**5. Measure of natural habitat provision**

Measures of habitat provision need to account for different types of habitat and their associated biodiversity. Dymond et al. (2008) showed how proportions of unique habitat remaining may be combined to give a national measure of habitat provision. The habitat measure is based on the contribution it makes to the New Zealand Government goal of maintaining and restoring a full range of remaining natural habitats to a healthy and functioning state. For measuring indigenous forest and grasslands, the historical unique habitats come from Land Environments New Zealand (LENZ) (Leathwick et al., 2003). Wetlands are at the interface of terrestrial and freshwater habitats, and therefore another habitat framework representing both aquatic and terrestrial biota (Leathwick et al., 2007) is used. As such, the measure of habitat provision for wetlands is applied separately from the indigenous forests and grasslands measure.

#### **5.1 Indigenous forest and grasslands**

We used LENZ at level II (suitable for national to regional scale) and the most recent land cover (2008) to characterise historic and present habitats. The measure of habitat provision for a land environment is defined as:

$$H\_i = P\_i \left(\frac{a\_i}{A\_i}\right)^{0.5} \tag{1}$$

where

210 Biodiversity Loss in a Changing Planet

Pakihi Bog Swamp Marsh Fen Gumland Seepage Inland

regional analyses suggest that wetland extent continues to decline, although at a slower rate, as land drainage and agricultural development continue (Grove, 2010; Newsome & Heke, 2010). Wetland mapping is a challenging task as wetlands are sometimes too small in area to be identified using common satellite resolution. Their extent can vary seasonally (e.g., dryness, wetness) and therefore can change markedly at the time of imagery acquisition. While satellite images are useful for providing information at national scale, automatic classification is not possible as vegetation types in wetlands are so variable, making them difficult to characterise through spectral signature. Thus wetlands have been mapped on a manual or semi-automated basis (Ausseil et al., 2007), and this requires a significant amount

Measures of habitat provision need to account for different types of habitat and their associated biodiversity. Dymond et al. (2008) showed how proportions of unique habitat remaining may be combined to give a national measure of habitat provision. The habitat measure is based on the contribution it makes to the New Zealand Government goal of maintaining and restoring a full range of remaining natural habitats to a healthy and functioning state. For measuring indigenous forest and grasslands, the historical unique habitats come from Land Environments New Zealand (LENZ) (Leathwick et al., 2003). Wetlands are at the interface of terrestrial and freshwater habitats, and therefore another habitat framework representing both aquatic and terrestrial biota (Leathwick et al., 2007) is used. As such, the measure of habitat provision for wetlands is applied separately from the

Saline

Historic extent

Current extent

0

of effort for all of New Zealand.

**5. Measure of natural habitat provision** 

indigenous forests and grasslands measure.

Fig. 5. Current and historic extent of wetland per class.

200

400

600

800

1,000

Area (x 1,000 ha)

1,200

1,400

1,600

1,800

*ai* is the area of natural habitat remaining in land environment *i,* 

*Ai* the area of land environment *i*, and

*Pi* is the biodiversity value of the *i*th land environment when fully natural.

The 0.5 power index is used to produce a function monotonically increasing from zero to one with a decreasing derivative in order to represent the higher biodiversity value of rare habitat. In the absence of comprehensive and detailed biodiversity information, Dymond et al. (2008) suggested using species-area relationships (Connor & McCoy, 2001) to estimate *Pi* as the land environment area to the power of 0.4. The varying condition, or degree of naturalness, of individual sites also needs to be taken into account in the habitat measure:

$$\mathbf{H}\_i = \mathbf{P}\_i \left( \frac{\sum\_{j=1}^n c\_{ij} b\_{ij}}{A\_i} \right)^{0.5} \tag{2}$$

where

*cij* is the condition,

*bij* is the area of of the *j*th habitat site in the *i*th land environment, and

*n* is the number of habitats in the *i*th land environment.

The condition of indigenous forest, subalpine shrublands, alpine habitats, and tussock grasslands above the treeline, are assumed to have a condition of 1.0. Tussock grasslands below the treeline have a condition of 0.8 and indigenous shrublands have a condition of 0.5. All other landcovers have a condition of 0.0.

Figure 6 shows the input layers (current land cover and land environments at level II) and the resulting habitat provision map. This map shows the contribution per hectare to the

habitat measure (i.e. each pixel represents 1 *ij ij n i ij ij <sup>j</sup> c b H c b* ).

#### **5.2 Freshwater wetlands**

Wetlands are at the interface between water and terrestrial dry environments. They have been categorised with freshwater environments in the past, and as such require a different definition of biogeographic units than the terrestrial environments. We replaced land environments data with biogeographic units defined by climatic and river basin characteristics (Leathwick et al., 2007). This framework was used to define priority conservation for rivers (Chadderton et al. 2004) and wetlands (Ausseil et al., 2011). A condition index for wetlands, similar to *ci* in equation (1), was calculated for all current wetland sites in New Zealand (Ausseil et al., 2011). This condition index reflects the major

Provision of Natural Habitat for Biodiversity: Quantifying Recent Trends in New Zealand 213

anthropogenic pressures on wetlands, including nutrient leaching, introduced species,

The measure of habitat provision for wetlands in a biogeographic unit now needs to account

1

*c b*

*n <sup>m</sup> ijk ijk <sup>j</sup> ik <sup>k</sup> ik*

Wetland habitats are defined at the class level (*m*=8 classes) using the wetland classification of Johnson & Gerbeaux (2004). As with *Pi* in equation (2), *Pik* is defined as the historical area

total area of class *k* wetlands in biogeographic unit *i*, weighted by the condition index *cijk* for each wetland site. If all the wetlands were in pristine condition, the sum would equal the

Figure 7 shows wetland habitat provision for each biogeographic unit. The colours represent

Though there are still large areas of natural habitat remaining in New Zealand, there continues to be ongoing loss. Prior to the settlement of humans, there were 23 million hectares of indigenous forest. Today, only 6.5 million hectares of indigenous forest are remaining. While the total area remaining is large, little of that is in lowland forest ecosystems, and over the last 20 years more lowland ecosystems have been lost. Despite continuing losses in lowland ecosystems, the net area of indigenous forest may well be increasing due to regeneration of indigenous shrublands in marginal hill country. Indigenous grasslands have a similar pattern of change. Over the last 170 years, 4.7 million hectares of indigenous grasslands have been lost. Though the total area of remaining grasslands is large, little of that is in lowland ecosystems, and over the last 20 years more lowland ecosystems have been lost. Wetlands are the most severely impacted ecosystems. Of the 2.4 million hectares of wetlands existing pre-Maori, only 250 thousand hectares are remaining – that is, only 10% of what was there originally. Again, lowland wetlands are mostly affected, with a higher proportion of swamps lost. Recent trend analyses shown in this chapter reveal that loss is still continuing, and is a precursor to negative impacts on

The habitat provision map for indigenous forest and grasslands show large spatial variability. High values are usually associated with rarer habitats in good condition, but also with habitats in very small land environments. For wetlands, the habitat provision map is

1

*<sup>P</sup> <sup>A</sup> <sup>W</sup>* 

0.5

(3)

*n ijk ijk <sup>j</sup> c b*

reflects the

imperviousness, loss of naturalness, woody weeds, and drainage pressure.

*i*

*cijk* is the condition index of wetland site *j* in class *k* in biogeographic unit *i,*

per wetland class per biogeographic unit to the power of 0.4. The sum 1

provision of ecosystem services and subsequent human well-being.

*bijk* is the area of wetland site *j* in class *k* in biogeographic unit *i,*

*n* is the number of class *k* wetland sites in biogeographic unit *i*.

*Aik* is the historic area of class *k* in biogeographic unit *i,*

*m* is the number of wetland classes, and

total areal extent in that class.

the value *Wi* from equation (3).

**6. Discussion** 

for different wetland classes, so is defined as

where

Fig. 6. Habitat provision per hectare from forests and grasslands.

anthropogenic pressures on wetlands, including nutrient leaching, introduced species, imperviousness, loss of naturalness, woody weeds, and drainage pressure.

The measure of habitat provision for wetlands in a biogeographic unit now needs to account for different wetland classes, so is defined as

$$\mathcal{W}\_i = \sum\_{k=1}^{m} P\_{ik} \left( \frac{\sum\_{j=1}^{n} c\_{ijk} b\_{ijk}}{A\_{ik}} \right)^{0.5} \tag{3}$$

where

212 Biodiversity Loss in a Changing Planet

Fig. 6. Habitat provision per hectare from forests and grasslands.

*cijk* is the condition index of wetland site *j* in class *k* in biogeographic unit *i,*

*bijk* is the area of wetland site *j* in class *k* in biogeographic unit *i,*

*Aik* is the historic area of class *k* in biogeographic unit *i,*

*m* is the number of wetland classes, and

*n* is the number of class *k* wetland sites in biogeographic unit *i*.

Wetland habitats are defined at the class level (*m*=8 classes) using the wetland classification of Johnson & Gerbeaux (2004). As with *Pi* in equation (2), *Pik* is defined as the historical area

per wetland class per biogeographic unit to the power of 0.4. The sum 1 *n ijk ijk <sup>j</sup> c b* reflects the

total area of class *k* wetlands in biogeographic unit *i*, weighted by the condition index *cijk* for each wetland site. If all the wetlands were in pristine condition, the sum would equal the total areal extent in that class.

Figure 7 shows wetland habitat provision for each biogeographic unit. The colours represent the value *Wi* from equation (3).

## **6. Discussion**

Though there are still large areas of natural habitat remaining in New Zealand, there continues to be ongoing loss. Prior to the settlement of humans, there were 23 million hectares of indigenous forest. Today, only 6.5 million hectares of indigenous forest are remaining. While the total area remaining is large, little of that is in lowland forest ecosystems, and over the last 20 years more lowland ecosystems have been lost. Despite continuing losses in lowland ecosystems, the net area of indigenous forest may well be increasing due to regeneration of indigenous shrublands in marginal hill country. Indigenous grasslands have a similar pattern of change. Over the last 170 years, 4.7 million hectares of indigenous grasslands have been lost. Though the total area of remaining grasslands is large, little of that is in lowland ecosystems, and over the last 20 years more lowland ecosystems have been lost. Wetlands are the most severely impacted ecosystems. Of the 2.4 million hectares of wetlands existing pre-Maori, only 250 thousand hectares are remaining – that is, only 10% of what was there originally. Again, lowland wetlands are mostly affected, with a higher proportion of swamps lost. Recent trend analyses shown in this chapter reveal that loss is still continuing, and is a precursor to negative impacts on provision of ecosystem services and subsequent human well-being.

The habitat provision map for indigenous forest and grasslands show large spatial variability. High values are usually associated with rarer habitats in good condition, but also with habitats in very small land environments. For wetlands, the habitat provision map is

Provision of Natural Habitat for Biodiversity: Quantifying Recent Trends in New Zealand 215

covenant is registered against the title of the land in perpetuity and there are obligations to manage the land in accordance with the covenant document. Over 70,000 ha are now protected by QEII covenants (Ministry for the Environment, 2007). Nga Whenua Rahui is a contestable fund to negotiate the voluntary protection of native forest on Maori-owned land. Legal protection is offered through covenants, setting aside areas as Maori reservations or through management agreements. About 150,000 ha of native ecosystems are now protected under this fund. The Nature Heritage Fund (NHF) is a third contestable fund for voluntary protection of nature on private land. Its aim is to add to public conservation land those ecosystems important for indigenous biodiversity that are not represented within the existing protected area network. Since 1990, the fund has protected over 100,000 hectares of indigenous ecosystems through direct land purchases, covenants on private land or fencing. The information on habitat provision could feed into the Department of Conservation (DOC) management system. DOC is responsible for managing biodiversity on the conservation estate, and is developing the Natural Heritage Management System (NHMS). DOC's statement of intent is to legally protect the best possible examples of each native ecosystem type, by fencing, reinstating water levels, replanting, controlling pest animals and weeds, and reintroducing native species to restore and maintain natural ecosystems. The framework proposed in this chapter is envisaged to help achieve this goal through accurate information on habitat extent and ecosystem loss, and provides a measure for comparing habitats within and across land environments where species level assessments may not be

Continuing loss of natural habitat may be due to a lack of market prices for associated ecosystem services (TEEB, 2010). Monetisation of habitat provision could partly redress this. In New Zealand, Patterson & Cole (1999) estimated natural forests and wetlands to both have total economic values of approximately 6 billion dollars in 1994. From this, it is possible to convert the units of habitat measure to economic value in dollars per year – Dymond et al. (2007) estimated this as 60 units to one (NZ) dollar per year (assuming areas are in hectares). This monetisation would permit the comparison of changes in habitat alongside changes in other ecosystem services in the same units. This reduces the complexity of results when analysing impacts of different land-use decisions. Using dollars also provides context for stakeholders unfamiliar with biodiversity impacts associated with habitat loss. The negative side of monetisation is that some stakeholders may be encouraged to make trade-offs on the basis of the monetisation alone, not realising the assumptions and limitations involved, or being aware of environmental bottom-lines. Indeed, the risk of valuation is to get the figure very wrong. There are numerous valuation methods, often based on subjective, hypothetical, and questionable assumptions, which can all give vastly different values (Spangenberg & Settele, 2010). Altogether, monetisation, although easy to comprehend, can be misleading and should be used with caution. It should be used in close consultation with decision-makers, so that they are fully aware of the pitfalls and

The measure of habitat provision is a landscape approach which makes several assumptions. First, it uses particular GIS databases, each of which has a certain level of sensitivity and accuracy. Land environments has been tailored to forest ecosystems, and does not encompass the full breadth of other ecosystem types. Biogeographic units were used for wetlands, assuming that freshwater species would be concentrated within defined hydrological boundaries. Second, it assumes that landscape morphology and pattern can be used as a surrogate for species. Though this overcomes the issues surrounding availability

assumptions behind the valuation, to avoid misallocation of resources.

possible.

Fig. 7. Habitat provision for freshwater wetlands in each biogeographic unit.

shown at the biogeographic unit level, mainly because wetland boundaries are difficult to depict at the scale shown here. The contribution to the national habitat measure comes mostly from biogeographic units with minimal conversion to productive land. Low values represent units where wetland areas have depleted or where wetlands have been degraded. This information can be used by decision-makers to prioritise the allocation of conservation funds. For example, the maps can be intersected with legally protected areas, like those from Walker et al. (2008), which target areas under private ownership with high natural values. Several legislative tools can be used to protect remnant habitats, including the establishment of conservation covenants like the Queen Elizabeth the Second National Trust (QEII), Nga Whenua Rahui, and the National Heritage Fund.

QEII's goal is to help New Zealand farmers protect open space on private land for the benefit and enjoyment of the present and future generations of New Zealanders. The

Fig. 7. Habitat provision for freshwater wetlands in each biogeographic unit.

Whenua Rahui, and the National Heritage Fund.

shown at the biogeographic unit level, mainly because wetland boundaries are difficult to depict at the scale shown here. The contribution to the national habitat measure comes mostly from biogeographic units with minimal conversion to productive land. Low values represent units where wetland areas have depleted or where wetlands have been degraded. This information can be used by decision-makers to prioritise the allocation of conservation funds. For example, the maps can be intersected with legally protected areas, like those from Walker et al. (2008), which target areas under private ownership with high natural values. Several legislative tools can be used to protect remnant habitats, including the establishment of conservation covenants like the Queen Elizabeth the Second National Trust (QEII), Nga

QEII's goal is to help New Zealand farmers protect open space on private land for the benefit and enjoyment of the present and future generations of New Zealanders. The covenant is registered against the title of the land in perpetuity and there are obligations to manage the land in accordance with the covenant document. Over 70,000 ha are now protected by QEII covenants (Ministry for the Environment, 2007). Nga Whenua Rahui is a contestable fund to negotiate the voluntary protection of native forest on Maori-owned land. Legal protection is offered through covenants, setting aside areas as Maori reservations or through management agreements. About 150,000 ha of native ecosystems are now protected under this fund. The Nature Heritage Fund (NHF) is a third contestable fund for voluntary protection of nature on private land. Its aim is to add to public conservation land those ecosystems important for indigenous biodiversity that are not represented within the existing protected area network. Since 1990, the fund has protected over 100,000 hectares of indigenous ecosystems through direct land purchases, covenants on private land or fencing. The information on habitat provision could feed into the Department of Conservation (DOC) management system. DOC is responsible for managing biodiversity on the conservation estate, and is developing the Natural Heritage Management System (NHMS). DOC's statement of intent is to legally protect the best possible examples of each native ecosystem type, by fencing, reinstating water levels, replanting, controlling pest animals and weeds, and reintroducing native species to restore and maintain natural ecosystems. The framework proposed in this chapter is envisaged to help achieve this goal through accurate information on habitat extent and ecosystem loss, and provides a measure for comparing

habitats within and across land environments where species level assessments may not be

possible. Continuing loss of natural habitat may be due to a lack of market prices for associated ecosystem services (TEEB, 2010). Monetisation of habitat provision could partly redress this. In New Zealand, Patterson & Cole (1999) estimated natural forests and wetlands to both have total economic values of approximately 6 billion dollars in 1994. From this, it is possible to convert the units of habitat measure to economic value in dollars per year – Dymond et al. (2007) estimated this as 60 units to one (NZ) dollar per year (assuming areas are in hectares). This monetisation would permit the comparison of changes in habitat alongside changes in other ecosystem services in the same units. This reduces the complexity of results when analysing impacts of different land-use decisions. Using dollars also provides context for stakeholders unfamiliar with biodiversity impacts associated with habitat loss. The negative side of monetisation is that some stakeholders may be encouraged to make trade-offs on the basis of the monetisation alone, not realising the assumptions and limitations involved, or being aware of environmental bottom-lines. Indeed, the risk of valuation is to get the figure very wrong. There are numerous valuation methods, often based on subjective, hypothetical, and questionable assumptions, which can all give vastly different values (Spangenberg & Settele, 2010). Altogether, monetisation, although easy to comprehend, can be misleading and should be used with caution. It should be used in close consultation with decision-makers, so that they are fully aware of the pitfalls and assumptions behind the valuation, to avoid misallocation of resources.

The measure of habitat provision is a landscape approach which makes several assumptions. First, it uses particular GIS databases, each of which has a certain level of sensitivity and accuracy. Land environments has been tailored to forest ecosystems, and does not encompass the full breadth of other ecosystem types. Biogeographic units were used for wetlands, assuming that freshwater species would be concentrated within defined hydrological boundaries. Second, it assumes that landscape morphology and pattern can be used as a surrogate for species. Though this overcomes the issues surrounding availability

Provision of Natural Habitat for Biodiversity: Quantifying Recent Trends in New Zealand 217

to thank Fiona Carswell and Garth Harmsworth from Landcare Research for their valuable

Ashdown, M and Lucas, D (1987). *Tussock grasslands landscape values and vulnerability*. New

Ausseil, A-GE, Chadderton, WL, Gerbeaux, P, Theo Stephens, RT & Leathwick, JR (2011).

Brower, A (2008). *Who owns the high country? The controversial story of tenure review in New Zealand*. Craig Potton Publishers, ISBN 9781877333781, Christchurch.

Chadderton, L, Brown, D & Stephens, T (2004). *Identifying freshwater ecosystems of national* 

Conservation International. (2010). *Biodiversity hotspots*. Retrieved 21 December, 2010, available from http://www.conservation.org/Documents/cihotspotmap.pdf. de Lange, PJ, Norton, DA, Courtney, S, Heenan, PB, Courtney, S, Molloy, BPJ, Ogle, CC,

plants of New Zealand. *New Zealand Journal of Botany*, Vol 42, pp. 45-76. Diaz, S, Fargione, J, Chapin, FS, III & Tilman, D 2006. Biodiversity loss threatens human well-being. *PLoS Biol*, Vol 4, No 8, e277. doi:10.1371/journal.pbio.0040277. Dickinson, KJM, Mark, AF, Barratt, BIP & Patrick, BH (1998). Rapid ecological survey,

Zealand. *Journal of The Royal Society of New Zealand,* Vol 28, pp. 83-156. Department of Conservation & Ministry for the Environment (2000). *The New Zealand* 

Dymond, JR, Ausseil, A-GE, Shepherd, JD & Janssen, H (2007). A landscape approach for

Dymond, JR, Ausseil, A-GE & Overton, JM (2008). A landscape approach for estimating the

Ewers, RM, Kliskey, AD, Walker, S, Rutledge, D, Harding, JS & Didham, RK (2006). Past and

http://biodiversity.govt.nz/picture/doing/nzbs/contents.html.

Zealand. *Ecological Economics,* Vol 66, No 2-3, pp. 275-281.

*rivers*. Department of Conservation Discussion Document, Wellington. Cieraad, E (2008). *How much indigenous biodiversity remains in land under indigenous cover?* Unpublished contract report LC0708/145, Landcare Research Ltd, Lincoln. Connor, EF & McCoy, ED (2001). Species-area relationship. In: *Encyclopaedia of Biodiversity*.

Applying systematic conservation planning principles to palustrine and inland saline wetlands of New Zealand. *Freshwater Biology* Vol. 56, No 1, pp. 142-161. Ausseil, A-GE, Dymond, JR & Shepherd, JD (2007). Rapid mapping and prioritisation of

wetland sites in the Manawatu-Wanganui region, New Zealand. *Environmental* 

*importance for biodiversity: Criteria methods, and candidate list of nationally important* 

Rance, BD, Johnson, PN & Hitchmough, RA (2004). Threatened and uncommon

inventory and implementation: A case study from Waikaia ecological region, New

assessing the biodiversity value of indigenous forest remnants: Case study of the Manawatu/Wanganui region of New Zealand. *Ecological Economics,* Vol 64, No 1,

conservation value of sites and site-based projects, with examples from New

future trajectories of forest loss in New Zealand. *Biological Conservation*, Vol 133, No

Zealand Environmental Council, ISBN 0477058272, Wellington.

comments on an earlier draft and Anne Austin for internal editing.

*Management,* Vol. 39, No 3, pp. 316-325.

CBD (1992). *Convention on Biological Diversity*. Rio de Janeiro, UNCED.

Academic Press, San Diego. Vol 5, pp 397-441.

*biodiversity strategy*. from

pp. 82-91.

3, pp. 312-325.

**8. References** 

of data, application is limited at the various levels and components of biodiversity. In other words, provisions can not be assessed at multiple scales (i.e. habit, community and/or species). Third, the condition of indigenous forest and grassland assumes that all sites are characterised by one condition, though condition could vary within large sites. For wetlands, the condition does vary per site, but it is based on landscape indicators. It is appropriate for a rapid assessment of sites, and can help for prioritising field visits (Ausseil et al., 2007), but does not necessarily reflect the true condition in the field.

The loss of indigenous forest is well characterised by the habitat provision analysis, but the gain of indigenous forest from regenerating indigenous shrublands is not. This is because both the LCDB and the LUM datasets focus on mapping change primarily between woody and herbaceous vegetation, and the subtle changes in the spectral signature of regenerating indigenous forest and mature forest are not accurately characterised or determined, making it difficult to decide whether indigenous vegetation is mature enough to be classified as forest. This is important because there are large areas of indigenous shrublands in New Zealand, approximately 1.6 million hectares. Much of these shrublands are currently regenerating to forest and could make a significant contribution to the areal extent of indigenous forest if this trend continues. If we assume a conservative time of 100 years to reach forest maturity and a uniform distribution of shrubland age, then we would expect about 1% of the shrubland area to change to indigenous forest each year – this amounts to 16 thousand hectares per year. Over 18 years this would equate to approximately 300 thousand hectares, which is six times the estimated current loss of indigenous forest. This fills an important information gap in our understanding of the changing areal extent of indigenous forest and indicates the importance of using objective mapping techniques to monitor change.

Conservation management in New Zealand is becoming increasingly strategic, systematic, and reliant on accurate information on which to plan and prioritise goals and actions. A range of sophisticated tools and approaches have been developed to support these efforts in the past ten years. These include measuring Conservation Achievement (Stephens et al., 2002), the Land Environments of New Zealand (Leathwick et al., 2003), and measuring provision of natural habitat (Dymond et al., 2008). In addition, these efforts have spawned considerable activity for acquiring underlying data, such as biodiversity value (Cieraad, 2008), land cover (the LCDB3 project), and threats to biodiversity (Overton et al., 2003; Walker et al., 2006). However, a national coordinated approach to conservation management taking into account species distributions is required. Overton et al. (2010) are developing a tool called Vital Sites to assess ecological integrity. This incorporates current and natural distributions of native species based on a modeling approach, pressures (e.g., pests or habitat loss) on biodiversity, and the effects of management on relieving pressures. It operates at two levels (species and landscape) and assessments of significance and priorities can be made at each separate level or by combining the two levels. This research tool will provide another step to helping achieve goals towards identifying the most vulnerable ecosystems in New Zealand requiring urgent protection and management.

#### **7. Acknowledgment**

This work was supported by the New Zealand Foundation for Research, Science, and Technology through Contract C09X0912 "An ecosystem services approach to optimise natural resource use for multiple outcomes" to Landcare Research. The authors would like to thank Fiona Carswell and Garth Harmsworth from Landcare Research for their valuable comments on an earlier draft and Anne Austin for internal editing.
