A state-owned enterprise, Timberlands West Coast Ltd. (TWC), proposed to sustainably manage beech forests on 0.1 million hectares of New Zealand's 2.8 million hectare beech forest estate. The scheme was to be reviewed by a Resource Management Act (RMA) commission hearing to establish whether or not it was sustainable in terms of the RMA, a scrutiny that would have been pivotal for the granting of a resource consent.

The New Zealand Labour Party adopted a policy against the scheme during September 1999, just weeks prior to gaining power (in coalition with the Alliance party) in a general election. Mr Jim Sutton, Labour party spokesman on forestry, immediately resigned his spokesmanship, and will no longer correspond on forestry matters.

One of the first acts of our new Labour/Alliance government (elected in November 1999) was to make TWC withdraw its application for consent for the beech scheme. This was achieved by removing sustainable beech management from TWC's mission statement. The action delighted some environmental groups, but dismayed other environmental groups and resource management professionals.

The Department of Conservation (DOC) prepared draft documents for presentation to the Commission, but the withdrawal of TWC from the hearing resulted in these documents not being presented. They have been obtained from DOC under the provisions of the Official Information Act, and have been re-typed and presented in these pages. Please note that the draft documents contained what appear to be typographical errors; they have been reproduced as obtained from DOC.

Evidence of Colin O'Donnell





1.1     My name is Colin Francis John O'Donnell, and I am employed as a scientist by the Science and Research Unit, Department of Conservation, 133 Victoria Street. Christchurch. I was awarded a BSc (Hons( in Zoology from Canterbury University (1980) and a PhD in Zoology from Otago University (1999). I have been employed as a conservation biologist by the NZ Wildlife Service between 1980 and 1987, and by the Department of Conservation between 1987 and the present.

1.2     My current position is Programme Leader for Animals Research for the Department of Conservation. My responsibilities include developing research strategies for threatened invertebrates, amphibians, lizards, freshwater fish, birds and bats, evaluating annual research needs, and prioritising needs with the aid of an advisory group.

1.3     I have been involved in studies of forest birds since 1978 when I worked for the NZ Forest Service on the West Coast, measuring bird densities in forests logged at different times. Between 1979 and 1982 I undertook various contracts for the Fauna Survey Unit of the NZ Wildlife Service, assisting with inventories of sites of national significance for wildlife throughout New Zealand and including West Coast and Buller District forests. From 1983 onwards I was responsible for forest bird research in the NZ Wildlife Service and particularly a study aimed at predicting the impacts of proposed future selection logging systems on forest birds. This work has continued during my employment with the Department of Conservation. Since 1990 I have been working in forest areas in Southland, Canterbury, West Coast and central North Island on a range of conservation problems related to threatened forest wildlife (yellowheads, parakeets, kaka, long-tailed bats, short-tailed bats). I have published over 40 refereed scientific papers and over 20 management reports.

1.4     I have had considerable experience in rating areas for their regional, national and international significance for wildlife. This experience includes rating Sites of Special Wildlife Significance and Wetlands of International Importance for the NZ Wildlife Service between 1980 and 1987, my membership of two International Union for the Conservation of Nature specialist groups for threatened species, and of Department of Conservation groups that prioritise research on theatened species and predators and advising on sites for national habitat restoration projects.

1.5     My specialist expertise is in the area of birds and mammals. Hence in this evidence I define "forest wildlife" as species of birds and mammals indigenous to New Zealand forests, and specifically to the Timberlands Application Area. There are other forest fauna present in the area that should be considered including lizards, invertebrates (for example land snails), and freshwater fish. Some of these species are distinctive to the region, however, I will not comment on these further here.


In this evidence I show that:

(a)     Forests of the Timberlands forests covered by these Consent Applications contain nationally significant populations of indigenous wildlife and particularly rare and threatened species. These species are Absolutely Protected by the Wildlife Act 1953. Reports commissioned by Timberlands record areas of "Outstanding habitat" through much of the unlogged forests covered by these Consent Applications.

(b)     I contend that in the context of Section 6 (c) of the Resource Management Act (1991) and related provisions in the Buller and Tasman District Plans the Consent Applications must show that the proposed logging regime will sustain wildlife within the forests proposed for logging by Timberlands.

(c)     Sustainability is defined in terms of ensuring the persistence of sufficient amounts of foraging, roosting and breeding sites for wildlife to maintain viable populations. Maintenance of foraging habitat would be meaningless without protection of a viable number of preferred nesting, roosting or breeding sites - and vice versa.

(d)     I am concerned that wildlife populations in the Timberlands Application Area would be reduced further by the logging proposed by the Consent Applications. The proposed logging does not appear to sustain old-age, large diameter "wildlife trees". These are needed to retain populations of sensitive hole-nesting birds and bats, particularly the threatened species present. Proposed harvesting rates of old-age trees need to be considerably more conservative if these wildlife populations are to be retained.

(e)     Tree stem size harvesting limits proposed by Timberlands appear to be too high to protect a significant proportion of wildlife trees at least in the short term (<100 years). Recruitment of new trees is a long-term process, and extraction rates need to match replacement rates.

(f)     Specific threats to long-tailed bats, kaka and parakeets are not being mitigated adequately in the Consent Applications because of a disproportionate negative affect of the logging of the old-age wildlife trees, at least in the short term.

(g)     Predicted impacts on wildlife need to be modelled before the Consent Applications are approved because of the high risk to threatened wildlife.

(h)     I complete the evidence with a set of recommendations aimed at resolving these problems. A precautionary approach to managing indigenous forests is well justified given the information presented in this evidence. I recommend that the Consent Applications not be approved at this time because the logging proposed seems unlikely to sustain wildlife populations. Revised Consent Applications would need to be considerably more conservative in terms of extraction rates for wildlife trees and there need to be strict, unambiguous rules to mitigate predicted threats. Proposed mitigation for wildlife is inadequate at the present time.


Indigenous forests within the Timberlands Application area can be termed as significant habitats for indigenous fauna under criteria widely used to implement Section 6(c) of the Resource Management Act 1991. In the context of the Act these forests contain representative fauna communities, rare and distinctive species, and wildlife are important in maintaining ecological processes. I will give examples below.

It has been suggested that the forests within the areas covered by these Applications cannot be nationally significant, otherwise they would have been protected by the West Coast Forest Accord (See Joint Staff Report to the Hearing, Page 34, Section 2.3.13) This is NOT correct, largely because detailed forest bird surveys for some of the unlogged forests within the Application areas were not available at that time. In addition, techniques for finding bats were not available in New Zealand at the time, and have only been available in the 1990's. Timberlands West Coast have redressed that balance somewhat in recent years by conducting intensive surveys of some forests not previously assessed in terms of their national significance for wildlife. The findings of these surveys are included in several comprehensive reports produced by Timberlands. These are cited in Appendix 1, and I refer you to Buckingham (1996, 1997, 1999), Buckingham & Brown (1996), and Buckingham & Nilsson (1994).

The authors of these reports quantify forest bird and bat conspicuousness in the forests using standard techniques that allow comparison with some other forest surveys in New Zealand. By examining the frequency of species of wildlife on standard counts it is possible to evaluate the relative significance of wildlife populations.

The Timberlands reports make numerous references to specific bird populations being among the best recorded in the country, and note habitat values that are "Outstanding" nationally. I examine some of the key national values below, but refer you to the Results and Discussion sections of the Timberlands Reports and other comparative references cited in Appendix 1 for considerably more detail.

3.1       Representative wildlife communities within the Timberlands Application Areas
3.1.1    The indigenous forests within the Timberlands Application Area are biologically diverse and support bird communities representative of forests dominated by podocarps (e.g., kahikatea, rimu, mountain totara), beech (red, hard, mountain and silver beech), and broadleaved evergreen trees (e.g., southern rata, kamahi, tawheowheo). The unlogged forests support excellent examples of the biological diversity found in these forest types nationally.

3.1.2     Significant populations of 24 forest bird species and two bat species occur (Table 1).

3.1.3     All species are Absolutely Protected Wildlife under the Wildlife Act 1953.

3.1.4     I identify 3 categories of indigenous forest wildlife that occur in indigenous forests within the Timberlands Management Area (Table 1). Eight species, such as the kaka, that are confined to indigenous forest throughout the year, or for a crucial part of it, for roosting, breeding, feeding and shelter are defined as "obligate forest dwellers". Species that are primarily forest dwellers (9 species, for example New Zealand pigeon) breed and spend most of their annual cycle within indigenous forests, but are less dependent on forests because they may occur in shrublands and modified forest. Populations of 9 further species, such as the grey warbler, are common in a range of habitats, including indigenous forests, and viable populations can survive in alterative habitats. These are termed "facultative forest dwellers".

3.1.5     It is the obligate, and to a lesser extent primary, forest dwellers that are most sensitive to changes in habitat, and all are threatened with extinction in the medium term if factors causing their decline are not halted. Key factors in their decline are degradation of optimum habitat, and predation and competition from introduced animals.

3.1.6     There is a large amount of resource material available characterising the fauna of the Buller and Tasman Districts. A reference list to important publications is provided in Appendix 1. This information base includes extensive surveys run by the Ornithological Society of New Zealand, the New Zealand Wildlife Service, Ecology Division of Department of Scientific and Industrial Research, the Forest Research Institute, the Department of Conservation, and Timberlands West Coast Limited.

3.1.7     Information in the references given in Appendix 1 indicates that significant habitats for wildlife occur in crown indigenous production forests within the Timberlands Application Area, as well as on nearby land set aside for conservation.

3.1.8     Wildlife surveys by Rhys Buckingham and others for Timberlands West Coast indicated that all forest species present in North Westland, including most threatened species, were present within forests designated for production. Buckingham found that bird numbers were amongst the best examples forest wildlife communities he reviewed. The wildlife reports note a large number of areas that were "Outstanding" for wildlife using definitions used by the former Wildlife Service for ranking nationally significant habitats. This was particularly the case for unlogged forests throughout the Maruia Working Circle and some of the Inangahua forests (for example, Orikaka, Ohikanui, Pells Creek, Larrys and Camp Creek) and parts of the Grey Working Circle, particularly Te Wharau and Paparoa forests. These forests provide the food, shelter and nesting habitat required by the obligate and primary forest dwellers mentioned in Table 1.

3.1.9     The lowland forests designated as production forests are particularly important for the maintenance of bird populations throughout the year. Research by myself and others in North and South Westland (see Appendix 1) shows that birds move seasonally to different food sources while others move to lower altitudes simply to avoid harsh winter climates. Specific food sources are usually only available in certain seasons. For example, kaka move from high and mid-altitude food sources (e.g. beech invertebrates) to low-altitude food sources such as fuchsia flowers in the early spring, mistletoe flowers in mid-summer, rimu fruit in autumn and rimu seed in winter. Usually there is only one or two major food sources available at one time, so if one source is removed then there will be a negative impact on the bird concerned. Therefore, key factors necessary to maintain diverse wildlife communities are the diversity of forest and vegetational patterns and the presence of environmental gradients (such as altitudinal and climate gradients).

3.2       Distinctive wildlife within the Timberlands Application Area
3.2.1     I define distinctive wildlife as species whose status in the North Westland area is unusual. These are species that are characteristic of, and relatively common in the region, including Timberlands Forests, but are rare or absent in much of the rest of the country. Examples of distinctive species are great spotted kiwi, New Zealand robin and weka.

3.2.2     The great spotted kiwi only occurs in forests of the north west of the South Island, and is classed as theatened nationally. It ranges from Arthurs Pass (where it is rare) through to North-west Nelson, including Paparoa Range and the Karamea area where it is common. Kiwi are present in the Paparoa, Te Wharau, Orikaka and Ohikanui forests in the Timberlands Application Area.

3.2.3     Robins occur widely in the South Island but they have a fragmented distribution. They have vanished from most forests on the East Coast of the South Island, and from the extensive forests of South Westland. The North Island subspecies has also vanished from much of that island. No one knows why, but North Westland is a special place because robins remain common in many of the forests. Some of the largest populations nationally occur in Maruia and Inangahua forests.

3.2.4     A similar example is western weka. While not confined to forests, Buller and Tasman Districts are the national stronghold for this species. Again, weka have disappeared from much of its former range in the South Island.

3.3       Nationally important sites for rare, theatened species within the Timberlands Application Area
3.3.1     Timberlands own Wildlife Reports recommended that "Logging should not be carried out in areas where numbers of kiwi, kaka, parakeet, robins and other threatened or endangered species are concentrated". Threatened species are defined as those which will become extinct in the short to medium term if factors causing their decline are not halted or reversed. Ten nationally threatened bird species, and at least one threatened bat species have been recorded within the Timberlands Application Area, although one of these may now be extinct there (Table 1 & Bell 1986, Molloy, Davis & Tisdall 1994). Numerous concentrations were recorded in the Application Area, and data for these can be found within the Timberlands Wildlife Reports cited in Appendix1. Seven of these species are absolutely dependent on forests for their food, breeding sites, and shelter. The Department of Conservation has recently confirmed that orange-fronted parakeets are a valid species and are classified Category A (in urgent need of conservation). The Department has reviewed information on its former distribution and found that one of the last records of the species was from the Maruia area. A specific search for this species in the Maruia area has been timetabled to commence during summer 1999-2000. I am pleased to report that recently Department of Conservation staff rediscovered the species in the Cave Stream area of Station Creek within the Application Area.

3.3.2     Conservation of these threatened species is deemed to be of very high priority by the Department of Conservation and significant research and management programmes are in progress which aim at developing techniques for restoring viable populations of these species. Birds such as kaka and kiwi and the two bat species have been declining steadily, and all habitats where they occur should be considered key sites for their recovery. However, their distributions are now fragmented, and many of the important populations tend to be concentrated in the lower-altitude forests. Ensuring the continued viability and health of these populations, and the forest processes that sustain them, should be paramount before making decisions as to whether logging or clearance should occur in any particular forest.

3.3.3     The threatened species also tend to be those least able to cope with changes in their habitat brought about predation and competition with introduced mammals, and modification or loss of habitat. They have specialised food and nesting requirements that need to be met. Therefore, they have become rare because they no longer have enough suitable feeding and breeding trees or adequate shelter to sustain their numbers. Individuals are preyed upon by introduced pests at a rate higher than the animals can replace themselves naturally, and introduced species out-compete them for resources they require to thrive. Those threatened species listed in Table 1 that are already rare are the most likely to decline if remaining habitats are further modified.

3.3.4     The Department of Conservation is having some success at developing techniques for restoring forest wildlife communities. Integrated programmes of pest control and habitat enhancement have commenced throughout the country and are achieving many positive results. For example, in nearby Nelson Lakes National Park, over 18 kaka chicks survived long enough to fly independently from their nests over the last two summers, where formerly virtually none survived.

3.4       Ecosystems processes and wildlife within the Timberlands Application Area
3.4.2   (sic)   Many complex interactions occur between wildlife and plants in indigenous forest. Many of these interactions are not well understood, nor are the consequences of increasing rarity or removal of plant species from the system on the indigenous biodiversity. Cavity nesting wildlife, such as kaka, rifleman, parakeets and bats, is dependent on natural processes of decay of forest trees to provide suitable nesting sites, while they in turn play a role in breaking down dead wood and contributing to nutrient cycling as a result of their foraging behaviour.

3.4.2     Species present in the Timberlands forests are important for maintaining many forest processes. For example, kaka, tui, bellbird and short-tailed bat are important pollinators of flowering plants in the forest. The New Zealand pigeon, along with tui and bellbird are important dispersers of forest seeds. If pollinators and seed dispersers are reduced in numbers then forest health may decline, seed banks may be reduced, and regeneration of plants limited. For example, kaka are a major pollinator of Peraxilla mistletoes, and yellowheads a major disperser of Peraxilla seeds. The rarity of these mistletoes in some areas reflects a loss of their dispersing bird species. Lack of dispersers means that new sites cannot establish, and regeneration of forests, which may be less biologically diverse, is affected.


4.1       Sustaining wildlife populations: the context of the Resource Management Act

1.     The above evidence highlights the outstanding wildlife values present within the Timberlands Application Area that are significant in the context of the Resource Management Act (RMA) 1991 particularly Section 6c. The values of these forests and the presence of rare and threatened species have been confirmed and detailed in the wildlife reports and surveys commissioned by Timberlands and acknowledged in their Consent Application.

2.     Given this evidence, I contend that the Timberlands Consent Application needs to demonstrate that the logging they propose will sustain wildlife populations at their current levels, particularly significant populations and threatened species, in perpetuity, in the context of the RMA (and the Rules of the Buller and Tasman District Plans). Crucial to this assessment is whether sufficient nesting and feeding habitat and shelter remain to maintain wildlife populations in the areas logged following timber extraction and whether the Specific Controls Timberlands propose to mitigate threats are adequate.

3.     I have read the Consent Application thoroughly and the bulk of my remaining evidence will assess whether sustainability of wildlife populations can be achieved.

4.     Timberlands refer to several consultation processes it has been involved in prior to lodging this Application (See Section 7 of the Application). The Department of Conservation has made many recommendations for mitigating impacts on wildlife but the key recommendations have not been adopted. I will outline these below.

4.2         Impacts of past logging techniques on forest wildlife communities
4,2,1     Data I present in this section serve as a warning about the severity of potential impacts of logging should the management regime proposed by Timberlands not sustain wildlife populations. Many New Zealand studies have compared bird populations in forests which have been logged, with nearby unlogged forest, or compared populations in forests logged at different times (see Appendix 2). These studies include clearfelling of forest where the majority or all merchantible (sic) stems have been removed, partial logging where a certain number of seed trees per hectare were left, and selection logging where usually about 25% of merchantible (sic) trees were removed (see Table 2). The evidence for native bird species being adversely affected by past-logging is very strong.

4.2.2     Overall, the numbers of native species and individuals declined significantly between unlogged and logged forests. However, the numbers of introduced birds increased significantly in logged forest, contributing to an overall increase in both total numbers and species in logged areas (e.g. McLay 1974). Increases in introduced species are undesirable. They prey on indigenous wildlife and may compete for food and nest sites.

4.2.3     Historically obligate forest-dwellers disappeared from logged forest, no matter how intensive the logging (Table 2). The occurrence of six key species in 17 comparisons between logged and unlogged forest is summarised in Table 2. Eight sites were cutover, 6 sites selectively logged and 3 sites partially logged. Kaka and rifleman were absent from 10 of the 13 sites where they formerly occurred. Kaka occurred in lower numbers at the remaining three sites and rifleman at one. Parakeets were absent from five logged areas and present in lower numbers at the other seven. Although brown creeper, tomtit and NZ robin were usually still present in logged forests, their numbers were frequently lower.

4.2.4     Of the threatened wildlife occurring within the Timberlands Application Area, published studies indicate that long-tailed bat and NZ Falcon sometimes forage in the open along forest edges, riparian strips, and indigenous shrubland, but return to the forest to roost and breed. A further eight endemic birds occur significantly more often in unlogged than logged forest (kaka, parakeets, New Zealand pigeon, rifleman, brown creeper, fantail, tit, and robin).

4.2.5     Conversely, seven species are usually present in significantly greater numbers in logged areas (dunnock, song thrush, blackbird, yellowhammer, chaffinch, redpoll and long-tailed cuckoo). Of these seven, all but long-tailed cuckoo are introduced species.

4.2.6     Hole-nesting species are significantly more abundant in unlogged forest. Logging has been particularly detrimental to kaka, parakeets and yellowhead in the past. In most instances these species have not been recorded in any type of logged forest at all.

4.2.7     Hole-nesters are particularly vulnerable to logging because they rely on the oldest and largest trees in the forest for breeding sites: old-age live trees and standing dead trees. There are 11 hole-nesting species in the Timberlands areas. These are listed in Table 1. They are dependent on processes of forest aging and timber decay to provide specific nesting sites used by wildlife. Different species require different types and sizes of cavity to provide the correct microclimate for raising young. Prolonged use of unsuitable roosts can reduce survival and productivity and may reduce the viability of a population.

4.2.8     My evidence below will demonstrate the importance of large cavity-bearing trees in the forest for threatened species as examples, and highlight the need for provisions for retaining these trees and ensuring suitable new trees reach the large size classes.

4.2.9     The logging regime Timberlands propose (sic) is new, and has never been tested for its impacts on wildlife, so the impacts may not be as severe as previous techniques. However, no-one knows the extent of impacts because logging of the types proposed has not been undertaken before. Therefore, given the potential risks outlined above, a precautionary approach is needed. In my opinion, experimental logging is very risky, because there is a high probability that threatened species will decline. Accordingly, an alternative approach is needed.

4.3         Predicting the impacts of logging on wildlife
4.3.1     A predictive approach should be developed for assessing whether the logging plans proposed by Timberlands can sustain the wildlife. An initial assessment should have been done before the Consent was considered because the potential impacts on threatened wildlife are high. This section of evidence describes some of the techniques that can be used to assess whether the logging will impact on wildlife.

4.3.2     Information on foraging requirements for wildlife provides the first step towards predicting the impacts of logging of preferred trees. The removal of any plant species used by birds would have a potential impact on them if the resource was limited. The degree of impact will depend on the level of extraction and whether or not the birds can switch to alternative food sources. The impact of removing plants that are used by wildlife in specific seasons and represent specialised food sources is likely to be high, because alternative food sources are probably rare.

4.3.3     In order to predict the impacts of new techniques, information on foraging and breeding requirements of wildlife species is also needed. This information needs to be compared against which trees are being proposed for extraction from a particular forest, and judgements need to be made about whether the remaining habitat is adequate to sustain the wildlife populations. As an example, myself and Peter Dilks (in publications in 1986, 1987, 1990, 1991) used this approach in South Westland to predict whether helicopter logging techniques proposed for that area would impact on wildlife communities. A simple model for predicting the impacts of selection logging on feeding habitat of indigenous forest birds was developed. The model compared the degree of overlap between the trees preferred by birds for foraging and those targeted for logging, and predicted the proportions of preferred habitat lost at different percentage extraction rates.

4.3.4     We found that tree species, which would have been a primary target for removal in South Westland, were critical to some forest birds. For example, kahikatea was important for N.Z. pigeon, rimu for N.Z. pigeon, kaka, kea, yellow-crowned parakeet, rifleman and brown creeper, and dead trees for kaka, kea and yellowhead.

4.3.5     As an example of how this approach to predicting impacts of logging worked we compared our predictions of "percent habitat loss" if silver beech were to be selectively logged with actual responses of forest birds to logging in Southland (shown in Figure 1). Eric Spurr collected the data on response to logging while working for the Forest Research institute. A comparison between the percent of habitat loss predicted and actual declines in yellowhead, parakeet and kaka numbers indicated that beech logging was highly detrimental to these key forest birds (Figure 1).

4.3.6     In the case of rimu removal, the model predicted, that if trees >100 cm in diameter were removed (that is, only 6.7% of the rimu present in the South Westland forests where the study was done) then over 33% of kaka habitat, 10% of tui habitat, 11% of pigeon habitat, and 2.5 % of parakeet habitat would disappear and alternative food sources would not have been available in the forests to balance the shortfall in food. If rimu >80 cm diameter at breast height ( abbreviated as dbh in the remaining evidence) were removed then the impact would be significantly greater: nearly 60% of rimu preferred by kaka would be removed, but 10% of parakeet trees and 30% of pigeon trees.

4.3.7     Another technique that can be used to predict impacts of logging is called Population Viability Analysis. This technique uses computer models to predict the effect of different factors on the productivity and maintenance of population levels in wildlife species. In 1991 colleagues and I (Seal et al 1991) looked at factors that would affect population viability of kaka. We found that because kaka breed very slowly, with a reduction in habitat quality of as little as 1% per year and an annual mortality rate of as little as 3% there was a high probability of extinction within 100 years. However, if both predation was reduced AND habitat quality increased then kaka populations could recover.

4.3.8     Past work on impacts of logging has usually considered impacts on feeding habitat of wildlife. New data collected over the last 5 years (e.g. parakeets, kaka, and bats) indicate that (i) preferred breeding trees could also be a primary target of logging in the future and (ii) rotation periods for management may not be long enough to sustain nesting sites in perpetuity.

4.3.9     In summary, while Timberlands propose using a new logging system, it is difficult to imagine, given the evidence summarised so far, that there will not be a risk of adverse impacts on some wildlife. I will give further evidence for this assertion below using specific examples of threatened species that, in particular, are sensitive to small degrees of habitat change. Predictive models, perhaps of the types I've used as examples in my evidence, should be developed before any logging be approved. Otherwise I would be far from confident that new logging systems being proposed would be sustainable for wildlife.

4.4         Will sufficient "wildlife trees" be maintained through the proposed Timberlands harvest cycles?
4.4.1     In the absence of modelling of the effects of logging on wildlife, as was previously recommended to Timberlands, and is again recommended above, I have undertaken an assessment of the potential impacts using data from other beech and podocarp forests to show which trees need to be retained for wildlife. I will then use Timberlands data on proposed harvest rates to predict whether there will be sufficient wildlife trees remaining following logging.

4.4.2     Maintaining "wildlife trees" through the harvest cycles is a crucial requirement of the proposed logging. In earlier submissions to the Ministry of Forests and Timberlands I have provided considerable detail on this matter.

4.4.3     By "wildlife trees" I mean those which provide food, shelter and nesting habitat for wildlife of the type referred to in evidence above. Without them the wildlife would not survive. Sufficient of these needs to be sustained in the logged forest as a minimum conservation requirement.

4.4.4     Wildlife uses a wide variety of trees but particularly important, especially to threatened species, are the older-aged trees, In a study in mixed beech forest in Fiordland Jane Sedgeley, Peter Dilks and I found that threatened species used 14 cavity bearing trees >80 cm dbh/ha (long-tailed bat, kaka, and yellow-crowned parakeet) for breeding and roosting. Most of these trees were red beech. Given the density of these trees in the Application Area, there is a high probability that any trees >80 cm dbh selected for harvesting by Timberlands would be trees used by these species. Given that it often takes a minimum of %gt;300 years for beech trees to reach a size whereby they are used for breeding it is very likely that silvicultural practices will slowly reduce the number of cavity bearing trees in the forest. Specific requirements for these species are outlined in Sections 4.9-4.12 below.

4.4.5     The key issue for sustaining wildlife is that enough trees in particular size classes survive (a) natural mortality and (b) removal by logging, to be recruited in the next age class.

4.4.6     The basic premise of sustainable management proposed by Timberlands is that the approximate forest structure will be retained through the harvesting process. Any trees logged should be surplus to those required to replace the existing structure. Timberlands undertake to harvest 50% of the surplus.

4.4.7     Rotation length is a crucial factor in managing potential impacts on wildlife. It takes silver beech trees 300-400 (up to 600) years and red beech 450-600 years to mature and reach a size that they provide breeding habitat for wildlife.

4.4.8     According to Timberlands data, in the Maruia Working Circle it will take on average c. 31 years for red beech trees 40-50 cm to grow 10 cm (and so be recruited into the next size class), c. 35 years for trees 50-60 cm, c. 39 years for trees 60-70 cm, c. 44 years for trees 70-80 cm and c. 50 years for trees 80-90 cm (Figure 2). These growth rates are taken from Table 5.2, p. 79 of the Maruia Sustainable Management Prescription. In reality the growth within a size class will not be linear like that shown in Figure 2 this because growth slows down as trees age. (sic)

4.4.9     These growth rates give us a strong indication of how long logging cycles should be for particular size classes for them to have the potential of being replaced. That is, at a minimum, a new cycle of logging trees 80-90 cm in diameter should not begin for approximately 50 years.

4.5         Predicting impacts of proposed Timberlands harvest on "wildlife trees".
4.5.1    I have used the Timberlands data on current densities of trees within the Maruia Working Circle to assess whether it is likely that sufficient wildlife trees will remain in the forest over the time scales it takes for each size cohort to naturally replace itself. The reality is that there are about 17 potential wildlife trees/ha in the Working Circle, and this is about the density that needs to be retained in the future to maintain current forest structure. The immediate risk is that there is a high probability that wildlife trees will be felled and that, even if these are replaced in 50 or 100 years time, the damage would have been done because in the intervening years either the wildlife in question was killed when the tree was felled, or the nesting site was lost so the wildlife had nowhere to breed.

4.5.2     In mature forest, Timberlands propose taking 15 trees >30 cm dbh/ha in total per logging cycle. (Consent Application and Table 5.5, p. 84 Maruia Sustainable Management Prescription). Thus over 35 years, and if the cycle is tied down to 15 years , then there will be up to 3 cycles/ha (that is, up to 45 trees/ha removed).

4.5.3     Crucial to determining the impacts of this level of logging are the recruitment and mortality of trees. There seem to be few data for these factors.

4.5.4     The recruitment rate can be viewed in two ways. Firstly in the context of new seedlings being established and their rate of growth, and secondly in the context of whether smaller size classes of trees grow sufficiently within the logging cycle to replace those larger trees that have died or been removed by logging (between-cohort growth rates, or "in growth" as Timberlands calls it).

4.5.5     The first measure of recruitment mentioned is a long-term process. It takes several hundred years for trees to reach the size of wildlife trees.

4.5.6     However, in the short term there will probably be severe impacts on the availability of wildlife trees because harvest and mortality appear to outstrip between-cohort replacement rates. Although trees are moving into and out of each size class all the time, the growth rates noted in 4.4.8 above, and the harvest rates proposed by Timberlands, indicate that it takes considerably longer than the 15 year logging cycle for smaller trees to replace larger ones.

4.5.7     Also crucial to assessing sustainability for wildlife is accurate forecasting of mortality. There are several ways to estimate this. We could use the reduction in stem density between size cohorts as an estimate of between-cohort mortality. These figures are an actual measure of the type of mortality that had occurred (Figure 3) and this is the forest structure Timberlands are aiming to maintain. In the Maruia, estimated mortality (reduction in density) was very variable, ranging from 0-43% for red beech and 0-100% for silver beech between stem-diameter size classes. Reduction in density between size classes averaged 24 14.8 (1 SD) %. As we would expect, mortality is lower in younger trees and higher in older ones. This is a conservative mortality estimate for red beech. Wardle (1984) estimated a higher mortality rate of 0.68%/year (34% over 50 years) and June & Ogden (1978) estimated 0.6&/yr (30% mortality).

4.7         What will happen to the wildlife trees in Maruia given these figures?
4.7.1     I look in detail in this section at potential impacts of harvesting in the Maruia Working Circle. This is where impacts on wildlife are likely to be highest because:
(a) Overall wildlife values, and the significance of populations of threatened wildlife species, is (sic) higher than in the Inangahua and Grey Working Circles (though there are some specific high values in these working circles too):
(b) Here tree growth rates are significantly slower (e.g. trees 80+ cm grow at 2.04 mm/yr, compared with 4.19 mm/yr (Grey) and 3.96 mm/yr (Inangahua);
(c) Mortality rates appear to be higher in the Maruia:
(d) Compared to the other working circles large diameter beech make up a greater proportion of the density (very likely why wildlife values are so high);
(e) Proposed harvest rates, particularly of red beech, are also much higher than the other working circles (the proposed harvest rate for red beech is 5.88/ha in Maruia compared to 3.405/ha in Inangahua and 1.635/ha in the Grey).

4.7.2     In the Maruia working circle there are currently 7 potential red beech wildlife trees per hectare in the 80-90 cm size class, and 6/ha in the >100cm class (Table 5.1, Maruia Sustainable Management Prescription). To retain the existing structure then approximately 4 of the 7 trees per hectare of a size of 80-90 cm dbh need to survive to replace the larger trees. Initially this leaves approximately 3 available for natural mortality and harvesting in the Maruia over the c.50 years it takes trees 80-90 cm on average to grow through (Figure 4). In addition, potentially there is "in growth" from the 70-80 cm class (if the harvest rate for that cohort isn't too great).

4.7.3     So, how many of these trees (80-90 cm) do Timberlands propose to fell, and how many do we expect to die naturally?

4.7.4     The cutting rates for red beech 80-90 cm are specified at 0.055/ha/yr. That means 2.75 trees/ha (out of the maximum available 3/ha) would be harvested over the 50 years it would take for smaller trees to grow large enough to replace them (ie short-term recruitment). This equates to 1.93 trees over the 35 years of the Consent application. If we use the average mortality estimate of 24% over the 50 years then in addition 1.47 trees/ha will die. If we use the reduction in density between 80-90 and 90-100 then the estimated mortality would be 36% over 50 years (2.52 trees/ha), so I would be underestimating natural mortality considerably. These predictions are summarised in Table 3 and Figure 4.

4.7.5     In my opinion the theoretical surplus for trees 80-90 cm should be 50% of trees available after mortality is taken into account. That is, 3 - 2.52 trees/ha = 0.482 = 0.24/ha/50 years. This rough calculation equates to 1 trees/4 ha in the 80-90 cm class per 44 years available for "sustained yield" if the wildlife trees are to be retained (compared with 11/4 ha proposed by Timberlands).

4.7.6     Theoretically the short-fall in harvestable trees should be made up for by in-growth of some of the trees from the 70-80 cm class. But noting my earlier calculation, that only a proportion of these trees would potentially grow through in 35 years because it takes on average 44 years for them to grow 10cm. the question is, will enough of these smaller trees be recruited into the wildlife classes to make up for the short-fall created by harvesting? Lets look at the predicted harvest rate for trees 70-80 cm.

4.7.7     There are only 11 potential recruits/ha (ie those currently in the 70-80 cm class) available to replace the 7 trees/ha in the 80-90 cm class in the short term - leaving 4 potentially available for mortality and harvest in this size class. For Maruia I estimate only 5.2 - 5.9 trees/ha will remain as potential recruits to replace the 7 trees after mortality and harvesting are accounted for - a short-fall of 1-2 trees/ha (Table 3). This would only get worse over the initial cycles because the new recruits from small trees will take so long to get large enough to replace them.

4.7.8     Based on my calculation I predict a significant short-fall in trees available to replace those 80-90 cm. Most, or all, of the ingrowth is required to replace stand structure. Most, or all, of the surplus trees in the 70-80 cm class will either die or be harvested, and not be available to grow through.

4.7.9     Modelling by Murray Efford of Landcare Research (Efford 1999) warns of the risks of using the types of model used by Timberlands for predicting permissive cut that might be sustained, especially when the models are so sensitive to the tree mortality figures entered into the model. The types of analysis above are designed to illustrate scenarios of what is likely to happen to wildlife trees. These should be modelled thoroughly, but it seems to me that there is very little room for error in the harvest rates proposed by Timberlands. Even optimistically, it seems that barely enough trees would grow through to the next size class, so harvest rates for wildlife trees would need to be far more conservative. The situation is worst for the larger trees (90-100 cm) because they grow much slower in the Maruia and have higher mortality. Only 13-28% of recruits will be available to replace these important trees in just 50 years (Table 3).

4.7.10     These results imply that there should be no, or a very low levels, (sic) of logging in the Maruia of trees 70-80, 80-90 and 90-100 cm if larger size cohorts are to be replaced in the short term. If the predicted pulse of new recruits eventually grows through, then these surplus trees could potentially come available in 100-200 years. Timberlands use the figure of 57 recruits/ha/year. However, these are all small trees entering the 0-10 cm size class (that is, long-term recruits). It will be >250 years before these trees are big enough to compensate for the loss of trees 80-90 cm (Figure3).

4.7.11     If we underestimate mortality and harvest more trees than are replaced, then the impact could be extremely adverse. If we are more precautionary and overestimate mortality and so reduce harvest rates now it simply means that more trees might be available for extraction in the future (in cycles beyond 35 years) if the compensatory growth and recruitment do eventually grow through. What the example in Table 3 shows is that there is a very high probability of wildlife trees being felled by the operation as presently proposed, that the 50% rule for harvesting only 50% of the increment will be invalidated, and, at least in the short term, these trees won't be replaced by recruits. This will create a severe bottleneck in the availability of nesting, roosting and foraging trees for threatened species.

        How good will Timberlands be at predicting which trees will die naturally
4.8.1     Timberlands argue in their consent that much of the tree mortality will be harvested, so that trees loss might not be as great as the sum of harvesting + mortality as shown in Figure 4 and discussed in the previous section. They intend to mimic natural mortality by felling trees in groups in a similar pattern to that of windthrow gaps. They also intend to harvest dying trees. However, I could find no data on how much natural mortality they actually intend to subhume (sic) into the harvest.

4.8.2     However, I see some problems with this approach. I suggest that it will be virtually impossible to predict which trees are going to die over the time frame of the logging cycles and on the scale proposed. Although the design of the patches to be felled matches the natural patterns, there is no way to predict in which year and where the natural wind-throw events occur. Often trees that fall over in red beech forest appear very healthy from the outside. Those that don't are likely to be hollow and of no value for harvest. For example, if Timberlands log a particular compartment in Year 1, how will they know which trees will be blown over in, say, Year 5 or Year 14. There is every chance that they would chose (sic) the wrong patch of trees - and mortality would then have to be added to the harvest rate (Figure 5) and the harvest rate for the compartment would be doubled. Theoretically this means that another compartment should be left without harvesting for another cycle, but what happens if there is high mortality in the remaining compartments. The technique still doubles the chance that wildlife trees will die.

4.8.3     Also, a high proportion of the older-aged wildlife trees in red beech forest die slowly and independent of windthrow patches(sic). They are scattered throughout the forest. It is very difficult to imagine how Timberlands would predict where trees that are going to die during a 15 year period would be. Again, if Timberlands log a particular compartment in Year 1, how will they know which trees will die in, say, Year 14. There is every chance that they would chose (sic) the wrong tree - and mortality would then have to be added to the harvest rate.

4.8.4     Standing dead trees are vital for threatened species and other forest birds. Even if the location of dying trees can be predicted, this would be contrary to one of the important mitigation techniques proposed by Timberlands - that is, leaving dead trees alone for wildlife. On one hand, they will target trees that are dying so that they capture some of the natural forest mortality in their harvest rates but, if they are successful, then the occurrence of dead trees that they seek to preserve will be reduced considerably after a few logging cycles and this will negate their exempting the current dead trees from logging.

4.8.5      If Timberlands focus on capturing the natural mortality then there will always be a bias toward logging old age trees, despite this not being the intentions.

4.8.6     Timberlands also argue for compensatory growth of trees once the gaps have been created. However, this would be biassed towards new trees growing within the new gaps and the benefits would not be seen for perhaps 100 or 200 years or longer. This extra growth appears to be entirely theoretical. Potentially Timberlands could collect these data now by comparing growth rates of trees in old gaps with non-gap trees. However, given the fact that growth rates only vary by a couple of mm per year in different situations, the additive effect of compensatory growth would be minor for old-age wildlife trees.

4.9         Predicting impacts of logging on long-tailed bats
4.9.1     There is an active recovery programme for long-tailed bats being run by the Department of Conservation. The Recovery Group, for which I am leader, is becoming more concerned about the conservation status of this species. Preliminary survey work on long-tailed bats has indicated that numbers may be much lower than was previously thought, and populations could be declining. Recent work investigating the genetics of long-tailed bat populations indicate that North and South Island populations are distinct, and there is greatest concern for the viability of South Island long-tailed bats. Intensive work has indicated that long-tailed bats are rarely encountered in many areas, even when considerable survey effort has been made. They are very rare or absent at several South Island sites where they were widespread in the 1960's.

4.9.2     Records of bats from indigenous forests in Buller and Tasman Districts and surveys run by the Department of Conservation, Timberlands WEst Coast, and private individuals over the last three summers indicate that some forests in the Application Area support some of the only significant populations of the threatened South Island long-tailed bat. Using DOC's standard monitoring technique we have recorded one of the highest bat activity levels recorded so far in New Zealand in the Maruia Valley, including the Maruia Working Circle, with the exception of the Eglinton and Dart Valleys in Fiordland. This means that the conservation of bats in the Maruia Working Circle is of very high priority.

4.9.3     Bat populations investigated in red-silver-mountain beech forest in eastern Fiordland have very specific roosting areas. While feeding over 50 km2 at night, they always return to traditional patches of forest as small as 2 km2 during the day. Within the roosting ranges they use clumps of 3-6 trees in patches at any one time, then move to a new clump of trees after a few days. The implication of this is that unless these roosting patches are identified and protected then there is a risk of localised logging wiping out a whole population.

4.9.4     Determining statistically whether or not an animal selects particular trees (compared to: if they avoid them or simply use them at random) is important for identifying critical elements of the forest structure for threatened species. For bats Jane Sedgeley and I determined if particular trees were important by comparing bats' roost sites with the characteristics of available sites. This study concluded that log-tailed bats actively selected the largest trees in the forest for roosting and breeding: over 70% of trees used are greater than 80 cm in diameter (Figure 6).

4.9.5     Sampling of trees available to long-tailed bats indicated that suitable trees were rare in the forest and were concentrated in lowland forests on river terraces and outwash fans on the valley floor. Ninety-five percent of roost trees were in mature, open-structured lowland forest on the relatively flat valley floor within 500m of the forest edge (Sedgeley & O'Donnell 1999a - see Appendix 2).

4.9.7 (sic)     Four tree species (including dead trees) were used as day roosts in beech forest in Fiordland. Long-tailed bats did not discriminate between tree species per se, but selected roost trees that were taller, which had relatively low canopy closure, larger stem diameters, larger trunk surface areas, and greater numbers of cavities than random trees. Red beech trees and dead trees were most likely to provide these preferred characteristics in Fiordland forests. Seventy-four percent of roost trees were c. 100 - >600 years of age. Such trees have been a target of logging in the past, and any sustainable management system used in the future needs to leave sufficient trees to ensure bat populations survive.

4.9.8     Little work has been done on determining the type of trees long-tailed bats roost in in podocarp-hardwood forests. However, this summer I have recorded roosts in large diameter (> 1 metre diameter rimu and totara trees in Te Kuiti and Geraldine respectively. In addition, Gillingham (1996), a Masters student at Massey University, found that that (sic) in the North Island long-tailed bats used rimu, miro, kahikatea and red beech, 50-140 dbh, for roosts.

4.9.9     Because long-tailed bats switch roost sites on a daily basis they require a large number of these specialised trees and cavities (>100 trees per colony).

4.9.10     Bats did not simply use any cavity. For example, each cavity used by bats for breeding has specific dimensions and stable thermal characteristics to enable them to rear young successfully (Sedgeley & O'Donnell 1999b and, for example, Figure 7). They move among roosts, which provide significantly different characteristics at in (sic) different weather conditions.

4.10         Predicted impacts of logging on short-tailed bats
4.10.1     Forests in Buller District support the only significant populations of the theatened South Island short-tailed bat known outside of the Eglinton Valley, Fiordland. In addition, these bats form a genetically distinct type, not found elsewhere. Recent surveys led by John Lyall of the Department of Conservation in Westland, and involving many staff, have found these bats in the Oparara and Paparoa areas, but based on these findings they could be present elsewhere in the Buller and Tasman Districts. There has been one potential report from the Glenroy Valley in the last five years. Technology for finding these difficult to find nocturnal bats has only been developed in recent years, so there is considerable work still be done in defining their distribution, and in characterising their habitats and requirements. I recommend specific searches for short-tailed bats are done on the Timberlands Application Area.

4.10.2     Two studies specifically indicate that logging could be detrimental to short-tailed bats. A study by Jane Sedgeley Jenny Christie and myself in mixed beech forest in the Eglinton Valley in Fiordland shows that short-tailed bats use the large diameter, cavity-bearing trees in the forest (Figure 8a). The proposed logging would likely target a significant proportion of these trees, Therefore such trees need to be identified and protected.

4.10.3     A second study of short-tailed bats I have been involved with was undertaken in a different forest type, podocarp-hardwood, on Codfish Island. Results show a similar dependence on large diameter trees, with hall's totara, rimu, southern rata and miro being used. (Figure 8b). However, a proportion of roosts was in smaller sized trees as well. These were most often used by solitary bats and by small groups that require smaller cavities for hibernation.

4.11         Predicted impacts of logging on kaka
4.11.1     Like the bats, kaka also uses only large diameter trees for nesting. Peter Dilks of the Department of Conservation has collected nesting information from mixed beech forest in Fiordland over the last year. There, kaka nests were all in red beech trees > 85 cm in diameter (Figure 9a)

4.11.2     Work by Terry Greene of the Department of Conservation in podocarp-hardwood forests in the central North Island shows that the largest trees are used (Figure 9b). Matai was the main cavity-forming species in that study, but rimu, miro, totara and hinau were also important.

4.11.3     Any logging proposal would need to ensure that trees suitable for kaka were left in the forest in sufficient numbers to sustain breeding of viable populations and that new trees were becoming old enough to enter the older age cohorts.

4.12         Impacts of logging on parakeets
4.12.1     In beech forest in Fiordland, yellow-crowned parakeet nest in red beech (74%), silver beech (20%), and standing dead trees (6%). In red beech 33% of nests are in trees 70-110 cm in diameter (dbh) and 46% in trees >110 cm dbh. Such trees are usually 400-600 years old. The Consent Application proposes logging a significant number of trees of the sizes selected by parakeets and their removal would have a major impact on this threatened species.

4.12.2     In Rowallan Forest in Southland Eric Spurr and co-workers undertook standardised 5-minute bird counts in forest logged at different times. Parakeets were recorded on 34% of counts in unlogged forest. This average reduced to 12% of counts immediately after logging, to 5% of counts five years after logging, and to fewer than 1% of counts in forest 15 and 25 years after logging.

4.12.3     Now that orange-fronted parakeets have been re-discovered in the Maruia Valley then they will likely select similar trees to yellow-crowned parakeets although this needs to be investigated before any logging were to be approved so that important trees can be identified and protected.

4.12.4     The example of parakeets shows that the effects of logging do not necessarily manifest themselves immediately. Longer-term attrition in the quality of habitat for parakeets would impact more slowly on viability of populations. A study by Rhys Buckingham in Southland showed that parakeets and yellowheads had not returned to red beech forests even 80 years after selection logging.

      In this section I assess the adequacy of mitigation proposed by Timberlands for sustaining wildlife.

5.1         Setting maximum sizes of harvest trees
Timberlands say that they mitigate impacts on theatened species by setting maximum diameter limits on trees to be harvested (Page 101 of the Application). These limits are 110 cm for red beech and 110 cm for hard beech. No limits are set for rimu, despite these being very important wildlife trees. Earlier evidence shows how important beech trees >70 cm are for threatened wildlife. There are very few trees in the >110 cm category, so few wildlife trees would be protected by setting the size limits proposed by Timberlands. Setting size limits needs to be set down as a Condition of any future Consent Application, but the limits should be significantly lower than those proposed by Timberlands. I recommend 80 cm dbh as the maximum size because this will ensure that the majority of wildlife trees will be protected. The evidence gathered to date indicates that setting the maximum size of tree to be harvested to 80 cm dbh would mean that the majority (ca. 80%) of cavity breeding and roosting sites for threatened species would be retained (based on studies in Fiordland beech forest). However, based on the information from podocarp hardwood forests, some medium-sized trees are required as well, at least for short-tailed bats. Identifiying these trees would be one task of the proposed bat research.

5.2         Harvesting of rimu
Timberlands state in their application that no upper limit is set for rimu because they don't provide the same habitat as large old beech trees (Page 101 of the Application AEE). This is not true, and is contrary to the advice they received in their own Wildlife Reports. Large diameter rimu DO provide important foraging habitat for threatened species such as kaka and parakeets. Some of this information is presented in section 4.3 of my evidence but I could provide the Commission with considerably more data if they wished. O'Donnell and Dilks (1986 & 1987) found that all forest bird species in Westland fed on either insects, fruit, fungus, epiphytes, seeds and shoots present on rimu. Kaka preferred rimu > 80 cm with 60% of feeding observations in trees of this size. Recent research on bats in the North Island and on Codfish Island shows that where rimu is present some trees in the 35-130 cm dbh class are used as maternity roost sites. (Gillingham 1996, B. Williams personal communication, C. O'Donnell unpublished data). Timberlands acknowledge that rimu has a "completely different and significantly longer life cycle" than beech. Therefore it is difficult to imagine how rimu can be sustained in perpetuity. (sic) Considerable (sic) more research is needed and the response (sic) of rimu to logging would need to be modelled rigorously.

In addition, it is inconsistent for Timberlands to propose exclusion of other podocarps (kahikatea and miro) from logging as a mitigation technique because they provide important wildlife trees but, at the same time, allowing the logging of rimu, which is probably more important for wildlife. Therefore, the exclusion of rimu from logging needs to be set down as a Condition of any future Consent Application.

5.3         Retention of dead trees
Timberlands indicate that initially they will leave standing dead trees as wildlife trees (Page 106 of the Application AEE). Standing dead trees are vital for threatened species and other forest birds for breeding and food. Protection of these trees is a positive move. However, Timberlands reiterate through their Application that they also will target trees that are dying in the future so that they capture some of the natural forest mortality in their harvest rates. If they are successful, then dead trees will disappear after a few logging cycles because the original protected dead trees will eventually fall over, leaving a significant habitat gap for threatened species.

If Timberlands are always focussing on capturing the natural mortality then there will always be a bias towards logging old age trees, despite the best intentions. Therefore, Conditions in any future Application need to state that Timberlands will retain standing dead trees as suggested, but also that the bias of the logging towards predicting natural mortality be removed. This will ensure that dying trees that could stand for many years and become important wildlife trees are retained.

5.4         Increasing risks of predation
Another major risk of the logging not mentioned by Timberlands will be increased risk of introduced predators entering forests and preying on indigenous wildlife. Introduced predators are a major threat to forest wildlife (see Appendix 3). Predators such as stoats and rats cause significant declines in both common and theatened species. The Department of Conservation is investing considerable resources into managing predator problems and conducting research aimed at developing cost-effective techniques for predator control. As new advances are made the Department is expanding the number of sites in which integrated pest control programmes are being used. Future logging systems would potentially increase predation risk. Predators can disperse along roads and increase in numbers both along roads and along forest edges themselves (References are given in Appendix 3). Logging systems proposed by Timberlands call for greater roading and the increase in forest edge habitats through creation of artificial gaps in the forest. In the Eglinton Valley in Fiordland there were greater numbers of stoats and rats along the beech forest edge than in the forest interior and male stoats used roads significantly more than other habitats, at least in some years. Edges and disturbed sites harbour greater numbers of introduced birds and rodents, resulting in increased food levels that would sustain and attract predators.

Therefore any future Consent Application should have in it a Condition that predator control be undertaken throughout the forests and particularly along rosds (sic) and forest edges. The Applicants would need to outline the type and intensity of predator control so its adequacy could be judged. I note that the Joint Staff Report to the Hearing states that most pest control would focus on possums (page 21). Rats and stoats are the primary predators of native birds, so control needs to focus equally on them.

5.5         Role and effectiveness for wildlife of reserves proposed by Timberlands
Timberlands propose creating a number of reserves that exclude logging. These are proposed to protect amenity, landscape, recreational and wildlife values. This is a positive move by Timberlands but in itself does not represent a mechanism for mitigating negative impacts of management on wildlife within the logging zones. In addition, the location and size of the reserves means that these will do little towards protecting viable populations of many wildlife species. For a viable wildlife population within a reserve to be maintained in perpetuity the reserve needs to be sufficiently large to contain what biologists call "Minimum Viable Populations". To remain viable, a population needs sufficient numbers to maintain their genetic diversity and social structure, and buffer them against random catastrophes and other sudden events that might reduce the population. There is a wealth of information on how to maintain population viability and I refer you to a good summary paper by M.S. Boyce in the Annual Review of Ecology and Systematics Volume 23, pages 481-506.

One way to begin determining how adequate an area is to protect a species is by determining home range requirements of the species. First determine how much space one breeding pair of a species needs, then how much space a social group needs. Some of the threatened species discussed in this evidence have large home range sizes. For example, most yellow-crowned parakeets in red beech forest in the Eglinton Valley range over a minimum of 100 to 200 hectares. Average range sizes for long-tailed bats in the same area average between 330 and 1589 ha depending on age and stage of the breeding season. Kaka have similar sized ranges covering hundreds of hectares. The Application outlines an intention to reserve 8 areas across the 18 forests for wildlife reasons. Most reserves in are (sic) in the order of 100 to 200 hectares in size so clearly would not support viable populations of these threatened species. Areas of this size are also unlikely to constitute forests of national significance.

The proposed reserves cover very little of the range of threatened species recorded by Timberlands in their wildlife assessments. For example, in their wildlife report, Timberlands noted that forests in the Inangahua Working Circle had some of the best kaka and parakeet counts recorded in New Zealand. Encounter rates of between 3 and 9 of each species per hour in Orikaka were outstanding, yet the proposed wildlife reserves cover less than 10% of the range of these species within the forest. High encounter rates were recorded in Ohikanui and Camp Creek. I estimated from the bird distribution maps in the Inangahua Working Circle that less than 5% of the range of threatened species was covered by the proposed reserves. This means that over 95% of the ranges of threatened species were within the forests proposed for production, and ways to maintain these species must be found. Contrary to statement (b) on page 18 of the Joint Staff Report to the Hearing I see no evidence that Timberlands have set aside viable reserves for known rare and threatened indigenous birds discussed in this evidence. The Joint Staff Report to the Hearing is also misleading in suggesting that only individual or small groups of rare or endangered birds may also occur outside the reserve area (Page 18 (d)). Timberlands own wildlife reports clearly show that the bulk of threatened species occur outside of the reserves. The Joint Staff Report recommendation that there be a condition for annual reporting of "accidental" discoveries of rare or threatened species once logging had commenced does nothing to safeguard these species, nor do the staff recommend any conditions on how they intend to take "action" to "avoid disturbance to these species.

5.6         Identification of bat zones
The Department of Conservation's earlier submissions to Timberlands pointed out the many threats its logging regime posed to threatened bats. Timberlands proposed as a Condition to the Consents that they will assess forest compartments for bats and reserve roosting trees when found (Page 106 of the Application AEE). This is a positive move, and will be necessary if bats are not to be impacted by the logging. However, Timberlands have probably underestimated how time consuming and expensive this will be. The only way to find bat roosts is by undertaking a long-term radio-tracking study. Long-tailed bats have very specialised roosting conditions and move around a number of trees (for example, approximately 400 trees in the Eglinton Valley) over their annual cycle. They stay on average only 1.7 days before moving onto another of their preferred trees. Therefore a roost might not be occupied on the day that felling occurs but its removal could be crucial. If bats use suboptimal roosts then breeding success and survival are low. I note that in the Joint Staff Report to the Hearing states (sic) that bats occur in "isolated small numbers or as single specimens" through the Timberlands estate. As far as I am aware no-one knows how many bats occur nor how they are distributed. However, evidence 4.9 above indicates to me that the population in the Maruia Working Circle is more significant than the Staff Report suggests. Therefore, a Condition of any Consents granted in the future needs to state that Timberlands identify bat roosting areas using a radio-tracking study well BEFORE any logging in an area commences (as opposed to after the event as suggested by the Joint Staff Report to the Hearing), probably several years, depending on the investment in the research, and the roosting areas identified need to be excluded from any logging.

5.7         "Improvement felling"
Trees that naturally age and eventually fall down need to be replaced by suitable cavity bearing trees over time. Therefore sustainable yield management systems need to ensure a large enough pool of trees be retained in smaller diameter classes so that they will eventually reach the larger size cohorts. Younger trees that have cavities developing in them should not be felled or pruned because otherwise future generations of large diameter trees may not contain cavities suitable for threatened wildlife. The Department of Conservation's earlier submissions to Timberlands pointed out the many threats of 'improvement fellings" for threatened species. I will not repeat these here because the consent application says there will now be no improvement felling (Page 40 of the Application). This is a positive move. However, the Prescriptions provided with the Consent Application still contain prescriptions for "improvement felling". Therefore, any Condition of any Consents granted needs to state that Timberlands will delete reference to "improvement felling" from their Prescriptions and Operations Plans so that there is no confusion among forestry workers.

5.8         Wildlife monitoring
Timberlands note difficulties in undertaking formal monitoring of bird populations, especially problems with isolating impacts of logging from potential changes brought about by pest control ((Page 112 of the Application). However, the logging programme proposed is highly experimental and the potential impacts on representative bird species unknown. If some logging were to be approved in the future a Condition of any Consents granted needs to state that Timberlands would contract an independent body to monitor forest bird numbers at stages through the programme. Regular review dates would be necessary. The problems Timberlands cite can be overcome by proper and rigorous experimental design. For example, large enough samples of 5-minute bird counts need to be undertaken in treated (logged) and untreated (unlogged) forests. The experimental design should include treated areas that include both logging and pest control and areas with logging but no pest control

1.     I am concerned about both the direct impact of felling trees on wildlife and the longer term impact of significant reduction in "wildlife trees" particularly on threatened species. Timberlands West Coast Limited have not demonstrated that the logging they propose will sustain wildlife populations in perpetuity within the Application Area as they should have done to fulfil requirements of the Resource Management Act 1991. While Timberlands have made some allowances for wildlife I consider these are currently inadequate to avoid, remedy or mitigate adverse effects.

2.     Therefore, given the evidence presented here, and the significant risks posed to theatened wildlife, I recommend that a precautionary approach to logging must be taken, particularly in the Maruia Working Circle. There are currently too many unanswered questions regarding whether wildlife populations will be able to be sustained within the forests proposed for logging.

3.     The consents should be declined at this time until further research on impacts on wildlife, particularly threatened species, has been conducted, and the implications of the proposed logging can be judged more clearly.

4.       Specifically, Timberlands should:

  1. Determine the availability of wildlife cavities and model their persistence.
  2. Undertake specific modelling of short and medium term, as well as long term. persistence of wildlife trees >80 cm dbh.
  3. Undertake research that identifies bat roosting areas in the Maruia working circle using a radio-tracking study and protect these trees.
  4. Determine proportions of trees that die naturally in gaps compared with individual deaths, and model the persistence of standing dead trees.
  5. Determine with confidence how much natural mortality would be subsumed by their proposals.
  6. Undertake an intensive survey for short-tailed bats.
  7. Determine the habitat requirements of orange-fronted parakeets and protect key sites.
  8. Design mitigation techniques that properly ensure the protection of theatened species listed in Table 1 of this evidence.
5.      If revised consents are applied for in the future, the application should include better conditions for mitigating impacts on wildlife. The applicants should include:
  1. Reduced maximum stem sizes of trees to be harvested to ensure wildlife trees are protected.
  2. Reduced extraction rates per hectare for wildlife trees.
  3. Protection of rimu within working circles.
  4. Specific conditions that ensure sufficient dead trees are left in future logging cycles to replace trees currently exempt from felling.
  5. Realistic provisions for identification of bat roosting areas.
  6. No provision for "improvement felling".
  7. Rigorous wildlife monitoring protocols.
  8. Mitigation of predator risks following new roading.
  9. Better use of existing wildlife data from Timberlands surveys to identify and protect concentrations of threatened species.
  10. More specific prescriptions for predator control.
5. (sic)     I also note that the proposed Consent Conditions written in the Joint Staff Report to the Hearing contain few conditions pertaining to wildlife habitat. These conditions do not adequately mitigate the concerns outlined above. I also note that not even all the conditions proposed by Timberlands in the Consent Applications are included in these Draft Conditions (for example, no logging of rata, miro, kahikatea, standing dead trees; tree size limits; rotation lengths; see pp xxv, xxvi & section 8 of the Application AEE).