Using Geographical Information Systems to identify and target sites for ...

Journal of Environmental Management (2002) 64, 25–34 doi:10.1006/jema.2001.0495, available online at http://www.idealibrary.com on

Using Geographical Information Systems to identify and target sites for creation and restoration of native woodlands: a case study of the Chiltern Hills, UK John T. Lee* , Neil Bailey and Stewart Thompson School of Biological and Molecular Sciences, Oxford Brookes University, Gipsy Lane, Headington, Oxford, OX3 0BP Received 16 November 2000; accepted 9 June 2001

Rare and threatened habitats in Europe must be restored and enhanced in accordance with the European Union’s Habitats and Species Directive. In the United Kingdom, conservation and expansion objectives for species and habitats are outlined in the Species Action Plans and Habitat Action Plans. Site identification for these measures has to date been ad hoc without consideration of either the existing ‘‘stock’’ of the natural resource or the ability of the surrounding land use to deliver the enhancement (enlargement) of a given habitat. Using a Geographical Information System, we outline a targeting system for creating new woodland in association with existing ancient woodland in the Chiltern Hills Area of Outstanding Natural Beauty. The aim was to create woodland blocks of at least 100 ha, as being of the most benefit to biodiversity. We identified existing patches of woodland between 20 and 50 ha as cores for habitat expansion and classified land use in terms of its suitability and proximity to these core areas for tree planting to meet the targets of the statutory body. Our results suggest that the targeting method employed is a useful tool for habitat restoration.  2002 Academic Press

Keywords: spatial targeting, woodland restoration, Geographical Information Systems, biodiversity.

Introduction With the publication of the Biodiversity Action Plan, the United Kingdom (UK) Government pledged to conserve and enhance the nation’s biodiversity (Department of the Environment, 1994). This is to be achieved primarily with the identification of priority habitats and species and the production of Habitat and Species Action Plans (HAP and SAP) to conserve and enhance their distribution. These plans are projected to cost between £29 million and £44 million per annum until 2010, costs which are in addition to existing financial commitments for conservation (Department of the Environment, 1995).

Ł

Corresponding author. Email: [email protected]

0301–4797/02/010025C10 $35.00/0

The proposed Agenda 2000 reforms to the Common Agricultural Policy (CAP) offer considerable opportunities for increases in nature conservation funding, particularly under Modulation whereby payments to farmers are likely to be increasingly linked to Agri-Environment Programmes (AEPs) (Ministry of Agriculture, Fisheries and Food, 1999). Currently, however, financial resources for conservation are severely limited. To maximise environmental benefits, resource targeting is needed, since it offers the best opportunity for maximising policy efficiency and enables policy makers with limited budgets to set priorities and direct resources (Cook and Norman, 1996). There are numerous examples of spatial targeting used in nature conservation: from the UK (Webster and Felton, 1993; Wilson, 1997; Yeo et al., 1998; Woodhouse et al., 2000), North and South America (Scott et al., 1993; Lathrop and Bognar, 1998; Ranta et al.,  2002 Academic Press

26

J. T. Lee et al.

1999; Stoms, 2000), Australia (Pressey et al., 1993) and South Africa (van Jaarsveld et al., 1998). To promote biodiversity in the wider countryside, and to fulfil the habitats and species objectives of the Biodiversity Action Plan in the UK, some form of spatial targeting is required (Wynne et al., 1995; Thompson et al., 1999). This will maximise the chances of generating ecologically valuable sites. Much of this resource targeting work has used Geographical Information Systems (GIS), since these are suited to the manipulation of large quantities of data, enabling a multi-layer, landscape-scale approach to habitat restoration. Furthermore, GIS allow integration of data collected from various sources and at a variety of appropriate spatial scales (Brown et al., 1998). The approaches used are either species-based (Brito et al., 1999; Dettmers and Bart, 1999) or focused at the habitat or ecosystem level (Russell et al., 1997; Thompson et al., 1999). For the research described here, GIS offered considerable advantages over manual map-based analysis methods in terms of both speed and flexibility. Here we present a spatially targeted method for restoration of native habitat (ancient semi-natural deciduous woodland1 ) using a land use database. Land use data were interpreted from aerial photographs for the part of the English County of Buckinghamshire within the Chiltern Hills, and 304 ancient deciduous woodlands were identified. A GIS was used to identify those woodlands which had the greatest potential for expansion based on their position in the landscape and their current size. This ‘aggregate with outliers’ approach was used to identify those parts of the landscape which would be best returned to, or created as, native woodland.

Data and methods Ancient semi-natural woodland Ancient semi-natural woodland as found in the Chiltern Hills Area of Outstanding Natural Beauty is listed under the UK Biodiversity Action Plan and forms part of the Lowland beech and yew woodland Habitat Action Plan. There are approximately 30 000 ha of the habitat remaining in the UK, 1

Ancient semi-natural woodland is land that has had continuous woodland cover since at least 1600 AD and has retained the native tree and shrub cover that has not been planted, although it may have been managed by coppicing or felling and allowed to regenerate naturally (English Nature, 2001).

comprising between 15 000 and 25 000 ha of ancient semi-natural woodland and the remainder of recent beech planting (UK Biodiversity Group, 1998). The habitat has a number of distinctive vegetation types reflecting the variations in soil type and topography, but the characteristic species are beech (Fagus sylvatica), ash (Fraxinus excelsior) and yew (Taxus baccata). Oak (Quercus robur and occasionally Q. petrea) is also common and an understorey of bramble (Rubus fruticosus) or holly (Ilex aquifolium) are characteristic under appropriate conditions (UK Biodiversity Group, 1998). These form the characteristic National Vegetation Classification (NVC) communities W12, W14 and W15 (Rodwell, 1991). Amongst the rarer plants found in ancient semi-natural woodland are the red helleborine (Cephanlanthera rubra) and species which have nationally restricted distribution, such as coralroot (Corallorhiza trifida) and the lesser hairy brome (Bromus benekenii) (UK Biodiversity Group, 1998). The Buckinghamshire local Biodiversity Action Plan identifies 21 woodland species listed in the UK Biodiversity Action Plan (Joanne Hodgkins, Buckinghamshire County Council, pers. comm.). The habitat requirements of these species are shown in Table 1. The table shows both those species known to be largely restricted to beech woodland, and those whose habitat requirements are satisfied by beech woodland, but which have a broader range. Although it is unlikely that all of these species will occur together in a single wood, it is desirable to create structurally and environmentally heterogeneous woodlands of high nature-conservation value. This can be best achieved by creating large woods through natural colonisation (Peterken, 1995; MacMillan et al., 1999). Large woodland patches are desirable, as there is a strong positive correlation between habitat patch area and species abundance; for example for forest birds (van Dorp and Opdam, 1987; Bellamy et al., 1996); plants (Simberloff and Gotelli, 1984) and mammals (Laurance, 1994). It has been suggested that it is desirable to create woodlands of at least 100 ha (Wynne et al., 1995). This can be achieved by additional tree planting in one of two ways. Either existing large woodlands can be increased in size by a small amount, or small woods can be expanded by a large amount. Given that there is a limited amount of conservation money available and hence a limited number of trees that can be planted, these two options offer contrasting benefits. In the first scenario, a large number of woodlands can be expanded to

Using GIS to target habitat restoration Table 1.

27

The 21 woodland species listed in the Buckinghamshire local Biodiversity Action Plan

Scientific name

Common name

Habitat requirements

MAMMALS

Muscardinus avellanarius

Dormouse

Mixed deciduous forest, especially edge and understorey shrubs.1

Plecotus auritus

Brown long-eared bat

Lightly wooded areas and sheltered valleys.2

Nyctalus noctula

Noctule bat

Holes in mature deciduous trees, especially beech.2

Coccothraustes coccothraustes

Hawfinch

Deciduous woodland with fruits of wych elm, hornbeam, beech, wild cherry, sycamore and maple.3

Luscinia megarhynchos

Nightingale

Woods and commons with dense thorn scrub thickets.3

Milvus milvus

Red kite

Mature oak woods on steep valley sides.3

Muscicapa striata

Spotted flycatcher

Woodland edge or forest glade, especially wooded streams and lakeside trees (3); open deciduous woodland.4

Pyrrhula pyrrhula

Bullfinch

Forest edge and scrub; also hedgerows, young forestry plantations.3

Streptopelia turtur

Turtle dove

Woodland with open areas for feeding.3

Turdus philomelos

Song thrush

Trees, bushes, woodland edges, hedgerows.3

Satyrium pruni

Black hairstreak

Broadleaved woodland with understorey; scrub & shrub; scrub, grasses, herbs, bracken; hedgerows.5

Thecla betulae

Brown hairstreak

Scrub & shrub; scrub, grasses, herbs, bracken; hedgerows.5

Ladoga camilla

White admiral

Broadleaved woodland with understorey, grass rides & permanent clearings.5

Stag beetle

Inhabits deciduous woodland and breeds in rotting wood such as tree stumps.6

Devil’s bolete

Calcareous soils with beech and oak.7

Knothole moss

Only one known location.8

Buxus sempervirens

Box

Woods & scrub on chalk and limestone.9

Cephalanthera rubra

Red helleborine

Beech woods on chalk and limestone.9

Epipogium aphyllum

Ghost orchid

Only three known sites.9

Orchis militaris

Military orchid

Only four known sites.9

Sorbus torminalis

Wild service tree

Woods, scrubs & hedgerows on clay or limestone.9

BIRDS

BUTTERFLIES

BEETLES

Lucanus cervus FUNGI

Boletus satanas MOSS

Zygodon forsteri VASCULAR PLANTS

Superscripts indicate references. 1 Mitchell-Jones et al., 1999; 2 Corbet and Harris, 1991; 3 Sharrock, 1976; 4 Cramp et al., 1993; 5 Pollard and Yates, 1993; 6 Crowson, 1981; 7 Harding et al., 1996; 8 Smith, 1978; 9 Stace, 1997.

100 ha but many of these may already make a significant contribution to regional biodiversity and increasing their area may have only minor benefits. The second option prioritises small woodlands for expansion, but each wood requires a large number

of trees so few woods can be expanded. By way of compromise, medium-sized woodlands should be targeted for expansion. This has the benefit not only of maintaining existing large woods (some of which may only be slightly smaller than 100 ha),

28

J. T. Lee et al.

but also of ensuring there are a sufficient number of small habitat patches, which have been shown to contribute to biodiversity (Fitzgibbon, 1997; Abensperg-Traun and Smith, 1999). Natural colonisation encourages maintenance of locally native species and genotypes (Peterken, 1995). However, natural colonisation is slow and unpredictable and may not be ideal for woodland enhancement (Hodge and Harmer, 1995). Consequently, woodland expansion may best be achieved by a rolling programme of tree planting in appropriate locations, followed by sympathetic management leading to species diversification and aided by later natural colonisation. This can be encouraged by targeting those parts of the landscape that are firstly proximal to existing medium-sized patches of woodland and secondly of suitable land use type for tree planting (Gilbert and Anderson, 1998). This forms the basis of the method described here.

Study area

between 700 mm and 850 mm of rainfall each year and has a mean annual temperature of approximately 10° C. The region is dominated by chalk and the escarpment faces north-west rising to 250 m, behind which the dip slope falls. This is characterised by open downland and rolling farmland, cut by a series of deeply dissected valleys. The soils are strongly associated with the underlying chalk geology and are typically shallow and nutrient-poor, resulting in characteristic, semi-natural vegetation communities dominated by woodland and grass. The area is predominantly rural, with over 75% of land being agricultural and some 20% wooded. Tree cover compares favourably with the national average of 7Ð5% but falls well short of the average for the whole of the European Union of 36%. Past patterns of agricultural practice (clear felling, followed by ploughing) have led to habitat loss and fragmentation, with few remaining woodlands in the Chiltern Hills more than 100 ha in size.

Data collection

The Chiltern Hills stretch 90 km north-east from Goring-on-Thames in Oxfordshire to Hitchin in Hertfordshire (Figure 1). The area receives

The land uses were interpreted from the most recently available aerial photographs taken by

Hitchin

London

Princes Risborough

High Wycombe

Goring-onThames

Henley-on-Thames OS tile SU79

Figure 1.

N

Approx 10 km

The Chiltern Hills Area of Outstanding Natural Beauty showing Ordnance Survey tile SU79.

Using GIS to target habitat restoration

Buckinghamshire County Council in 1995. The outlines and land use codes of the polygons were transcribed to hardcopy from the aerial photographs. The boundaries and polygon attributes were then digitised on-screen with a GIS (Arc/Info ver. 7Ð1) on 1:10 000 raster base maps supplied by the UK Ordnance Survey (OS). The positional and attribute accuracy of the habitat patches was validated using digital outlines of Sites of Special Scientific Interest provided by English Nature and that of deciduous woodlands from Forest Enterprise.

Methods Fragments of semi-natural ancient woodland (N D 304) were identified from the Ancient Woodland Inventory of the English County of Buckinghamshire (Farino, 1986) (Figure 2). These included all those in the County that occur within the Chiltern Hills AONB and cover an area of 5382Ð6 ha. For the purposes of the research, all of these fragments were assumed to be beech woodland. The spatial targeting of restoration was based on two objectives. First, the requirement of the lowland beech and yew woodland Habitat Action Plan (HAP) to increase the current extent of

this rare habitat by 10% (UK Biodiversity Group, 1998). Second, the desirability of woodland patch expansion to form buffers or links between native deciduous woodland, particularly where site size can be increased to 100 ha (Wynne et al., 1995). Although there is an argument for increasing the size of the habitat in Buckinghamshire by more than the recommended 10%, we aimed to increase the reserve by 538 ha in keeping with the targets of the HAP. We targeted sites for tree planting based on their current land use, size and proximity to existing patches of beech woodland. Whilst the values used in the methodology are specific to the study area, the approach can be modified for use in other regions or habitats. The stages of the methodology are explained below and in Figure 3. 1. Select woodland patches of between 20 and 50 ha in size as woodland expansion ‘nuclei’. Medium-sized woods were targeted for expansion. The size range was chosen for two reasons. First, these sizes are intermediate between the extremes of less than 1 ha and more than 100 ha. Second, the 50 woods within the size range are a realistic number around which to target for woodland expansion given the desire for an additional 538 ha of woodland to be created.

Ancient woodland AONB boundary

Princes Risborough

High Wycombe

N Approx 5 km

Figure 2.

29

Ancient woodland patches within Buckinghamshire used in the study.

30

J. T. Lee et al. Stage 1 – Select woodland patches between 20 and 50 ha.

< 20 ha

20–50 ha

> 50 ha

Stage 2 – Buffer woodlands by various distances.

200 m

400 m

750 m

Stage 3 – Encapsulation of a minimum of 70 ha of woodland by the buffer. In this case, 75 ha of woodland are included (45 + 15 + 10 + 5). Woodland of 45 ha

Patch B = 10 ha

Patch A = 15 ha

Patch C = 5 ha

Stage 4 – Ranking of woodlands according to the number of additional patches required to create a minimum of 70 ha. Woodland A requires 2 patches and is ranked first, whereas woodland B needs 4 patches.

Woodland A

Figure 3.

Woodland B

Schematic diagram of the woodland conservation and expansion method.

2. Buffer each of the woodland nuclei by varying distances of 200 m, 300 m, 400 m, 500 m, 750 m and 1000 m to determine the proximity of each to other woodland patches. A buffer is a zone of specified width around an area. The range of buffer distances was chosen to reflect the range of inter-patch distances in the study area. The mean edge-to-edge distance of each woodland patch to all other woodlands was 658Ð1 m (SD = 291Ð1 m) and the minimum distance was 179Ð4 m. 3. For each woodland nucleus, record the buffer distance needed to encapsulate a minimum woodland area of 70 ha inclusive of the nuclei. This ensures that at least 70 ha of existing ancient woodland occur within the buffered matrix and only 30 ha need to be planted

with trees to make up the required 100 ha. As the buffer distance increases, the area of the woodland nucleus, together with the additional proximal woodland patches will reach the targeted 70 ha. This is the first stage of the restoration targeting since those nuclei that encapsulate 70 ha of woodland with the shortest buffer distance are those that are least isolated and therefore should be targeted first for restoration. 4. For each buffer distance, record the number of additional woodland patches that are required to make up the 70 ha target. Priority was given to those nuclei that needed the least additional patches to make up the required 70 ha. This identified those woodland nuclei that should be targeted in the second tier for restoration.

Using GIS to target habitat restoration

5. Identify suitable locations and land use types for restoration, based on their proximity to targeted woodlands. A land use database was used to target those parts of the landscape that offer the best opportunity for tree planting based on their position between woodland nuclei and proximal woodland patches. Three land uses were thought to be suitable for restoration to native woodland, namely: deciduous woodlands not listed as being either ancient or ancient semi-natural, arable land and improved, or agricultural, grassland. The rare and threatened calcareous grassland habitat was not considered for woodland planting because of its priority status in Annex I of the European Union’s Habitats and Species Directive (European Economic Community, 1992). 6. Target sites for tree planting based on their area. Potential new woodland was targeted interactively with the GIS to generate woodlands of more than 100 ha. Areas of at least 30 ha of suitable land use were identified to ensure that all expanded woodlands would be a minimum of 100 ha. Areas adjacent to roads or railways were avoided as these are deleterious to woodland species’ movement.

31

(a) 0

2 km

N

(b)

Results

>100

91–100

81–90

71–80

61–70

51–60

41–50

31–40

21–30

11–20

Figure 5. Ordnance Survey map tile SU79 showing patches of woodland prioritised for expansion( ) together with areas of other ancient woodland ( ) and other deciduous woodland( ). The area is shown (a) before and (b) after targeting of land for tree planting.

6–10

100 90 80 70 60 50 40 30 20 10 0

0–5

Number

Fifty woodland nuclei of between 20 and 50 ha were selected at the first stage and the targeting approach we devised would create 19 new woodlands with a minimum area of 100 ha. The linking of small woodlands into larger patches resulted in the initial 304 fragments being reduced to 255. The number of patches in all size categories below 90 ha decreased, particularly in the 20–50 ha range (Figure 4). This is to be expected as these woods

Size (ha)

Figure 4. Comparison of woodland patch size, before ( ) and after ( ) targeting of new woodland.

were prioritised for expansion. The restoration targeting identified 532 ha of potential new native woodland. This would involve improving 33 ha of existing deciduous woodland to link existing patches of ancient wood, planting trees on 383 ha of arable land and 117 ha on improved grassland. The extent of the original and expanded woodland in Ordnance Survey tile SU79 is shown in Figure 5.

32

J. T. Lee et al.

Discussion Geographical Information Systems have been widely used in nature conservation, particularly in the analysis of species’ habitat requirements (Young et al., 1987; Congalton et al., 1993; Austin et al., 1996; Radeloff et al., 1999). However, there are fewer examples of their usage in habitat expansion and restoration programmes and these are dominated by species-based approaches such as those involving large mammals (e.g. Mace et al., 1999; Mladenoff et al., 1999). Nonetheless, there are examples of habitat-based targeting of restoration measures in which the GIS are used to integrate, store and manipulate data from disparate sources (Veitch et al., 1995; Ranta et al., 1999; Wickham et al., 1999). The lowland beech and yew woodland HAP estimates that there are currently about 30 000 ha of this rare habitat in the UK (UK Biodiversity Group, 1998). The required 10% increase in this habitat could be achieved by random tree planting, yet to maximise the ecological benefit it is argued that a spatially targeted, regional-scale restoration offers the best approach. The targeting method used here is based upon the size and spatial position of existing native woodlands, which act as foci for woodland expansion through tree planting, and as such integrates principles of landscape ecology, woodland ecology and habitat restoration. The approach is beneficial for three major reasons. First, ecologically desirable larger patches of native habitat are created, which maximises biodiversity both through the eventual establishment of a large core area for woodland species and through the creation of greater variety of habitat in large patches. Second, both natural colonisation by woodland plants and faunal movements leading to metapopulation persistence are encouraged by targeting sites for restoration which are adjacent to existing woodlands. Third, the approach implicitly maximises woodland patch connectivity, by interpatch planting on suitable land use types. This aids species movement and hence colonising ability through the establishment of stepping-stones (Opdam et al., 1985). Connectivity can be further enhanced by judicious tree planting to utilise the existing network of hedgerows in the landscape. Additional benefits can be identified. Much of the ecological benefit provided within any habitat can be linked to its structural diversity. Using a rolling programme of tree planting, both species and structural diversity can be maximised within the newly created woodland patches. The GIS

is a valuable tool within this managed rolling programme, as it can be used to prioritise areas for planting both spatially, (in relation to the existing distribution of the beech woodland resource), and temporally (in relation to the age of the existing beech woodland resource). Careful site selection using the GIS can therefore provide a woodland reserve network of variable age structure, thus removing the monoculture appearance of previous woodland planting schemes. Similar benefits can be obtained if the GIS was used to select sites for planting which maximised the diversity of shape within the newly planted woodlands. Selecting those sites which provide a variety of edge-towoodland-area ratios can again help maximise the ecological interest within the whole of the reserve network. Using the GIS to select those potential planting sites which both follow natural contours and are adjacent to existing areas of beech woodland may enhance the landscape value by removing the hard edges which are often a feature of woodland planting schemes. This could be extended to exclude those areas of flat land currently intensively farmed, which would require high levels of compensation to encourage farmers to give their land over to tree planting. It is acknowledged that the success of newly planted trees depends more than upon selection of sites with suitable land use type and proximity to existing large patches of woodland. To ensure success in generating woodland of both high nature-conservation value and longevity, other factors are important such as soil type, altitude, slope, infrastructure and overall landscape character. For example, soil type and underlying geology are significant in dictating the type of woodland which evolves and the species assemblage (the wood may be dominated either by acidophilous or neutophilous species). Similarly, topography may be significant in determining the success of newly planted trees, although the altitude range in the study area is not extreme and presents no constraint to woodland.

Conclusion At a time when financial resources for nature conservation are extremely limited, a targeted method that maximises policy efficiency is essential. Geographical Information Systems enable spatial targeting by identifying those parts of the landscape for restoration that offer the best chance of increasing biodiversity. The method discussed here is beneficial for three reasons. First, it allows

Using GIS to target habitat restoration

multiple scenarios to be run to assess a variety of potential solutions. Secondly, it uses widely available datasets in a desktop, decision support system. Thirdly, it allows planners access to information ensuring that a more informed decision is made. Although there are some data, which would be valuable in further constraining woodland expansion by planting, these are expensive to acquire and may not add greatly to the process. Whilst the method described is both flexible and transparent, it is based upon heuristic, rather than robust scientific, principles. As such, it would be of most benefit as a tool for exploring options within a framework of objectives, rather than as a solution to a given policy; it is only a first step in the process of habitat enhancement. Nonetheless, the approach offers significant advantages in targeting land for woodland expansion and could be modified to suit local conditions. Although the study area is in the UK, the method has wider potential application for targeting restoration and creation measures, particularly in conjunction with agri-environment programmes.

Acknowledgements The authors would like to thank Gail Murray and Carol Hatton of WWF-UK for their help and advice, WWF-UK for part funding this research; Mary Webb of Oxford Brookes University for the aerial photography interpretation; Joanne Hodgkins at Buckinghamshire County Council for her help and advice with the species data and Corinna Woodall of English Nature for allowing access to valuable data. All maps are copyright the Ordnance Survey, UK and are used with permission.

References Abensperg-Traun, M. and Smith, G. T. (1999). How small is too small for small animals? Four terrestrial arthropod species in different-sized remnant woodlands in agricultural Western Australia. Biodiversity and Conservation 8(5), 709–726. Austin, G. E., Thomas, C. J., Houston, D. and Thompson, D. B. A. (1996). Predicting the spatial distribution of buzzard Buteo buteo nesting areas using Geographical Information Systems and remote sensing. Journal of Applied Ecology 33, 1541–1550. Bellamy, P. E., Hinsley, S. A. and Newton, I. (1996). Factors influencing bird species numbers in small woods in south-east England. Journal of Applied Ecology 33(2), 249–262. Brito, J. C., Godinho, R., Luis, C., Paulo, O. S. and Crespo, E. G. (1999). Management strategies for conservation of the lizard Lacerta schreiberi in Portugal. Biological Conservation 89(3), 311–319.

33

Brown, N. J., Swetnam, R. D., Treweek, J. R., Mountford, J. O., Caldow, R. W. G., Manchester, S. J., Stamp, T. R., Gowing, D. J. G., Soloman, D. R. and Armstrong, A. C. (1998). Issues in GIS development: adapting to research and policy-needs for management of wet grasslands in an Environmentally Sensitive Area. International Journal of Geographical Information Science 12(5), 465–478. Congalton, R. G., Stenback, J. M. and Barrett, R. H. (1993). Mapping deer suitability using remote sensing and Geographic Information Systems. Geocarto International 3, 23–33. Cook, H. F. and Norman, C. (1996). Targeting agrienvironmental policy: an analysis relating to the use of Geographical Information Systems. Land Use Policy 13(3), 217–228. Corbet, G. B. and Harris, S. (1991). The Handbook of British mammals, 3rd ed. Oxford: Blackwell Scientific (for the Mammal Society). Cramp, S., Perrins, C. M., Brooks, D. J. and Dunn, E. (1993). Handbook of the birds of Europe, the Middle East and North Africa: the birds of the Western Palearctic. Vol. 7: Flycatchers to shrikes. Oxford: Oxford University Press. Crowson, R. A. (1981). The biology of the Coleoptera. London: Academic Press. Department of the Environment (1994). Biodiversity: The UK Action Plan. London: HMSO. Department of the Environment (1995). Biodiversity: the UK Steering Group Report. Volume 2: Action Plans. London: HMSO. Dettmers, R. and Bart, J. (1999). A GIS modeling method applied to predicting forest songbird habitat. Ecological Applications 9(1), 152–163. English Nature (2001). http://www.english-nature.org.uk European Economic Community (1992). Council Directive on the Conservation of Natural Habitats and of Wild Fauna and Flora (‘The Habitats and Species Directive’) (92/43/EEC). Brussels, European Economic Community. Farino, T. (1986). Botanical assessment of Chiltern ancient woodlands in Buckinghamshire and Oxfordshire. Part 1: explanation. Peterborough: Nature Conservancy Council. Fitzgibbon, C. D. (1997). Small mammals in farm woodlands: the effects of habitat, isolation and surrounding land-use patterns. Journal of Applied Ecology 34, 530–539. Gilbert, O. L. and Anderson, P. (1998). Habitat creation and repair. Oxford: Oxford University Press. Harding, P., Lyon, T. and Tomblin, G. (1996). How to identify edible mushrooms. London: Harper Collins. Hodge, S. J. and Harmer, R. (1995). The creation of woodland habitats in urban and post-industrial environments. Land Contamination and Reclamation 3, 86–88. Lathrop, R. G. and Bognar, J. A. (1998). Applying GIS and landscape ecological principles to evaluate land conservation alternatives. Landscape and Urban Planning 41(1), 27–41. Laurance, W. F. (1997). Responses of mammals to rainforest fragmentation in tropical Queensland: a review and synthesis. Wildlife Research 24(5), 603–612. Mace, R. D., Waller, J. S., Manley, T. L., Ake, K. and Wittinger, W. T. (1999). Landscape evaluation of

34

J. T. Lee et al.

grizzly bear habitat in western Montana. Conservation Biology 13(2), 367–377. Macmillan, D. C., Harley, D. and Morrison, R. (1998). Cost-effectiveness analysis of woodland ecosystem restoration. Ecological Economics 27(3), 313–324. Ministry of Agriculture, Fisheries and Food (1999). CAP Reform Agreement: an information document. London: Ministry of Agriculture, Fisheries and Food. Mitchell-Jones, A. J. (1999). The Atlas of European mammals. London: T & A D Poyser. Mladenoff, D. J., Sickley, T. A. and Wydeven, A. P. (1999). Predicting gray wolf landscape recolonization: Logistic regression models vs. new field data. Ecological Applications 9(1), 37–44. Opdam, P., Rijsdijk, G. and Hustings, F. (1985). Bird communities in small woods in an agricultural landscape: effects of area and isolation. Biological Conservation 34, 333–352. Peterken, G. F. (1995). An overview of native woodland creation. In The ecology of woodland creation (R. Ferris-Kaan, ed.). Chichester: Wiley. Pollard, E. and Yates, T. J. (1993). Monitoring butterflies for ecology and conservation: the British butterfly monitoring scheme. London: Chapman & Hall. Pressey, R. L., Humphries, C. J. Margules, C. R., VaneWright, R. I. and Williams, P. H. (1993). Beyond opportunism–key principles for systematic reserve selection. Trends in Ecology and Evolution 8(4), 124–128. Radeloff, V. C., Pidgeon, A. M. and Hostert, P. (1999). Habitat and population modelling of roe deer using an interactive geographic information system. Ecological Modelling 114, 287–304. Ranta, P., Niemela, J., Blom, T., Joensuu, E. and Siitonen, M. (1999). The fragmented Atlantic rainforest of Brazil: size, shape and distribution of forest fragments. Biodiversity and Conservation 7, 385–403. Rodwell, J. S. (1991). British Plant Communities. Volume 1: woodlands and scrubs. Cambridge: Cambridge University Press. Russell, G. D., Hawkins, C. P. and O’Neill, M. P. (1997). The role of GIS in selecting sites for riparian restoration based on hydrology and land use. Restoration Ecology 5(4S), 56–68. Scott, J. M., Davis, F., Csuti, B., Noss, R., Butterfield, B., Groves, C., Anderson, H., Caicco, S., D’erchia, F., Edwards, T. C., Ulliman, J. and Wright, R. G. (1993). GAP analysis: a geographic approach to protection of biological diversity. Wildlife Monographs 123, 1–41. Sharrock, J. T. R. (1976). The atlas of breeding birds in Britain and Ireland. Berkhamsted: Poyser. Simberloff, D. and Gotelli, N. (1984). Effects of insularization on plant-species richness in the prairie-forest ecotone. Biological Conservation 29(1), 27–46. Smith, A. J. E. (1978). The moss flora of Britain and Ireland. Cambridge: Cambridge University Press.

Stace, C. A. (1997). New flora of the British Isles, 2nd ed. Cambridge: Cambridge University Press. Stoms, D. M. (2000). GAP management status and regional indicators of threats to biodiversity. Landscape Ecology 15(1), 21–33. Thompson, S., Larcom, A. and Lee, J. T. (1999). Restoring and enhancing rare and threatened habitats under agri-environment agreements: A case study of the Chiltern Hills Area of Outstanding Natural Beauty (AONB), UK. Land Use Policy 16(2), 93–105. UK Biodiversity Group (1998). Tranche 2 Action Plans: volume II–terrestrial and freshwater habitats. Peterborough: English Nature. Van Dorp, D. and Opdam, P. F. M. (1987). Effects of patch size, isolation and regional abundance on forest bird communities. Landscape Ecology 1(1), 59–73. Van Jaarsveld, A. S. Freitag, S., Chown, S. L., Muller, C., Koch, S., Hull, H., Bellamy, C., Kruger, M., EndrodyYounga, S., Mansell, M. W. and Scholtz, C. H. (1998). Biodiversity assessment and conservation strategies. Science 279, 2106–2108. Veitch, N., Webb, N. R. and Wyatt, B. K. (1995). The application of Geographic Information Systems and remotely sensed data to the conservation of heathland fragments. Biological Conservation 72, 91–97. Webster, S. and Felton, M. (1993). Targeting for nature conservation in agricultural policy. Land Use Policy 10(1), 67–82. Wickham, J. D., Jones, K. B. H. R. K., Wade, T. G. and O’Neill, R. V. (1999). Transitions in forest fragmentation: implications for restoration opportunities at regional scales. Landscape Ecology 14, 137–145. Wilson, G. A. (1997). Selective targeting in Environmentally Sensitive Areas: implications for farmers and the environment. Journal of Environmental Planning and Management 40(2), 199–215. Woodhouse, S., Lovett, A., Dolman, P. and Fuller, R. (2000). Using a GIS to select priority areas for conservation. Computers, Environment and Urban Systems 24, 79–93. Wynne, G., Avery, M., Campbell, L., Gubbay, S., Hawkswell, S., Juniper, T., King, M., Newbery, P., Smart, J., Steel, C., Stones, T., Stubbs, A., Taylor, J., Tydeman, C. and Wynde, R. (1995). Biodiversity Challenge. Sandy: RSPB. Yeo, M. J. M., Blackstock, T. H. and Stevens, D. P. (1998). The use of phytosociological data in conservation assessment: a case study of lowland grasslands in Wales. Biological Conservation 86, 125–138. Young, T. N., Eby, J. R., Allen, H. L., Hewitt III M. J. and Dixon, K. R. (1987). Wildlife habitat analysis using LANDSAT and radiotelemetry in a GIS application to spotted owl preference for old growth. GIS’ 87: Second Annual International Conference on Geographic Information Systems. San Francisco, 595–600.