kg C or
N ha-1 y-1		 T CO2
ha y-1
CH4 animals 240 5.9 330 8.4 495 12.5
N2O soils 7 3.6 1 0.4 5 2.5
N2O off site 1 0.4 1 0.4 3 1.4

In the poorly drained but productive Tokomaru soil the nitrous oxide emissions were equivalent to 68% of methane emissions; while in the adjacent Manawatu soil nitrous oxide emissions were 9% or 31% of methane emissions depending on nitrogen fertiliser use.

5.3 Definitional issues

Some issues arise in relation to definition for grazing land, e.g. whether the definition is based on land-use or land-cover; at what point land would change its classification if it were retired from use as pasture; how frequently the classification would have to be updated. Such factors will have an impact on area estimation.

The treatment of CO2 fertilization has been raised in relation to Articles 3.3 and 3.4 as there is a view that this is not a "human-induced" effect and so should not attract credits. Ross et al (1996) showed that, in the short-term, elevated CO2 may not necessarily lead to greater carbon sequestration by New Zealand temperate pastures, but is likely to result in greater carbon turnover. However, based on observed effects on grassland over many decades, elevated CO2 could lead to increased soil carbon levels in grassland (Ross et al 2000b).

5.4 Socio-economic factors influencing the activities considered

Changes in grazing land area and its use are driven by commodity prices (wool, meat, and dairy products). With increased profits in dairying, the area under sheep, and sheep/beef farms is declining slightly as farmers convert to dairying. The area under irrigation is increasing and, with increased long-term profits in plantation forestry, some of the hill and high country sheep pasture is being converted to Pinus radiata.

Withdrawal of government subsidies on fertilisers in the early 1980s resulted in a short-term decline in pasture fertility and an increase in unimproved pasture area. Similar effects will continue to result from changes in input costs (e.g. fertilisers and application costs), land values, and the availability and cost of irrigation in some areas. Climate variations can affect the profitability of grazing land as seen in the retirement of pasture lands on the West Coast following recent strong El Nino - La Nina variations resulting in floods and droughts.

It should be noted that some pastoral farms around major cities are initially converted to arable/vegetable cropping and this high-class land is then subsequently sub-divided for life-style blocks.

5.5 Applicability of a land-based accounting system

A national soil carbon inventory for grazing land has been produced from site-based soil carbon measurements combined with land area estimates (Tate et al 1997).

Agribase farm surveys (SONZA and MAFStats) give acceptable areas for pastures. However, these surveys neither consider spatial referencing (location within farms) nor they are consistent with each other. A National Soil Carbon Monitoring System (CMS) should soon be operational and will improve this situation.

All estimates of soil carbon used here are based on assumed steady state values for current management and climatic conditions. While this is currently believed to be a good approximation for grazing lands, many of these parameters may change with the evolution of new management practices in response to economic, social, technological, political and genetic developments. Thus carbon accounting methods must be able to treat situations in which transient behaviour is expected.

5.6 What data, publications and reviews are available

Tate et al (1997) derived soil organic carbon estimates from the NZ Land Resource Inventory, the Soil Map of Stewart Island, and the New Zealand Soils Database (NZSD), and plant biomass carbon estimates from the Vegetative Cover Map (VCM) of New Zealand. They also observed a 1:1 relationship (R, 0.68) (see Fig. 5.1) over a wide range of soil carbon contents for pasture soils sampled recently and 30-50 years ago, indicating the majority of NZ pasture soils are at or near steady-state and estimates of soil carbon made from the NZSD are likely to be representative over time (provided no major land-use change occurs).

Long-term fertiliser applications under cropping and pasture have been shown to either: increase soil carbon (Jackman 1960; Nguyen and Goh 1990; Haynes and Williams 1993); have no effect (Campbell and Zentner 1997; Saggar unpublished data); or decrease soil carbon (Lambert et al 2000). The increase in soil carbon due to fertilisation is often attributed to an increase in net primary production (NPP), and thereby increased crop residue or root growth. In well-established pastures carbon is allocated equally between above- and below-ground (Saggar and Hedley 2001). Soil fertility and moisture availability influence this allocation (Saggar et al 1997, 1999a). The degree of accumulation of organic carbon depends on the relation between carbon input and decomposition rates, and results from a complex interaction of ecosystem properties. As the rate of carbon input approaches that of decomposition, soil carbon accomplishes a steady-state. Any perturbation that disrupts this steady-state disproportionately affects carbon input or decomposition resulting in change of soil carbon (Jenzen et al 1998). Fertilised pastures translocated twice as much carbon to roots as non-fertilised pastures (Saggar et al 1997, 1999a). Higher proportions of the carbon inputs were decomposed in the fertilised than in the non-fertilised soils (Saggar et al 2000), resulting in only 4% increase in carbon concentration in the 0-10 cm depth of fertilised soils over the 16-year period.

Figure 5.1. Carbon contents (%) of 43 topsoils sampled recently (CR) and 30-50 years ago (CA) in North Island and South Island pastures. Values for soils drained since the original analyses are shown as dot.gif (70 bytes)

graph6.gif (2284 bytes)

Table 5.3. Size of pools relevant to carbon cycling in pastures managed by different grazing systems (data from Brock et al 1990). Within pool differences were significantly different between each management.

Pool Rotational grazing Set stocking Combination grazing
Residual herbage dry matter (kg ha-1) 2345 3780 3010
Microbial biomass C (µg/g dry soil) 747 843 899
Total soil organic matter (%) 0-7.5 cm1 7.24 7.68 8.07
Soil organic matter light fraction (%) 0-7.5 cm 0.131 0.169 0.146

1Soil organic matter was estimated by loss on ignition

The carbon storage of managed grassland can be manipulated by fertiliser application and through grazing (Table 5.3). Nitrogen fertiliser will increase NPP of grassland and, assuming there is no change in utilisation (the amount used by animals), then the increased carbon inputs will result in increased carbon storage in soils. Data on the relationship between N fertilisation and carbon storage are difficult to interpret because in most cases stocking rate is not held constant. Results from modelling suggest a 35% increase in carbon sequestered with a 100 kg ha-1 increase in N (Table 5.4)

Studies by Saggar et al (1997, 1999a, 2000) suggest that fertiliser addition not only resulted in increased production and increased below-ground carbon inputs to the soil but enhanced the decomposition of soil organic matter, and thus increased the rate of carbon loss. Ghani et al (1996) also found that three Waikato soils under higher nutrient input dairy contained significantly less soil carbon (~1.5% to 0-75 mm depth) than soils under sheep-beef. Thus, it appears that addition of nitrogen is unlikely to improve the soil carbon stocks in majority of the developed pastoral soils. Added nitrogen, however, may enhance nitrogen mineralisation rate and contribute to additional nitrous oxide loss.

Table 5.4. Effect of nitrogen fertilization on carbon density. Data from Table 1 of Thornley et al. 1991.

 

N (kg ha-1 y-1) C sequestered at steady state (kg C m-2)
5

11.9
15

12.5
30

13.2
100

16.1

However, grazing animals can significantly modify the proportion of any increased aboveground carbon inputs (resulting from fertiliser or irrigation application) moving to soil carbon by determining the relative amounts of NPP going to animal production (product and respiration) or to the decomposer pathways (and hence into the soil carbon pool). In practice, the aim of the grassland manager will be to use the extra dry matter resulting from N fertilisation for animal production thus reducing the carbon storage potential of the fertiliser application. Approximately 80% of the carbon ingested by sheep is lost to the atmosphere through respiration and methane production. Maximum storage would result therefore from increased carbon inputs from fertiliser with no animal utilisation but this defeats the object of grassland farming. As utilisation increases, the potential for carbon storage therefore declines to a low level; however, the relationship between utilisation and storage is non-linear (Thornley and Verbene 1989) offering the possibility of manipulating carbon storage by grasslands through management without necessarily compromising farm returns.

Using a carbon accounting model (ACSD), a Dutch group has recently estimated changes in carbon stocks from 1990 to the present for several countries, including New Zealand. The estimated changes resulted from a range of land-management scenarios, including increased productivity of pastures, erosion control, manure application, etc. The areas being used for grazing land were close to those in this report, but in the absence of actual New Zealand data for 1990 carbon stocks, default values from Australia and the US were used for New Zealand. Inaccuracies in such global calculations of New Zealand carbon stocks and changes need to be viewed with caution by our representatives at international fora. This emphasizes the importance for New Zealand to make its own robust carbon stock and change estimates.

5.7 Sampling techniques and issues

There are issues with the LCDB database concerning spatial resolution and thematic resolution, e.g. the proportions shown as bareground include large areas of sparsely vegetated grasslands, typically at high altitude. In addition pasture is not well differentiated from tussock in some areas. Improved remote sensors and development of topographic corrections are expected to address this issue.

Soil carbon data is generally available for either 0 - 0.1, 0 - 0.3, or 0 - 1 m depths. There is some potential for bias in the available data due to avoidance of eroded areas, dung and urine spots, etc. In some cases soil carbon data are only available for 0.075 m or 0.15 m soil depths.

Appropriate sampling techniques need to take account of potential changes in bulk density due to animal treading and accumulation of undecomposed root mass (Saggar and Tate 2001. There is a high degree of within-paddock variability due to an uneven excretal distribution by grazing stock (Speir et al 1984).

There is very limited root carbon data available for NZ pastures (Barker et al 1988; Saggar et al 1997, 1999a, Stewart and Metherell 1998, Saggar and Hedley 2001) although roots represent a minor component of total below-ground carbon.

5.8 What process models are available to complement data

Smith et al (1997) assessed the performance of nine models in simulating the dynamics of soil organic C. However, not many of the models they studied predicted the inputs of carbon to the system from the growing plants. They suggested that some form of coupling between the soil carbon models and plant growth models is needed for prediction of carbon sequestration.

In New Zealand, two physiologically based models are available for the estimation of herbage production. Both are extended versions of the mechanistic physiological model of pasture growth first developed at Hurley by Johnson and Thornley (1983, 1985). One model has been further developed specifically to look at the impact of climate change on pasture yield, and is included in the CLIMPACTS climate impact assessment programme. Output from the model includes the amount of carbon removed by harvesting, carbon returned to the soil surface in the form of litter and carbon allocated to roots. Within the CLIMPACTS framework, the carbon inputs to the Rothamsted carbon turnover model are obtained from the pasture-growth model (Warrick et al 2001).

A second model has been developed by Woodward (2001) specifically to look at pasture quality and herbage intake. This model has detailed above-ground processes (reproductive development, litter production, litter decomposition) but does not consider below-ground processes. This omission is currently being addressed by linking the Woodward pasture model to the Century soil model. An empirical, annual time-step model 'OVERSEERTM' provides average estimates of the fate of nutrients N, P, K and S, ignoring year-to-year variability due to climate (Ledgard et al 1999). This model calculates the nutrient outputs and balance by finding the difference between the nutrient inputs and outputs. Other soil models (Daisy, Socrates) are available to predict soil carbon changes but have not yet been coupled to pasture growth models in New Zealand. This is an area that needs future development.

5.9 What regional and national scale inventories are available

Data are available from:

  • Agribase farm surveys: SONZA, MAF surveys;
  • National Soil Carbon Monitoring system (CMS) - soil carbon data for grasslands are the most comprehensive of any land cover type in New Zealand (Davis et al 1999);
  • New Zealand Soils Database (NZSD) (Tate et al 1997);
  • LCDB and VCM databases.

There is only a limited amount of systematic data for the range of unimproved, improved and tussock grassland soils in New Zealand. While NZSD contains the majority of the grassland soils data, land management history is not well documented (e.g., length of pasture, fertiliser history and stock management).

The LCDB and VCM databases have not attempted to estimate pastoral area for 1990. There is a prospect that retrospective use of 1990 Landsat imagery can address this issue.

5.10 What is known of 1990 carbon stocks and current trends

Use of pasture land is being affected by an economically driven shift from sheep to dairying and by reversion of marginal hill country pastures back to shrub or their conversion to planted exotic forestry. However, no significant change in total grassland area is indicated by MAF Statistics (http://www.maf.govt.nz/statistics/primaryindustries/).

The uncertainties in area and carbon stocks are not well quantified. However, Tate et al (1997) (Fig. 5.1) suggest that uncertainties introduced by the use of NZSD would not be large.

Table 5.4. Estimated 1990 Carbon stocks and Recent Trends for Grazing Estimated for 1990

Estimated for 1990

Recent Rates of change per yeara
Area (Mha) Vegetative C stock (Tg C) Soil C stock

(Tg C)

 Total C stock (Tg C) Area (Kha) Veg C Stock (Tg y-1) Soil C stock
(Tg y-1)		 Total C stock (Tg y-1)
13.41 187 1598 1785 -40 -0.262 Probably3 0.0 Possibly4 –2.7 -0.26 ?

a is the average change during 1990 and 1999.
1Including improved and unimproved grasslands, tussock or Danthonia grasslands (SONZA, 1990).
2Directly calculated from data in the table for loss of pasture vegetation following conversion to other land uses from data in the table.
3Tate et al (1997) and Saggar (unpublished data) from long-term comparison of carbon stock at many sites
4Lambert et al (2000) from long-term data at AgResearch Ballantrae Research Station.

5.11 Projections for the first commitment period where feasible

Data from Lambert et al (2000) suggest that soil carbon in the 0-0.075 m depth may be decreasing by 200 kg C ha-1 y-1 in some NZ hill-country pastures. These data do not identify whether the carbon loss results from diffusion below 0.075 m depth, soil erosion, or respiration. If it is assumed that the loss process is to the atmosphere and that similar losses occur generally under intensive grazing conditions then New Zealand grasslands may lose up to 2.7 TgC y-1.

Results of Tate et al (1997) indicate a 1:1 relationship exists between soils sampled recently and 30-50 years ago, and suggest that New Zealand's established pastoral soils are at steady-state. Recent studies (Surinder Saggar, unpublished data) also show no change in soil carbon levels after 6 years of continuous fertilization. Further evidence for the general stability of soil carbon under pasture comes from Carran and Theobald (2000b) who show no change in soil carbon over 23 years in a pasture under intensive management.

5.12 Data and research needed to reduce uncertainties

Inconsistencies in estimates of the area under each land use will need to be resolved. This can be aided by comparison of statistical data sets and re-examination of modelling approaches. While preliminary estimates have been made of national scale NPP and soil respiration (Tate et al 2000b), more work needs to be done to make these robust.

Techniques need to be developed to scale site-specific above- and below-ground carbon dynamics to regional and national levels (Tate et al 1997). A promising approach is through combined use of remote sensing and process-based models.

Better quantification is required of carbon losses from grazing land by erosion and studies of carbon sequestration options on degraded and eroded land. Erosion is considered further in section 7.4.

More work is required on validation of steady-state soil carbon levels in established grazing land. There is a scarcity of data on long-term changes in soil organic carbon in pasture soil profiles under different management practices that could be used to test steady state soil carbon levels. Changes in soil carbon need to be based on a better knowledge of steady state conditions for different land-use, land-cover and land management options. Determination of below-ground carbon inputs in pastures is needed to verify predictions from coupled pasture growth models and soil organic matter models.

Pasture responses to elevated CO2 introduce another level of uncertainty. In the short-term, elevated CO2 may not necessarily lead to greater carbon sequestration by New Zealand temperate pastures, but it will result in greater carbon turnover (Ross et al 1996). However, based on effects on grassland over many decades, elevated CO2 could lead to increased soil carbon levels (Ross et al 2000b)

As noted in the previous section, management of rich organic and peat based soils raises some specific issues that require further study.

Study of New Zealand's nitrous oxide emissions from (principally) grazing land is a related aspect of land-use and pasture management, from both a scientific and policy perspective. NzOnet was established as a collaborative network of researchers in 1999 at MAF's invitation (Cameron et al 2000) to focus work on nitrous oxide. A research plan has been developed to reduce uncertainty in New Zealand's nitrous oxide emission rates by measuring emission factors for seasons and regions where data do not exist, and developing internationally defensible, process-based models, that are developed and refined from emergent data-sets. This work should be linked to models of the carbon budget in pastures and has the potential to develop an integrated view of pasture - greenhouse gas issues.

Previous PageTable Of ContentsNext Page

Contact for Enquiries

MAF Information Services
Pastoral House
25 The Terrace
PO Box 2526
Wellington, NEW ZEALAND

Fax: +64 4 894 0721
Contact this person

 
Print this page Print this page