Chris Gleed-Owen & Steve Langham. 21 March CGO Ecology Ltd Adder Status Project (England) Mar

Transcription

1 The Adder Status Project a conservation condition assessment of the adder (Vipera berus) in England, with recommendations for future monitoring and conservation policy Chris Gleed-Owen & Steve Langham 21 March 2012 CGO Ecology Ltd Adder Status Project (England) Mar

2 This document produced for: Dr Tony Gent Chief Executive Officer Amphibian and Reptile Conservation Trust 655A Christchurch Road Bournemouth Dorset BH1 4AP On behalf of: Paul Edgar Senior Environmental Specialist (Amphibians & Reptiles) Natural England 2nd Floor, Cromwell House 15 Andover Road Winchester Hampshire SO23 7BT Corresponding author: Dr Chris Gleed-Owen MIEEM Director and Principal Ecologist CGO Ecology Ltd 5 Cranbourne House 12 Knole Road Bournemouth Dorset BH1 4DQ CGO Ecology Ltd Adder Status Project (England) Mar

3 Note on sensitivity All of the data used in this analysis has been obtained with permission from data-holders, and much of it has conditions restricting its use. This report should not be circulated widely without permission, until sensitivity issues have been addressed and tailored for a wider audience. Please ask the corresponding author if in doubt. CGO Ecology Ltd Adder Status Project (England) Mar

4 Contents Executive summary 1. Background and rationale 1.1. Brief 1.2. Overview of adder status in England 1.3. Existing adder range data 1.4. Previous and ongoing status assessment exercises 1.5. Growing concern about adder declines 1.6. General approach 2. Conservation condition assessment 2.1. Background and context 2.2. Aims and rationale 2.3. Approach and methods CASM 2.0 model development GIS mapping 2.4. Data and assumptions Collation of adder range data Collation of population dynamics data Collation of land cover, geology, OS grid, and other data 2.5. Tools and techniques 2.6. Limitations and constraints 2.7. Results Range mapping Current conservation condition Historic loss of conservation condition 2.8. Deductions and relevance Conservation condition headlines Current adder distribution Loss of conservation condition at the national level Loss of occupancy within Vice Counties Loss of occupancy within Natural Character Areas CGO Ecology Ltd Adder Status Project (England) Mar

5 2.9. Recommendations for further work Further modelling Field testing results 3. The importance of hibernacula 3.1. Background and context 3.2. Aims and rationale 3.3. Approach and methods Questionnaire design Assumptions and limitations 3.4. Results 3.5. Deductions and relevance 3.6. Recommendations for further research 4. Management practices and factors affecting status 4.1. Background and context 4.2. Damage caused by habitat management and restoration 4.3. Damage caused by controlled burning 4.4. Other causes of adder decline and extinction Unexplained absence from some upland areas Cynical ploughing Pheasant-rearing Direct public pressure Genetic problems Mitigation failure 4.5. Examples of good management practice 4.6. Examples of bad management practice 5. Recommendations for conservation policy 5.1. Overview 5.2. Recommendations for research 5.3. Developing systems for monitoring adder status 5.4. Developing volunteer capacity 5.5. Recommendations for Natural England policy 6. Acknowledgements CGO Ecology Ltd Adder Status Project (England) Mar

6 7. References 8. Appendices 8.1. Excerpt from MoA between ARC and Natural England 8.2. Condition Assessment Model developed by Natural Acuity Ltd 8.3. Map of known current adder distribution in England 8.4. Map of known historic adder distribution in England 8.5. Map of predicted current adder distribution in England 8.6. Map of predicted historic adder distribution in England 8.7. Map of known current and historic adder distribution in England 8.8. Map of predicted current and historic adder distribution in England 8.9. Map of known current and historic adder distribution in Northern England Map of predicted current and historic adder distribution in Northern England Map of known current and historic adder distribution in Central England Map of predicted current and historic adder distribution in Central England Map of known current and historic adder distribution in Southwest England Map of predicted current and historic adder distribution in Southwest England Map of known current and historic adder distribution in Southeast England Map of predicted current and historic adder distribution in Southeast England Map showing paucity of adder records (predicted current and historic distribution) in the southern Pennines Map showing suspicious predictive patterns ( buffer syndrome ) in Herefordshire area Map showing forestry plantations in the Breckland area of East Anglia, illustrating apparent flaws in the predictive model Map showing predictive model working well in Dorset and the New Forest Table of CORINE Land Cover Types Table of component datasets included in current analysis Line graph showing Make The Adder Count data for 16 sites monitored from Questionnaire survey results: plant species associated with adder hibernation sites Questionnaire survey results: site features associated with adder hibernation Questionnaire survey results: other reptile species associated with adders Questionnaire survey results: incorporation of adders into site management plans Map showing the predicted current ( ) range of the adder in England Comparison between condition assessment based known survey data and predicted condition using the CASM 2.0 model to compensate for inadequacy of survey effort CGO Ecology Ltd Adder Status Project (England) Mar

7 8.30. Comparison between current and historic condition assessment for the adder in England Table showing monad loss between historical and current epochs per Vice County Table showing monad loss between historical and current epochs per Natural Character Area CGO Ecology Ltd Adder Status Project (England) Mar

8 8.33. Executive summary The adder is widely believed to be declining across England. This view is based mainly on anecdotal evidence and time-series data from individual sites, with no regional or national evidence to support it. Nevertheless, with a groundswell of opinion supporting the notion that conservation policy could be failing adders, there are sufficient grounds to investigate perceived declines and review conservation policy towards adders. Adder distribution is well-known at the hectad (10km) level across Britain, but a finer resolution is needed to measure occupancy and changes in occupancy at a meaningful level. This level of occurrence data exists in many areas, but is held by numerous organisations, and its coverage and quality is variable. Several initiatives have attempted to gather better data on adders at different spatial scales, but these are limited in their scope and/or sample size. Empirical adder data will always be insufficient to map occurrence comprehensively. New approaches are therefore needed to analyse known occurrence, and use proxy methods to predict likely occurrence. The Adder Status Project (ASP) draws together empirical data on adder occurrence, and through spatial analysis, geological and land cover correlation, models predicted occurrence during two epochs: present ( ) and historical (up to 2005). A mathematical model was constructed to perform a conservation condition assessment for the adder in England. This Condition Assessment Model (CAsM) uses statistical analysis across multiple datasets to generate a representation of adder occupancy that would be impossible to generate using traditional survey methods. By varying input data, this model can be used to measure changes in range, population size and habitat suitability across time, i.e. between a historical baseline and the present day. In common with all mathematical models, quality of CAsM output relies heavily upon the quality of input data. Whilst the CAsM algorithms are based upon sound statistical principles, the degree of verification of the model outputs has been constrained by the short timescale of this project. Early results from the model are in line with other objective analysis (NARRS), and subjective assessments of the putative decline in adder range and numbers. The ability to generate a national snapshot for the adder represents a significant advancement in our understanding of the adder s status in England. Recommendations are made for improving the model and the source data, and for testing its findings via field survey and further consultation. Consideration is given to factors affecting adder status, and ways of improving conservation policy are discussed. Further work is needed before policy and monitoring systems can be designed adequately. CGO Ecology Ltd Adder Status Project (England) Mar

9 1. Background and rationale 1.1. Brief As part of a Memorandum of Agreement (MoA) between Amphibian and Reptile Conservation (ARC) and Natural England (NE), entitled Developing Amphibian and Reptile Conservation in England, ARC has committed to delivering a project examining adder status across England. ARC instructed CGO Ecology Ltd to deliver this project on their behalf. The timescale was tight, with the MoA schedule excerpt received on 3/2/12, and a project deadline of mid-march 2012, six weeks later. The MoA schedule excerpt was treated as the brief from ARC (see Appendix 1). The required actions are as follows: Undertake a research project to evaluate the national status of the adder and develop monitoring metrics and protocols, involving: Collation of existing data and data sources; Further analysis of existing data and data from ongoing projects; Develop dialogue, perhaps via meetings/workshops involving a small number of key players ; Map of actual distribution plus link to NCA; Predictive mapping exploring different models to define current status. The project was a collaborative effort between CGO Ecology Ltd and Natural Acuity Ltd who delivered the data analysis and GIS modelling. NE s Senior Environmental Specialist (Amphibians & Reptiles) (Paul Edgar) provided guidance. Various individuals with adder expertise were consulted, and numerous organisations were approached for assistance. The project is referred to hereafter as the Adder Status Project (ASP). As the project was executed in a very tight timeframe, it must be stressed that the outcomes are works in progress, and that the many constraints prevented a comprehensive solution to all aspects of the brief. However; we believe that the results to date show what can be achieved, and that the approach warrants further development Overview of adder status in England Adders have always been widespread across Britain, albeit with a patchy distribution, especially in the English Midlands. Unlike grass snakes, they are not a vagile (mobile) species; and being tied to specific habitat patches, they are easily lost from a changing landscape. Even when suitable habitat is created, adders are less able than other reptiles to colonise new areas. At the local level, they may occupy a fraction of the habitat available to them, and seem more sensitive to change than the other reptile species. This under-occupation is generally a result of historical declines and local extinctions, but also implies that adders have a low ecological plasticity. CGO Ecology Ltd Adder Status Project (England) Mar

10 Adders have declined significantly over the last century in Britain. The causes are primarily habitat loss, fragmentation and degradation of various sorts, especially due to competing land uses in the latter part of the twentieth century, but also due to persecution (Cooke & Arnold, 1982; Cooke & Scorgie, 1983; Hilton-Brown & Oldham, 1991; Reading et al., 2004). Adders have always been scarce in some English counties, and in those with the fewest populations, may now be extirpated. Conversely, in some parts of England, adders seem to be thriving; especially in certain upland areas. They also seem to be widespread and common in many areas of Scotland and Wales (Arnold, 1995). As well as habitat loss, adders have suffered the additional problem of centuries of persecution in areas where they come into contact with human populations. The worst effects have probably been experienced on grouse moors and other game-keeping areas where gamekeepers traditionally kill adders; but as Britain s only venomous snake, adders still generate fear and loathing in many people. The Wildlife & Countryside Act 1981 (as amended 1988, 1991) now affords adders legal protection from deliberate killing and injury. Nevertheless, their habitat remains unprotected (unless incidentally protected for other reasons), and adders are often killed as an incidental result of other activities. One of the biggest concerns today, among professional and amateur adder experts alike, is the lack of consideration given to adders in habitat management practices. Large tracts of seminatural land across England are managed proactively for nature conservation, and yet adders (like other reptiles) are rarely incorporated into the design of such landscapes. Moreover, adders are sensitive to local habitat change, and seem to be suffering more than the other reptile species from land management practices. In particular, damage to winter hibernation areas can be fatal to individuals, and even populations. With many English populations comprising fewer than ten adders (Baker et al., 2004), the loss of any individuals can be a significant impact at the population level Existing adder distribution data Although early accounts of the adder s occurrence in Britain exist (Leighton, 1901), the first attempts to map national reptile distribution in Britain were those of RHR Taylor, half a century later. Taylor collated literature and anecdotes, using them to map herpetofauna distributions at a crude Vice County level initially (Taylor, 1948), and later as dot maps (Taylor, 1963). Malcolm Smith s (1951) book included maps too. It is clear from these older accounts, and from local treatments, that some parts of Britain have never been well-endowed with reptiles. Areas will clayey soils and lowland agricultural landscapes, such as Warwickshire, Cambridgeshire and Huntingdonshire, seem to have always been largely devoid of adders. The sparse adder populations in such counties tend to be isolated and inherently vulnerable. The first modern grid-based atlas was produced by the Biological Records Centre (BRC) following a period of pro-active data solicitation (Arnold, 1973). Arnold mapped records at the hectad (10km) level, and split data into pre- and post-1960 categories (no information was given on the number of records). This gave the first accurate graphic depiction of adder distribution patterns across Britain. The distribution of adders was clearly patchy, with bold concentrations of post-1960 black dots in some areas (e.g. the south coast and East Anglia), and distinct clusters of pre-1960 open circles in others (e.g. central and northern England). CGO Ecology Ltd Adder Status Project (England) Mar

11 The hectad approach hides local variation in recording density, but Arnold s map gave a fairly convincing impression of a heterogeneous overall distribution. It also did not appear to be biased by human population centres; in fact, far from it. Furthermore, whilst a schematic map can be misleading if sample sizes are uneven in each category, the clustering of open circles in some places (with no modern records) gave a convincing impression of decline. Arnold s (1973) maps were later included in Frazer s (1983) book, but a full revision of the data was not published two decades later. After much new data collection, data cleaning and other improvements, the BRC published a revised atlas (Arnold, 1995). Recording had been ramped up in some areas, as the overall number of records greatly increased. A total of 3354 adder records were now included, covering 1014 hectads across Britain. The data was now divided into three categories to depict the most recent record from each hectad: pre-1900 (small open circles), (open circles), and 1970 onwards (black dots). Clearly, the data-gathering campaign had had the effect of filling gaps, and some of the previously-apparent clusters of extinctions were now reduced in size. Arnold s (1995) atlas has yet to be superseded in print, and recent general treatments of the British herpetofauna (Beebee & Griffiths, 2000; Inns, 2009) have only included low-resolution distribution maps. In recent years the internet has challenged print as the medium of choice for disseminating biological records. The National Biodiversity Network (NBN) Gateway ( now provides free access via the internet to data from the BRC, ARC, other wildlife entities, and numerous Local Records Centres (LRCs). Many other data portals have sprung up too, including ARC s Add an Adder campaign dedicated to collecting adder distribution records ( The rise of the internet brings its own problems however, including difficulties in verification. Other data-gathering exercises such as ARC s rare species monitoring programmes have gathered large amounts of adder data in some parts of the country, and the NARRS and Make the Adder Count surveys have attempted systematic survey on a more widespread basis. At a local level, initiatives coordinated by voluntary amphibian and reptile groups (ARGs) have sometimes yielded significant volumes of adder data recently, for example in Surrey and Kent. In terms of published local accounts, many have been published over the last century. The last fifteen years has seen a flurry of county atlases, including Kent (Philp, 1998), Nottinghamshire (Wright et al., 2004), Surrey (Wycherley & Anstis, 2001), Sussex (Barker & Elliott, 2001), and Warwickshire (Clemons, 1996) Previous and ongoing status assessment exercises Status is the state or position something is in, usually with reference to some other point(s). In nature conservation, it is used as a general term meaning how well something is doing. Hilton- Brown and Oldham (1991) defined it as distribution and abundance, but in an increasingly target-driven environment, terms such as conservation status and favourable conservation status are widely used. These originate from the EC Habitats Directive 1992, and combine measures of species range, population, habitat availability, and future prospects. A subtle departure from conservation status is conservation condition, which comprises range, population and habitat status (without the prospects element). The trajectory of these factors is what defines how well a species is doing at any given point. The future prospects can (theoretically) be influenced in response to a conservation condition assessment. CGO Ecology Ltd Adder Status Project (England) Mar

12 Assessments of adder status have been attempted in the past, but they have all been lacking in numerical detail, and some have been rather cursory. The distribution-mapping exercises described above form the basis for any status assessment; and the BRC atlases (Arnold, 1973, 1995) provide an excellent starting point for examining clues to adder declines. Additional quantitative and qualitative data are needed, as well as intellectual examination of the issues affecting status, and development of suitable measurables or metrics. Taylor (1963) noted the scarcity of adders in central England, identifying urbanisation and agricultural intensification as the primary causes. Prestt et al. (1974) reviewed changes in British reptile status in the 1950s and 1960s. Again, adders were identified as scarce in central parts of England, and habitat loss was asserted as the key problem. Cooke and Arnold (1982) addressed the status of the widespread species, discussing apparent declines in relation to habitat loss since the 1930s up to the early 1970s. The West Midlands, East Anglia were among areas apparently experiencing adder declines. Cooke and Scorgie (1983) carried out a questionnaire-based survey of perceived abundance and distribution for the widespread species, on behalf of the Nature Conservancy Council (NCC). Responses were divided into NCC regions, with some adjustments aimed at equalising coverage. The adder was deemed widespread and common overall; common in southwest and northeast England, rare in central and northwest England. They calculated an index of change, which was negative for almost all regions, particularly the West Midland and East Anglia. Habitat loss was the worst offender, followed by persecution. Only two respondents cited cases of adder populations increasing. Interestingly, Cooke and Scorgie s summary of regional changes in status of the reptiles maps reported adder declines in only three out of eight English regions; with no change in all other regions. Grass snake and common lizard declined in six regions, and slow-worm was thought to be declining in seven regions. In other words, the slow-worm was perceived as the most threatened widespread reptile; and the adder the least. An almost identical exercise was carried out by Hilton-Brown and Oldham (1991), this time based on Vice Counties. They found that adders were still widespread and common in the Southwest, but scarce or rare in other parts of England. Participants reported declines across eastern and northwest England. Habitat loss due to forestry and public pressure was the top reason, followed by agriculture; persecution did not feature highly. In this study, the adder was thought to be declining in six out of eight English regions; with only West Midlands and the Southwest exhibiting no change. Comparison of Cooke and Scorgie s (1983) results with the subsequent Hilton-Brown and Oldham (1991) results is interesting. Both studies only depicted the adder as widespread and common in the Southwest. In all other regions, adders were less than common. Three regions depicted adder status differently in the later study: the Northeast was widespread but not common (rather than widespread and common); East Anglia was scarce (rather than widespread); and the West Midlands were widespread (not scarce). Despite this apparent improvement in the West Midlands since the previous study, participants in the later study recorded no change there. Questionnaires are subjective by nature, and some of the differences appear to reflect observer bias. Nevertheless, whereas the adder was declining in three English regions in the earlier study, they were now declining in six regions; a dramatic change. The first national survey of first-hand field sightings was coordinated by Leicester Polytechnic from (Swan & Oldham, 1993). It did not prescribe survey methods, but involved standardised recording cards. Geographical coverage of recorders and survey effort was variable, but the survey generated 817 new adder records, from 44% of previously-recorded hectads in Britain. Once again, adders were observed to be much sparser in the English CGO Ecology Ltd Adder Status Project (England) Mar

13 Midlands, and most common in areas of low human density. The results were presented as hectad dot maps, and contributed significantly to Arnold s (1995) BRC atlas. Nine British counties (not Vice Counties) did not yield adder records. Most records were from counties with upland moorland and lowland heathland habitats (Swan & Oldham, 1993). They also demonstrated that the paucity of records from the southern English Midlands could not have been an artefact of recorder density. Three-quarters of adder records (74%) were associated with woodland, heathland and moorland (Swan & Oldham 1993). In 2003, English Nature commissioned a questionnaire survey to identify long-term surveillance data for the adder and slow-worm, that might be used to assess their status (Baker et al, 2004). It collected information on 172 adder sites, of which about half were known to the recorder for over 15 years, and a third constituted systematic surveillance. About 40% of respondents were unable to estimate population size, but there was sufficient evidence to suggest a recent national decline in adder populations. Interestingly, non-systematic surveillance results concurred with systematically-collected data. Most worrying was the small size of many populations (a third contained fewer than 10 adults), including those with the greatest declines. Public pressure and persecution were perceived by respondents as negative impacts; isolation and habitat damage were also linked to population declines. Rates of decline were not measured. The report was limited by a disparity between types of data and collection methodologies, and by the unrepresentative site sample, but it does raise valid concerns, particularly for the adder. The first field-based status-monitoring scheme for adders - Make The Adder Count (MTAC) - was initiated in Surveyors record peak adder counts for around a hundred sites each spring, thereby collecting data on population trends and associated habitat factors. The sample size is too small to detect national changes, but effective at tracking local trends and influencing land managers. Now in its eighth year, and after a lull in coordination, MTAC is being re-launched. Preliminary analysis suggests that there is no obvious national trend when data are averaged, but individual sites are much more variable (J. Baker & J. Cranfield, pers. comms.). Summary data are presented in the appendices. In 2007, the National Amphibian and Reptile Recording Scheme (NARRS) was launched. It trains volunteers to survey randomly-selected monads (1km squares) across the UK, using standard methodologies and effort. With sufficient sample size, it could generate useful occupancy rates for individual species. After five years, reptile surveys have been performed on over 300 monads; with a target of 400 by the end of 2012 (the end of a six-year EC reporting cycle). Wilkinson and Arnell (2011) presented the results of the first three years ( ), in which adders were recorded from 7% of monads (15). They generally lived on isolated patches rather than large continuous blocks; and few sites were on nature reserves or protected land. Several current research projects are examining various aspects of adder ecology and status that should provide useful insights over the next few years. The Institute of Zoology s ongoing genetic work has already been mentioned. Research at DICE (University of Kent) is researching reptile detectability to improve the effectiveness of NARRS and other monitoring schemes. Current doctorate research at Manchester Metropolitan University is researching spatial habitat associations of the adder, by comparing analysis methods for identifying adder habitat associations at different spatial scales in the Peak District, Derbyshire. It will construct a spatial CGO Ecology Ltd Adder Status Project (England) Mar

14 model, using detailed ecological, physical and biological habitat data, to predict adder habitat and test it against adder presence/absence at various sites (D. Carter, pers. comm.). Ongoing research at Reaseheath College is estimating population size by capture-markrecapture methods (K. Palmer, pers. comm.) Growing concern about adder declines The last decade or so has seen a groundswell of opinion about adder declines. Compared to the rare species (sand lizard and smooth snake), adders have never justified conservation concern at a national level; but increasingly, alarm bells have been ringing. Some of the first evidence came from Sylvia Sheldon s (then unpublished) time-series data in the Wyre Forest, Shropshire, showing a severe decline in adder counts from around 250 in the early 1990s to 120 in When an estimated number of undetected individuals were included, the total population was thought to have dropped from 350 to fewer than 200 (Sheldon, 2011). Sheldon attributed the declines to housing development, forestry operations, infrastructure work, pheasant rearing, and several other factors. Heavy plant movement and damage to the ground where adders were hibernating was a major problem: spoil deposited on active adder hibernacula has buried many adders alive. She was also among the first to identify conservation management as a deleterious influence: Conservation work for other species has in some cases destroyed adder hibernacula. More liaison is required before embarking upon such projects. (Sheldon, 2011). These words still ring true today. In the following years, similar stories of decline and hibernacula damage were voiced by Tony Phelps, and by many other individuals. Their reports have generally been unpublished and anecdotal, but the consensus is that adders have been quietly disappearing from many sites. In addition to probable absence from Hertfordshire and Huntingdonshire, the adder may now be extirpated from its last site in Nottinghamshire. Surveys in 2011 and 2012 have found nothing (Gleed-Owen et al., unpublished). Adders are also thought to be recently extinct from Warwickshire (J. Clemons, pers. comm.), and virtually extinct from Cambridgeshire, Greater London, Lancashire and Oxfordshire. Reliable time-series datasets such as Sheldon s are rare however, and the last decade has seen a growing realisation that this gap must be filled. Long-term monitoring is needed in order to confirm the declines that many people believe are occurring, and to convince decision-makers that action is needed. Baker et al. s (2004) questionnaire survey identified other time-series datasets, and concluded that many recorders were observing decline in their local adder populations. The small size of and vulnerability of many populations was also worrying. The most important outcome of Baker et al. s study was the initiation of MTAC in One of its key aims is to raise awareness about adders among land managers, and to influence their management of adder hibernation areas in particular. As well as ARC s survey schemes, new initiatives have kept the momentum going: Surrey Amphibian & Reptile Group (SARG) s The Adder Project and The South East Regional Adder Project, and Kent Reptile and Amphibian Group (KRAG) s Adders in Decline Project,. Most recently, a conference dedicated to adder conservation was organised by KRAG in Chatham in December It resulted in near-unanimous agreement that the adder was the British herpetofauna species most in need of additional conservation assistance. Despite the sand lizard and smooth snake being much rarer, these species are probably improving in status now, whereas the adder is widely believed to be declining nationally. CGO Ecology Ltd Adder Status Project (England) Mar

15 The Chatham conference reiterated the need to get the message across to land managers that adder hibernation sites must be protected. It also precipitated a parliamentary question that lead indirectly to the initiation of the current project. Although controlled burning and other damaging activities are still widespread, outright destruction of adder habitat for development is relatively unusual these days. And where it does take place, mitigation measures must be employed to minimise the impacts (albeit sometimes questionably). Developers must avoid killing reptiles where possible (most legal offences pertain to individual animals), and it is commonplace to rescue animals prior to destructive works (however temporary), and ensure no-net-loss of occupied habitat. Temporary destruction of vegetation cover, for the purposes of creating another type of vegetation cover, seems to circumvent these procedures however. Habitat restoration practitioners do not rescue animals prior to such works, nor do they mitigate the impacts on reptiles in many cases. This contradiction seems unfair to many, and is counterproductive to relationships between stakeholders, let alone a cause of avoidable damage to adder status (and other protected species). It might be unsustainable to carry out the same intensive capture programmes that developers have to; but some level of mitigation must be achievable. The argument often asserted is that temporary loss is a price worth paying to achieve overall gain. Leaving the shaky legal ground aside, this assertion does not necessarily hold true for reptiles. Temporarily killing adders does not make them come back in greater numbers later. Furthermore, the argument than the end justifies the means does not obviate a practitioner s legal and moral responsibilities. Risk-avoidance measures should be employed as a component of all restoration works; and adders, now more than ever, should receive special consideration. General statements on how restoration works will benefit adders are meaningless without input from someone with local reptile expertise. Text-book prescriptions and guidelines are no substitute for specific knowledge of the adders on a site. In this respect, it can be invaluable to engage the advice of a reptile adviser, for example from a local ARG. There are many cases where this works well already, and alleviates the bad press that conservation practitioners are increasingly receiving. A reptile adviser can help a land manager to assess potential risks to reptile populations, and tailor management and restoration works to minimise risks. In particular, specific locations known to be used by hibernating adders should be marked for retention, to avoid killing, injury or damage to hibernacula. Simply by marking out individual bushes or localised areas of ground, contractors can be instructed to avoid them. In many cases, negative impacts would be mitigated, and sometimes even entire adder populations saved. To some, this strategy might seem precious; and that saving individual adders is not practical in the bigger picture; yet with many populations down to single figures, any losses could contribute to local extinction. Adders are already localised on many suitable landscapes such as the New Forest, and sometimes occupy only a fraction of large nature reserves. This impaired ability for recovery has led some observers to invoke genetic hypotheses, such as inbreeding depression. A current study by the Institute of Zoology (Zoological Society of London) and other partners is examining this issue. Ironically this contrasts with our rarest reptile the smooth snake, which is almost ubiquitous on some southern heathlands. Hence it is reasonable to expect that existing management practices could at least make allowances for adders, before they decline so much that statutory means are their only saviour. CGO Ecology Ltd Adder Status Project (England) Mar

16 Anecdotal evidence has not systematically been gathered to assess the extent of hibernaculum damage issues, but there is a growing groundswell of opinion that something needs to be done to reduce the putative impacts. Too often, the people with local knowledge of adder populations are not in contact with the relevant practitioners; or if they are, their concerns are overlooked. Despite many admirable exceptions, land managers can be defensive and unreceptive. They may feel (quite justifiably) that their job is difficult enough already, without listening to single-interest groups. Nevertheless, habitat management must involve reconciliation of conflicting interests and prescriptions a difficult task when confronted with contradictory guidance from statutory agencies, funding bodies, powerful lobbies, and decision-makers. Even when land managers are sympathetic to adders, there are few mechanisms for incorporating them into existing management plans and strategies. Rather than formalising adder-sensitive policies, it is might therefore be more realistic to employ friendly diplomacy on a one-on-one basis to receptive individuals. Unfortunately this could never solve a national problem General approach The ASP project examines the following work streams: Conservation condition assessment The importance of hibernacula Management practices and factors affecting adder status Monitoring systems and policy improvement. CGO Ecology Ltd Adder Status Project (England) Mar

17 2. Conservation condition assessment 2.1. Background and context As defined by the brief, one of the key aims of ASP was to make a provisional assessment of the status of the adder in England overall, and in relation to National Character Areas. This is necessarily an interim assessment, owing to the tight timeframe of the project, and the limitations of the available data. Our understanding of the situation facing the adder at present in England can only be informed through a robust assessment of its conservation condition. This can then be used to inform conservation policy, and provide a baseline for monitoring the future trajectory of adder status Aims and rationale The current project aims to measure the current conservation condition of the adder in England. These are necessary precursors for developing appropriate surveillance systems to monitor changes in condition over time, and make recommendations for conservation policy. Conservation condition is a component of conservation status. Conservation status is defined by the EU Habitats Directive as comprising four factors: population dynamics, range, suitable habitat, and future prospects for a species. There is a paucity of guidance concerning the appropriate metrics for each factor, and this decision is left to the relevant expert taxonomic group or reporting lead. The future prospects factor is somewhat different from the other factors, as it is not directly observable or derivable from field data. The three determinable factors can be grouped together to define the conservation condition of a species. Current conservation condition can be assessed independently of future prospects, and compared against targets if set. Conservation Status = Conservation Condition Assessment + Future Prospects Favourable Reference Value = Conservation Condition Target + Future Prospect Goals 2.3. Approach and methods To make a conservation condition assessment, we have employed statistical analysis of multiple datasets; and generated visualisations of the data using an open GIS platform, for the purposes of presentation and error checking CASM 2.0 model development The Conservation Condition Assessment Model (CASM) is an information management, analysis and exploitation system. It is used to determine the likely conservation condition for a species, using the assumption that available survey and distribution data will never be comprehensive. CASM can be used to calculate the likely trend in conservation condition for a given species across time. CASM 2.0 is an iterative development from the CASM 1.0 model, which was used for the assessment of the conservation condition assessment of all native reptile and amphibian species found within the county of Surrey (Langham, 2010). CGO Ecology Ltd Adder Status Project (England) Mar

18 CASM uses a federation of Microsoft Access databases as the backbone of its information management system, although any database platform could be used. These databases are a combination of information laid out in structured tables; queries that assist in the structuring of the data; forms for selecting variables such as geographic extent and species of interest; reports for outputting analysis; and code modules for the logical sequencing of analytical steps. The CASM 2.0 federation of databases are currently co-located on a single computer. However, the architecture of the model allows for remote compilation of data sources, such that third-parties could be tasked with ongoing information management, and CASM 2.0 could remotely access this data, provided that sufficient connectivity was available GIS mapping The simplest presentation tool for data with a spatial dimension is to present it on a map. With digital mapping freely available via the internet (including good-quality aerial photography), this platform seems the most sensible for wider dissemination. Traditional GIS platforms are more versatile but frequently require dedicated and expensive software. By generating Keyhole Markup Language (KML) files, CASM output can be presented and viewed using web-based GIS platforms such as Google Maps, Google Earth, and Bing Maps, in addition to traditional GIS platforms Data and assumptions All the potentially-available datasets were considered, and following collation of those available to us, the most appropriate data were used. The basis for our analysis was adder occurrence data ( biological records ). We opted for basing our model on a monad (1km square) level of resolution. This precluded data with lower-resolution spatial references Collation of adder range data A major task was to collate adder range and population data. As discussed earlier, there is a large volume of existing distribution data, and yet it is insufficient to fulfil the role to which it is being employed here. Not all adder populations are known, not all have been recorded, and no single dataset contains all the available data. Adder occurrence data is typically patchy in geographical coverage, due to a combination of presumed bias towards human population density and real spatial variation in adder distribution. The measurement of species range and population dynamics relies upon empirical collection of biological data. Range data is composed of biological records, the definition of which is widely accepted to be a sighting or other definite record of a species occurrence at a specific time and place. A legitimate biological record must comprise four basic components: species name, location, date, and recorder. Most biological data contains more parameters than this, but these are the most important. In the current analysis, data were collated from as many sources as possible to gain maximum coverage for the whole of England. There are many constraints with this, especially due to the short timeframe. We approached the NBN Gateway, and all of the individual LRCs. We also circulated a request to all individuals ARGs, via ARG UK, requesting access to any additional data, and assistance in verifying some datasets. CGO Ecology Ltd Adder Status Project (England) Mar

19 We attempted to collate data for all reptile species, as this would help us to identify false negatives, and to calculate likely absence probabilities for monads where the occurrence of other reptile species was sufficiently high and without adder records. The data fields we required were: reptile species, year of record, monad (1km square), and verification status. We treated all datasets received from the NBN Gateway and LRCs as verified, although we realise this is not strictly possible. The alternative would be to exclude data, as there was insufficient time available for a comprehensive verification review. We selected the monad as the spatial unit of mapping, modelling and analysis. This gives a hundred times greater resolution than the hectad (10km square), and allows a much greater precision than has hitherto been attempted. The benefits for distribution mapping, analysis and predictive mapping are yet to be fully exploited, but the monad approach is clearly achievable. Note that the permissions obtained for some of the component datasets restrict the audience and publication of this report. It should not be circulated widely without permission. The table in appendix 8.22 lists the component datasets which yielded over 100,000 unique reptile sighting records Collation of population dynamics data Population dynamics data for the adder is scarce. MTAC is the only national monitoring project that gathers numerical adder population data from many sites. Although around 270 sites are contained within the MTAC dataset, many are not monitored consistently year on year. Fewer than a hundred are monitored each year on average (J. Baker, pers. comm.), and only 17 sites have continuous records from 2005 to 2011 (L. Cranfield, pers. comm.). The line graph in appendix 8.23 shows these (with one anomalous site removed). It is evident that MTAC and other local datasets (e.g. Sylvia Sheldon s Wyre Forest data) are reliable at mapping local trends in adder population size. Their scalability, and fitness for monitoring the trajectory of adder populations at a national scale, are more questionable. It remains to be seen whether this is a sample-size issue, and whether or not an increased sample size would help. MTAC population dynamics data was not directly included in the current analysis. Range change data was identified as more useful proxy measure of population dynamics at a national scale; although local data is discussed later Collation of land cover, geology, OS grid, and other data These datasets were accessed by Natural Acuity Ltd. The Land Cover module contains various land cover datasets covering the British Isles Currently these tables comprise the CORINE 2000 Land Cover and CORINE 2006 Land Cover datasets. Which have been reformatted from seamless vector data into a monad grid format. For each monad, it provides the area for each of 44 land cover types (see appendix 8.21). Only the cover for England was utilised in the current exercise. The CASM model currently utilises the Ordnance Survey s OSGB36 CRS (covering the British Isles), but could be adapted to include the entire European Union. The Geology module holds data from the British Geological Survey. For each monad across the United Kingdom, both bedrock ( solid ) and superficial ( drift ) geologies are recorded. This geology database is effectively a proxy for soil types in most cases (to be added to CASM in future developments). Where superficial drift exists, it forms the surface geology, and CGO Ecology Ltd Adder Status Project (England) Mar

20 determines the nature of the overlying habitats. Where superficial drift is absent, the bedrock is the surface geology, and determines the habitats. The Area Template database holds a variety of spatial templates, such as the monad list for the country of England, monad lists for the Watsonian Vice Counties (VCs), and England s National Character Areas (NCAs). Templates can be used to define the area of analysis, or to categorise results by sub-areas (NCAs and VCs). Other datasets that could be employed in future modelling, but which are not incorporated at present, are climate, elevation and aspect data Assumptions and data manipulation Occurrence data was gathered for all reptile species, most of it ostensibly verified, and therefore directly useful for mapping adder distribution in relation to other reptile species. It was assumed that the verification was reliable, and therefore the data were accurate; but as discussed later, there are thought to be some errors within some datasets. Two of the datasets were unverified, and as such were only used to map the distribution of reptile recording effort, not adder occurrence. Some reptile data were excluded from the analysis as they did not have the required spatial resolution (1km accuracy). All records were normalised to monad resolution. Some records were given an improved spatial resolution, where the site descriptor(s) allowed selection of a specific monad rather than the 10km resolution given in the source records. Some records had tetrad (2km) resolution, which were converted to monad using the southwestern monad within the tetrad. It was assumed that this was not a significant bias. The selection of two epochs ( current and historical ) is arbitrary but sensible, according to the data available. The current epoch s length of six year ( ) was selected to meet the EU reporting timetable, but it appears to work well as it provides sufficient records to derive significant statistical correlation; whereas constraining to an individual year may not. In construction of the CASM predictive model, we assumed that known adder presence could be used to imply likely adder presence in neighbouring squares up to 2km away, provided that the land cover and geology were deemed to be suitable. The figure of 2km was derived from the spatial analysis of the Surrey adder dataset, and as such it is likely to be a conservative assumption when applied to areas where survey may be less comprehensive. The model was allowed to make its own assessment of which land covers and geology were associated with adder presence, and use these to inform predicted presence. We used insights from The Adder Project in Surrey to inform our national population estimate element of adders. In Surrey, adders typically occupy only a few hectares in each occupied monad; with an estimated population of 24 adders per monad. We used this figure as a baseline assumption for our national population calculation Tools and techniques CASM outputs are in two forms: a Species Analysis Report, and GeoFiles. A Species Analysis Report is the direct model output, in the form of a run-log file. It calculates the number of monads known to be occupied in each reporting epoch; the number of monads predicted to be occupied; and correlations to land cover types and geology. CASM uses statistical analysis approaches to generate an estimate of monad occupancy across an area. The benefit is that it defines explicitly those monads (or other area units) that are likely to be occupied. CGO Ecology Ltd Adder Status Project (England) Mar

21 KML geofiles are generated automatically by the model, which can be viewed in Google Earth, Google Maps, Bing Maps, or other open-source or bespoke GIS viewers. No bespoke software is necessary to view the precise geo-distributions generated by the model. Due to the statistical significance tests and correlation algorithms used, datasets for the same characteristics (e.g. Land Cover) do not need to be consistent in units nor type for different epochs. The model adapts to whatever data format is presented Limitations and constraints There are insufficient land cover datasets to accurately depict trends over time. For instance, comprehensive land cover data is not available prior to 1990; hence analysis for historical periods must rely upon recent land cover data. As semi-natural habitats have generally declined historically, the likely effect is an underestimation of the predicted baseline (historical) distribution. As with all models, the quality of output relies heavily on the quality of input data. Although great effort has been applied to obtaining and verifying reptile distribution datasets, the ASP study has been conducted under challenging time constraints. The model could be improved if more time were available to obtain more comprehensive and verified input datasets. We assumed that known adder presence could be used to imply likely adder presence up to 2km away, but a larger buffer (e.g. 5km?) may have been justifiable in some under-recorded areas, although it would run the risk of over-prediction in well-recorded areas. The model s assessment of land cover importance is also apparently biased towards land covers found in the most heavily-recorded areas. This means that the model spuriously values habitats such as urban areas, because urban areas are present close to most areas where people see adders. Conversely, the model underplays the value of habitats in more remote parts of the country, and causes an under-prediction effect on some habitats such as moorland and forestry Results The results are presented as GIS maps, report headlines and tables in the appendices Range mapping The results of the data collation and predictive mapping exercise are presented in appendices 8.3 to 8.20, as GIS maps using Keyhole Markup Language (KML) files on a Google Map aerial photography platform. The occupancy data are presented for the whole of England as four types: Known current distribution ( ) Known historic distribution (pre-2006) Predicted current distribution ( ) Predicted historic distribution (pre-2006) The known datasets are the actual empirical data that have been gathered and visualised. For each epoch, the predicted distribution contains the known data, plus a buffer of neighbouring squares with suitable habitat up to 2km away where monads are deemed suitable in terms of land cover and geology (based upon the CASM model analysis). CGO Ecology Ltd Adder Status Project (England) Mar

22 One constraint that became apparent during development of the project is that some of supposedly-verified data from local data providers appears to be flawed. In a few counties where adders are known to be very rare or extinct, there are supposed adder records that are probably spurious. Further verification would therefore be needed to improve these component data in future. The net results are not significantly affected overall, but the apparent occupancy of some Vice Counties is probably spurious Current conservation condition CASM s outputs yielded the following summary data for adders in the whole of England, using data from the current epoch ( ). Population Occupied Monads 9,237 Population Estimate 221,700 Habitat Suitable Monads 31,894 Range Occupied Vice Counties 58 (all) Full condition assessment results, and a comparison between known-data and predicted-data methods, and current and historical epochs, are presented in appendix 8.29 and This gives a comparison between condition assessment based on known survey data, and that based on predicted occurrence using the CASM 2.0 model, using the latter to compensate for the inadequacy of empirical data. The historic loss of adders from England is set out in detail here for the first time. There is no baseline assessment with which to compare the results of the current exercise directly, as no previous status investigations have produced comparable data. Broad comparisons are possible, but as acknowledged by all previous workers, those attempts were based upon questionnaire data rather than empirically-based data Historic loss of conservation condition The best means of measuring change in conservation condition using the currently-available data is to measure loss in occupancy at the monad level. By calculation of the number of occupied monads in the current and historic epochs respectively, it is possible to calculate the number of monads lost, and therefore the percentage loss. Appendix 8.31 shows the loss of occupied monads per Vice County. Appendix 8.32 shows the loss of occupied monads per Natural Character Area Deductions and relevance The Adder Status Project has enabled the first ever empirically-derived numerical assessment of the adder s conservation condition on a national scale. Although the model predicts imperfectly, and there are limitations to the component data, the overall approach is effective and seems broadly accurate. At a time of growing concern about adder declines, the achievement is a milestone for adder conservation. A similar exercise was previously performed for Surrey (Langham on SARG website, ); but ASP is breaking new ground. Furthermore, the proof of principle CGO Ecology Ltd Adder Status Project (England) Mar

23 attained by this project means that its methods can probably be employed for any other reptile species (and perhaps other species of restricted mobility), at any other spatial scale Conservation condition headlines Our assessment predicts that adders are currently living in 9,237 kilometre squares across England. With an assumed average population size of 24 adders per monad, this gives a national population estimate of 221,700 adders. The number of occupied monads is therefore 7% of the total monads in England. Coincidentally, this the same figure (7%) as the preliminary occupancy rate for adder in the NARRS Widespread Species Surveys (Wilkinson & Arnell, 2011). From the predictive modelling of suitable monads around historical records, the model predicts that at least 31,894 monads are suitable for adders, although only 9,237 are predicted to be occupied. From our preliminary assessment therefore, we calculate that adders occupy only 29% of the available habitat (at monad level) in England. Obviously these figures are preliminary and have their limitations, but they provide a first attempt at determining the under-occupancy that many surveyors report in the field. CGO Ecology Ltd Adder Status Project (England) Mar

24 In summary, our preliminary assessment of the adder s conservation condition in England is: Occupancy* = 9,237 km 2 Habitat = 31,894 km 2 Population = 221,700 adders. *Monad occupancy cannot be used directly as a range measure, as it does not specify which monads are occupied. Vice Counties offer a better spatial metric for range assessment. The predictive approach works well, and effectively fills some gaps in recording effort. The comparison in appendix 8.29 shows that without predictive modelling, the occupancy estimate would be 2,160 monads, and a population of only 51,800 adders across England. The number of Vice Counties (VCs) currently occupied is difficult to assess using CASM. Flaws in the data resulted in an occupancy (albeit tiny) in several VCs where adders are thought to be extinct, such as Nottinghamshire and Hertfordshire. The likely source of error is misidentified (and incorrectly verified) records Current adder distribution The range measurement aspects of ASP have great potential for wider dissemination, at national, regional and local levels. Distribution maps provide ecologists with a simple tool with which to gain useful insights into the overall status of any species. Notwithstanding the limitations of incomplete recording, a species distribution map, particularly if temporal, is the first line of attack in considering the status of a species. It is what determines to a large degree whether a species is widespread, common, scarce, or rare. The range maps included here are deliberately presented so that a coarse national view can be made, as well as a close inspection of sub-regional patterns. At a casual glance, it is readily observed that the overall distribution pattern is very similar to previous distributionmapping exercises such as the BRC atlases (Arnold, 1993, 1995), and lower-resolution treatments given by Hilton-Brown and Oldham (1991) and others. The adder is most widespread in Southwest England, the southern coastal counties of Southeast England, some of the upland areas of Northern England, and parts of East Anglia. Calling it common however, as some workers have done, would be inaccurate and highly misleading. There are parts of the county, such as the coastline of the Southwest peninsula and some other coastal areas of Southeast England and East Anglia, where the adder seems to be almost ubiquitous. Yet even here, it is not possible to make any conclusions, based on range data alone, about local population sizes or fitness. Those considerations can only be enlightened by local field study, and by research projects examining aspects of population ecology. Thus ongoing genetic work (T. Garner et al.), and ecological work (e.g. D. Carter, K. Palmer) will be of fundamental importance to understanding factors influencing local and national status. The English Midlands are, on the whole, devoid of adders; with only scattered and localised occurrence. This agrees with all previous assessments, and is a natural progression from the known historical situation. There are some weaknesses in the data and the predictions, but the model (and data) clearly works well in other areas. Appendix 8.17 shows the paucity of known and historic data in the southern Pennines (the northern Pennines are completely devoid of data), and this is duly reflected in the predictive mapping, although the model does recognise that many monads in CGO Ecology Ltd Adder Status Project (England) Mar

25 this area are suitable for adder occupancy. Given that some other upland areas are liberally populated with adders (e.g. North York Moors, southern Lake District), the apparent absence of adders from most of the Pennines warrant further investigation. Some of the possible reasons for this are discussed later, but gaps in the distribution map such as this must be viewed with some suspicion until adequate explanations are established, or field testing shows that adders are indeed absent from the majority of the Pennines. In the meantime, field survey of the Pennines should be treated as a field survey priority. Another flaw is the buffering syndrome evident in some lowland agricultural counties such as Herefordshire. Where there is a known current or historic presence, the CASM predictive mapping has predicted presence in all neighbouring squares with what it considers suitable habitat, up to 2km in all directions (the arbitrary parameter imposed in the model s design). To the human eye, it does not appear obvious what the model has found attractive about the neighbouring squares (based as it is on the CORINE land cover data), that it does not find suitable in all other agricultural areas. Possible reasons for this buffering effect could be under recording in an area (such that predictive clusters do not overlap as elsewhere), or inaccurate identifications, such as grass snakes, creating correlation in monads which are not in reality occupied by adders. Also, one of the apparent limitations of the model is that it seems unwilling to predict likely occurrence in forestry plantation areas, even where there is extensive current and historic data (noting that all empirically-observed records fall into predicted occupancy). An example is the Brecks of Norfolk and Suffolk, where current and historic data is largely limited to the south of the area; but with forest rides and open heathland and scrub areas evident on aerial photographs in the north, it seems inexplicable that CASM cannot predict current or historic occurrence in all contiguous blocks of habitat. Part of the problem is the imposed limit of 2km buffering in all directions; but there are many localised examples of forestry where CASM has not buffered at all, due to an inability to find statistically-significant correlation between the characteristics of the monad and local adder distribution Loss of conservation condition at the national level The distribution maps clearly show a reduction in range from historic (black) areas to the present (red) areas. The comparison maps and tables in appendix 8.30 put this in stark perspective. Between the historical (pre-2006) and current ( ) epoch, overall occupancy is estimated to have declined from 15,154 to 9,237 monads, which is a range reduction of 39%. The estimated population reduced from 363,700 to 221,700 adders The rate of habitat loss cannot be calculated, because historical land cover data is not available. The VC occupancy rate was 100% in the historical epoch; meaning all 58 English VCs were occupied. However, as mentioned already, the current occupancy rate is apparently 100%, because the occurrence data requires some cleaning. The likely number of currentlyoccupied VCs is thought to be 52-55, equating to a 5-10% loss of VC occupancy Loss of occupancy within Vice Counties The table in appendix 8.31 lists all 58 English VCs, and compares the historic and current predicted monad occupancies for each. The percentage loss is presented in the final column. The VCs are sorted according to percentage monad loss, in descending order. Once again, the results are striking; 57 out of 58 VCs have undergone a measured loss of adder occupancy at the monad level. CGO Ecology Ltd Adder Status Project (England) Mar

26 The effect is particularly striking for the VCs with the highest rates of loss. Evidently the rate of loss has not been similar across the country; far from it. The rates of monad occupancy loss vary from 0% to 92%. Cheshire, for example, has lost 92% of its occupied monads; the dubious honour of being the worst county for adder declines in England. Cheshire % Incredibly, the lowest rate of loss is apparently in West Lancashire (0%), which had and still has 3 monads. This should be viewed with caution, however, and probably represents spurious modern or historic records. West Lancashire % The greatest historical occupancy rate was in Dorset, where the model predicts that adders used to occupy 1155 monads, but half of these are now lost. Dorset % The VCs with the largest current numbers of occupied monads are Surrey (799), West Sussex (723), East Sussex (603), Dorset (592) and South Devon (584). Surrey stands out as having the lowest loss rate (15%) of these, the most important counties for adders. Fourteen VCs have undergone a loss of over 50%. More than three-quarters (76%) of VCs have endured a loss of over 30%. By any measure, this is drastic loss of adders on a nationwide scale Loss of occupancy within Natural Character Areas Using a Natural Character Areas spatial template, the monad occupancy figures were calculated for each NCA. The table in appendix 8.32 presents the list of NCAs, historic and current monad occupancy figures, and percentage rates of loss. Once again, the findings are dramatic. The difference in occupancy rate loss between NCAs is even wider than for VCs, ranging from 100% loss (8 NCAs) to 0% loss (18 NCAs). Conversely, this is easier to understand and explain though. Adders are strongly governed by habitat conditions, and NCAs are by definition a product of landscape character. Therefore the NCA system should instinctively be a good mirror for patterns of adder distributions Recommendations for further work The natural next stage for ASP is to set out a strategy for refining the current condition assessment, improving adder status-assessment systems, and for building capacity to deliver the source data. A large amount of data was not accessible in the timeframe available. Also, much of the available data was not verified (e.g. ARC s Add an Adder dataset), despite the likelihood that much of it is valid. Known limitations of the CASM model should also be explored via a second stage of model development and system design. Given the apparent usefulness of the NCA system in understanding adder distribution in England (more so than the VC system), the relationships between adder occupancy rates, NCAs and VCs would warrant further investigation by statistical methods. CGO Ecology Ltd Adder Status Project (England) Mar

27 Further iterations and/or designs of the CASM model should also explore ways of weighting habitat predictors using the NCA system. Sample sizes would be too small in some cases; thus it might require grouping some NCAs together. By integrating additional land cover datasets, it would be possible to see how the adder has declined in six-year reporting periods from 1990 to the present. This would provide an indication of the rate and variability of decline across numerous, albeit recent, epochs Further modelling The CASM 2.0 model is a remarkable achievement in the time available. It has identifiable flaws, and there are probably other flaws that are as yet unidentified. We propose that alternative algorithms for the model are explored and assessed. The rationale is partly objective, and partly subjective. By comparing the predictive distributions, we are able to evaluate them against our own understanding and perception of adder occurrence. Whilst there is still much to learn about adder ecology and habitat use, all adder surveyors would agree that predicting their occurrence is possible with a better-than-random outcome. The sum total of previous experience draws an adder surveyor towards bracken-covered banks, the southern edge of gorse clumps, sunny woodland edges, humps, banks and lone birch trees, in the search of adders. This is because experienced adder surveyors know (or think they know) that adders are much more likely to be in such places, at least in early spring. There are also certain habitat types that seem to be closely associated with adders. Yet the CASM model has selected its own habitat and land cover criteria based upon association with known occurrence. The latter is obviously biased towards human population centres, and this has had a distinct effect in its weighting of certain categories. This could be mitigated by the implementation of variable buffering based upon spatial analysis of localised observations. Further model iterations and designs could explore subjectively-predicted correlations with adders occurrence. For example, the CORINE land cover categories include at least twelve (out of 48) that would catch the eye of an experienced adder surveyor: Road and rail networks and associated land Mineral extraction sites Sport and leisure facilities [golf courses] Pastures Land principally occupied by agriculture, with significant areas of natural vegetation Agro-forestry areas Broad-leaved forest Coniferous forest Mixed forest Natural grasslands Moors and heathland Transitional woodland-shrub Beaches, dunes, sands Bare rocks Peat bogs CGO Ecology Ltd Adder Status Project (England) Mar

28 The premise for selecting each of these subjectively is that they either have their own merit directly, or they are likely to be associated with adder-friendly habitats. The current model design automatically calculates statistically-significant habitat associations and correlations. Several possible options for next-stage model development are: 1) Subdivide England into parcels of different human density, and evaluate habitats for each 2) Weight each monad according to distance from major population centres, or an index of visitability 3) Manual selection of important habitats from the CORINE list (as above) to be weighted more heavily 4) Automatic habitat evaluation, but recalculated for each Natural Character Area. The first two are more robust and desirable statistically, and probably easier to test independently. Option one could involve division of England into several categories of human density. The second would probably involve a weighting applied for every monad, according to its location. The third option is highly subjective, and would be difficult to test without extensive field survey. The fourth option would probably involve too little data to allow reliable statistical calculations. With further thought, these and other options could all be explored Field testing results It is envisaged that this project might lead to field testing of some of the conclusions made and insights gained. This is unlikely to be feasible in 2012 if based upon standard springtime adder counting techniques. Other habitat-based aspects of the research, such as testing HSI parameters, or trialling hibernaculum-mapping systems, could be progressed later in the year, assuming involvement of suitably-experienced participants. Field testing of the CASM predictive modelling of range is one aspect that needs to be explored. For example: are adders really absent from large areas of the Pennines? Field results are numerous from some other upland areas of Northern England, so it seems natural that a gap on the adder map as large as the northern Pennines must reflect a real absence on the ground. The only way to prove this is through targeted field survey. Field survey should also be carried out in predicted occupied squares adjacent to known current data where possible. As discussed earlier, we suspect that some of the predictive buffering is flawed (and the reasons are identifiable); such as in Herefordshire and fenland parts of Lincolnshire. In other areas, such as Dorset, the predictive mapping seems reliable; and it would be worth carrying out field testing of predicted occurrence (as well as random squares not predicted). It should be noted that field survey conducted by Surrey ARG in 2011, using an earlier version of the CASM to prioritise monads for survey, yielded successful new monad occupancies for adder in 19 monads out of 19 surveyed. CGO Ecology Ltd Adder Status Project (England) Mar

29 3. The importance of hibernacula 3.1. Background and context It is well-known that adders have a high degree of site fidelity, more so than other reptiles; and that they are tied to particular habitat patches. This is particularly true of hibernation areas. Often the same location, usually a pre-existing subterranean void, may be used by the same individual for many years. Adders also have a tendency to hibernate communally. These factors mean that adders are more vulnerable to stochastic effects than other reptiles, and a single localised event can effectively wipe out a whole subpopulation. It has become increasingly recognised that better understanding of adder hibernacula, and better awareness of their value, is crucial to the conservation of adders. Adders generally return to the same hibernaculum (hibernation place) each winter throughout their adult life. Their home range during the spring breeding season and summer feeding season might extend for hundreds of metres in any direction, but they are able to follow scent cues to return to their traditional hibernaculum. Many surveyors can identify the localised areas where adders hibernate within a larger site. It is less common to know the exact location of a hibernaculum. Hibernacula are typically in mammal burrows or crevices under the roots of a bush or small tree. Emergence in the early spring (February onwards) is a dangerous time for adders; and a hibernaculum under a bush affords cover from avian predators, whilst allowing insolation from the low sun. Dead bracken and dead gorse litter also provide camouflage for emerging adders. Males emerge first, and complete spermatogenesis during this period; females tend to emerge a few weeks later. Hibernacula are often on a south-facing slope or bank, or the southern edge of a patch of scrub, although they can be on flat open ground. The location is always in direct view of the sun in early spring (from approximately hours in February). This optimises thermoregulation opportunities. Tens of adders have been known to occupy the same hibernaculum, and a mass emergence of adders from a communal hibernaculum is probably one of Britain s most impressive wildlife spectacles. However, communal hibernation leaves adders vulnerable to persecution, predation, and accidental damage On the plus side, site fidelity and communal hibernation are phenomena that enable population monitoring with consistency that is not possible for any other reptile species Aims and rationale Existing knowledge on hibernaculum locations and characteristics is held by a wide number of people, and it is not easy to access this combined knowledge. Accessing this information is therefore a natural prerequisite to gaining further insights, disseminating knowledge, and developing management best practices. As part of ASP, we aim to improve the understanding of adder hibernacula, and hibernation areas in general. In particular, we aim to develop ways of recognising hibernation areas, and where possible, specific hibernacula locations. This should assist their protection and improve awareness Approach and methods CGO Ecology Ltd Adder Status Project (England) Mar

30 A short pilot questionnaire survey was composed as a way of gathering information informally, that might inform the design of further research. The approach attempted to define what makes a location attractive to hibernating adders, and whether any correlates might help us to identify such sites. The assumption was made that the availability of suitable hibernation sites is one of the most important determinants of adder survival Questionnaire design A multiple-choice questionnaire was circulated to six experienced adder surveyors with detailed knowledge of adder hibernation sites. Several of the participants were chosen for their knowledge of many adder sites. The questionnaire asked four questions, with multiple choice answers (marking all that apply): 1) Do the adders hibernate under, or close to, any of the following vegetation types? (gorse, bracken, birch, oak, broom, heather, marram, Molinia). 2) Do the adders hibernate in any of the following site features? (vertical cliff, steep slope, gentle slope, flat ground, woodland edge, scrub edge, rock/scree, stone wall, embankment/'sea wall', localised hump/mound). 3) To your knowledge, which other reptile species are present on site? (common lizard, slowworm, grass snake, sand lizard, smooth snake). 4) Are adders mentioned in the site's formal management plan? (Yes, specific management; yes, it mentions them; no, but it mentions other reptiles; reptiles are not mentioned; there is no management plan; don't know) Assumptions and limitations Questionnaires are inherently subjective, like all hypothesis-base science. The questions were designed to test hypotheses implied by them. The list of possible answers was not exhaustive, and participants were invited to insert their own additional options. At least one obvious answer was accidentally omitted from question 1 (bramble). Nevertheless, the answers to questions were categorical; therefore the results are easy to analyse, and should be fairly robust. The questionnaire assumed that the participants knew some adder sites very well, knew where adders hibernated on those sites, and that their memories were accurate. A possible limitation of the data is that the character of the sites included might not reflect the diversity of adder sites across England. However, the geographical scope was wide, and although should be fairly representative. Another constraint is the likelihood that participants are better at responding to a list of options than inserting their own answers in the Other category. Hence the frequencies for the additional categories are subjective and probably under-estimated Results The six respondents (Lee Brady, Jon Cranfield, Chris Gleed-Owen, Steve Langham, Chris Monk, Kevin Palmer) provided data on 109 adder sites between them. The data covered the counties of Cornwall, Derbyshire, Dorset, Essex, Hampshire, Kent, Norfolk, Oxfordshire, Somerset, Surrey; plus two Scottish sites, in Lothian and Perthshire. The results are presented as percentages of the total number of sites, in appendices CGO Ecology Ltd Adder Status Project (England) Mar

31 3.5. Deductions and relevance Not unexpectedly, bracken is the most important plant species in adder hibernation areas (59% of sites). Adders are cryptic, and their dorsal zigzag pattern is crucial to their camouflage. Bracken aids crypsis, especially for ginger females and juveniles. Gorse and birch are evidently also of importance in hibernation areas. Both are regularly used as the location of hibernacula. The relatively low figure of gorse (22% of site) is perhaps surprising though. Gorse litter also provides good camouflage for adders, especially grey, beige and brownish animals. Nevertheless, as long as bracken is present on site, gorse is probably not such a critical factor. Birch trees were present on 43% of hibernation areas, which is not surprising, as adders often hibernate beneath them. Other trees were also significant, such as oak (10% of sites) and conifers (9% of sites). Bramble was surprisingly well correlated with adder hibernation areas (30% of sites), as was Molinia tussocks (18% of sites). Heather and/or grassland were present on the majority of sites, although occurrence of the latter was not specifically measured. The character of adder hibernation sites was somewhat elucidated by the second question. In terms of slope, 57% of adder hibernation areas were on gentle slopes; with 32% on steep slopes. Aspect was not questioned (though is presumably south-facing in the vast majority of cases). Flat ground was used in 31% of cases however, showing that aspect is not crucial. Localised humps/mounds feature heavily as hibernation locations (59% of sites); a singularlyimportant insight for recognising hibernacula. Embankments were used on 21% of sites, and stone walls and rubble were used on 7% and 6% of sites respectively. Woodland edge and scrub edge are clearly important ecotones for adder hibernation (39% of sites each); whereas single trees/bushes were only used on 12% of sites (lower than expected?). The implication of scrub blocks being important on 39% of adder sites is obvious for management. Likewise, woodland edge is equally valuable, and has implications for management. Clearly, the ecotones provided by localised changes in topography and vegetation cover are crucial to adder hibernation. They offer optimal aspect, shelter, microclimate, drainage, location, protection and perhaps other characteristics that are important to successful hibernation and emergence. The early-spring emergence period is particularly important for males completing spermatogenesis. Bracken, gorse, and probably leaf litter from oak and birch, provide excellent camouflage for sluggish emerging adders. Bushes and trees provide cover from aerial predators, yet do not obscure the low sun. Slopes, banks, raised humps, and loose material provide well-drained ground. As expected, there is good evidence that other reptile species are closely correlated with adder presence. Other reptile species were reported from 89% of adder sites. Conversely, it is perhaps surprising that no reptiles were reported from 11% of them. This might be due partly to incomplete detection, as most of these sites are monitored for MTAC using visual search alone, without artificial refugia. Syntopy with common lizard characterises a remarkable 86% of adder sites. In other words, if adders are present, common lizards are nearly always present too. Slow-worms (55% of sites) and grass snake (50% of sites) were also strongly correlated with adder sites. The inclusion of adders in management plans gave a diverse set of responses. Adders were mentioned in 54% of site management plans, although the data was skewed by almost half the sites being from Derbyshire, where they fall within a single management plan. Only 7% of CGO Ecology Ltd Adder Status Project (England) Mar

32 sites had specific measures for adders. Reptiles were not mentioned in 11% of plans, although this is likely to be an underestimate Recommendations for further research This was a valuable exercise, and the chief recommendation would be to circulate the survey more widely, with some amendments to the questionnaire. A larger sample size (perhaps sites?) could be achievable via existing monitoring networks with a little effort. The questionnaire might benefit from a more specific question asking what types of void are used by adders as hibernacula. Many surveyors would not be able to answer this question; but in some cases it would be possible to give multiple responses for one site. 4. Management practices and factors affecting status 4.1. Background and context Etc etc 4.2. Damage caused by habitat management and restoration Stories abound among the herpetological community about land management practices damaging adder sites, particularly hibernation areas. Habitat management for nature conservation is increasingly blamed. Historically, developers and commercial interests such as farming, forestry and moorland estates were the main antagonists; but today it is often nature conservation bodies that are accused of acting insensitively. Wildlife NGOs, local authorities, statutory agencies, and private land managers have all been criticised in recent years for carrying out insensitive management and restoration that either killed hibernating reptiles, or destroyed their habitat without appropriate mitigation. Typically these reports come from nature reserves, which by their island nature are often close to human population centres, rather than in the upland wilds of Britain. Hence, there is a disproportionately greater reaction to perceived damage on nature reserves than that which takes place each year on grouse moors. One of the commonest allegations is that wholesale scrub removal has taken place on a nature reserve, without consideration of hibernating adders, when simple avoidance measures could have been employed. With a little local knowledge of hibernacula locations, it is possible to identify localised patches of scrub, bracken, birch trees or other features that could be retained to avert damage to the adder population. The most damaging practice is the use of mechanical means. Tractor-mounted flails pulverise the vegetation cover, and effectively scarify the ground. If this takes place in adder hibernation areas, there is a real chance of killing adders and destroying their hibernacula. There seems to be a constant flow of such stories each winter and early spring (examples are discussed later), including one from Dorset in 2011 where two female adder corpses were found among flailed gorse chip (Gleed-Owen, unpublished). Even hand-cutting of scrub can damage hibernacula by exposing the ground. This leaves emerging adders vulnerable to aerial predators, removes shelter from the wind, and alters the microclimate; an important consideration for a reptile that emerges in February. On the whole though, hand-cutting is far preferable to mechanical removal; and with appropriate dialogue and awareness, individual bushes can easily be marked for retention. CGO Ecology Ltd Adder Status Project (England) Mar

33 The important plant species used by hibernating adders that are most commonly affected by winter scrub clearance on lowland heath are common gorse and silver birch, but other species such as Scots pine and bramble are also affected. Bracken treatment is also a potential risk to adder populations, especially if rolled or removed by other mechanical means. Bracken provides excellent camouflage for emerging adders in early spring (as does dead gorse litter), and is an important hibernation site choice. Hence, if ground is stripped or scarified, this can kill basking adders, and compromise hibernacula. From late spring onwards, adders typically move to summer feeding grounds away from their hibernation areas; shading from bracken would be a problem otherwise. Bracken rolling and other mechanical techniques are still a potential risk to adders though. Modern conservation management is by no means the only cause of damage to adder status. There are other traditional land management practices, primarily for cultural and economic purposes, that have caused incalculable damage to adders over the last century or more Damage caused by controlled burning Controlled burning of heather and gorse has been practiced on vast upland moorland estates in northern England, Wales and Scotland for over a century; primarily to provide short heather for grouse-rearing. Heather-burning (or muirburn in Scotland) is a landscape-shaping activity on a vast scale. It not only controls the vegetation cover across whole landscapes, but the nature of the other wildlife that lives on it. It is a practice with such a dramatic effect that it can be seen clearly from aerial photographs on any internet search provider. The Moorland Association states on its website that controlled burning can take place legally between 1 st October and 10 th April, with the burning cycle repeated every seven to twenty-five years ( The size of individual burns was traditionally many hectares, sometimes a square kilometre; and a casual scan of aerial photographs on the internet will show a striking patchwork of strips and scars belying a century of burning. The combined effect of hundreds of thousands of burns over such a sustained period must have been the erasure of reptiles, including countless adder populations, from vast areas of upland Britain. The more ubiquitous upland reptiles (common lizard and slow-worm) may have fared better than the adder, but at least one survey (Gleed-Owen, unpublished data, North Yorkshire) shows that even the slow-worm can be extirpated from a cyclically-burnt landscape, and only the common lizard is able to survive at a more or less continuous low-density presence. Estates are now under pressure to reduce burn sizes, in response to nature conservation concerns. Current Natural England policy restricts individual burns to one or two hectares. Nevertheless, the damage has been done for the adder; and its range has been reduced as a result. Whilst historical records may be lacking to prove former presence in many areas, the implication from un-burnt areas where adders still thrive is compelling. For example, the southern end of the Peak District is well populated with adders, whereas the rest (comprising grouse moors) is a large blank on distribution maps (e.g. Arnold, 1995, and data presented here). Controlled burning is also the norm in lowland heath areas such as the New Forest, where the Forestry Commission burns heather and gorse on a roughly fifteen-year cycle, to encourage new growth for livestock grazing. The practice is carried out during the winter and early spring (to avoid bird nesting season), but often encroach well into March when reptiles are emerging (including the European-protected CGO Ecology Ltd Adder Status Project (England) Mar

34 smooth snake). Regardless of whether reptiles are killed above ground, burning inevitably kills hibernating reptiles too (even with a fast shallow burn). It also removes vegetation cover, leaving them unprotected, and with no food source. Reptile survey data from across the New Forest shows that adders are absent from significant areas where smooth snakes are present (ARC, unpublished). Whether this is down to burning primarily, or other factors (Brusher Mills?), is hard to prove. The Forestry Commission now keeps GIS records of its burning schedule each year, and these could potentially elucidate the effects on reptiles. A scoping project is currently under way by ARC (funded by Natural England) to investigate the effects of controlled burning on reptiles. If it were more vagile and ecologically plastic, it might be able to capitalise upon shifts in policy, such as reduced burn size on grouse moors. As it stands, the trend towards small burns is likely to increase the area of habitat available to adders, but not likely to improve the adder s ability to recover Other causes of adder decline and extinction It is worth considering some of the other ongoing factors causing adder habitat loss, and adder declines. Permanent historic loss of habitat aside, agriculture and forestry are probably two of the biggest influences on adder status. Forestry operations are possibly the greatest influence, as they operate on a cyclical basis that can make or break local adder populations. Forestry also has great potential for reversing adder declines, should deforestation ever occur Unexplained absence from some upland areas Some upland areas of England appear to be devoid of adders, despite being superficially suitable landscapes. Almost all of the Pennines, the northern Lake District, and Dartmoor have no recent or historic records (Arnold, 1995, and data presented here). Have adders been extirpated from these areas, or were they never colonised? On aerial photographs they appear lighter, which might reflect a predominance of Sphagnum bog (i.e. waterlogged) and bare rock, rather than dark heather. Unsuitable habitat therefore seems one of the most obvious explanations for the absence of adders, but probably not the whole story. Overgrazing is well-known as being a serious problem in some upland areas, including the Lake District, but would have to have been a permanent problem since nineteenth-century biological recording began. Persecution has probably always been a significant problem in some areas, but does not readily explain a landscape-wide effect. Recent historic burning for grouse management is not evident on aerial photographs for these upland areas, and does not seem the obvious explanation; although natural or deliberate largescale burning in the distant past is a possibility (on a scale not known since records began) Cynical ploughing In some lowland areas, such as the South Downs in Sussex, there have been allegations of cynical ploughing by farmers, to remove wildlife liabilities from previously-unploughed land, prior to the designation of the South Downs National Park. Historic agricultural intensification, particularly after the Second World War, was certainly one of the biggest direct causes of adder habitat loss and extinction. Indeed, several thousand years of agricultural expansion in Britain (and associated deforestation) is undoubtedly the single biggest contributor to the adder s decline towards its current status Pheasant-rearing CGO Ecology Ltd Adder Status Project (England) Mar

35 Pheasant-rearing has been blamed for adder declines in some previous studies, and in anecdotal accounts. Pheasants are known to predate adders Direct public pressure Public pressure has been cited repeatedly as a cause of adder declines, due to disturbance, particularly in early spring (Baker et al., 2004; Hilton-Brown & Oldham, 1991; Sheldon, 2011). Fog-walker often seem to get the blame, although it is difficult to see how such a factor could have a large-scale effect Genetic problems Inbreeding effects have been discussed recently as a potential cause of adder declines in isolated populations, and a possible reason for their impaired ability to colonise new areas. A current study by the Institute of Zoology (Zoological Society of London) and other partners is examining microsatellite DNA from adder populations across England Mitigation failure Although mitigation should occur whenever adder populations are directly affected by development-related activity, this is a very grey area. As discussed above, there are far more situations where appropriate mitigation does not take place, than the relatively few situations where it does. In England, mitigation is fairly universal, but there are still many local authorities without dedicated ecologists or biodiversity officers that might offer guidance in this respect. There are still some places, however, with a de facto vacuum of mitigation activity, including industrial areas within upland landscapes. Even where mitigation does occur in England, the practices and result are often questionable in the eyes of many observers. The most significant case in point is the recent translocation of 291 adders from the Thames Gateway port development in Essex to Wilshire Wildlife Trust s Blakehill Farm reserve near Cricklade in Peak count in 2010 was eight adders. It is too early to say whether this was a success or failure, and to tell whether the missing adders have survived or not; but this case has generated a large amount of public debate. The key question is whether or not reptiles should be moved over such great distances, particularly if local alternatives were available. Secondly, it has been alleged that not all local alternatives were exhausted. Indeed, the large scrub and grassland site owned by Corporation of London at Wanstead Common in Southeast London was considered, but permission seems to have been refused by the owners. Hence, even if followed, there is no agreement on best practice at present. Natural England has withdrawn its Reptile Mitigation Guidelines, pending review; and other guidance is out-of-date and not fit for purpose. It is worth mentioning Scotland in this context. The primary factors causing historical and ongoing adder declines are afforestation and muirburning, but a general absence of development-related reptile mitigation in Scotland is also a likely contributory factor (C. Cathrine, pers. comm.) Examples of good management practice There is little benefit in criticising bad practice without presenting examples of good practice. Good examples should be used to promote best practice, and bad examples should present opportunities for improvement. CGO Ecology Ltd Adder Status Project (England) Mar

36 It is unusual for adders to be considered in, let alone explicitly incorporated into, local site management plans. Adders are often featured in Local Biodiversity Action Plans (LBAPs), but these as meaningless unless put into practice. There are some notable exceptions however. In the southern Peak District of Derbyshire, fifty or so known adder populations (all subject to MTAC monitoring) fall within the draft management plan that covers a complete 2,000-hectare estate, and which excludes damaging management from known adder hibernacula (C. Monk, pers. comm.). This is an admirable achievement, for which the credit is due to the local ARG coordinator and adder surveyor in question. In Norfolk and Suffolk, the Forestry Commission gathers brash, needle-litter and topsoil into windrows (low linear banks) after clear-felling operations. As the ground cover matures over the next few years, and reptiles begin to colonise clear-felled areas, these windrows offer opportunities for adders to bask and hibernate. Up to fifty adders have been counted on one windrow (J. Cranfield, pers. comm.). Many clear-felled forestry areas are eventually replanted and cropped for timber in due course; but in the meantime, windrows assist the adder metapopulation. ARC operates similar practices in Surrey, on former forestry plantation that it is restoring to heathland. Scarifying the ground allows the heather seed bank to regenerate, and the scraped material can be gathered into windrows that are focal points for adders and sand lizards (M. Preston, pers. comm.). These examples have a positive outcome for reptiles only if the locations are reptile-free beforehand. Mechanical vegetation removal and scarification are questionable when carried out on areas already populated by reptiles. Another positive example is Essex County Council s adder-friendly habitat management at Hadleigh Country Park. The site manager (and MTAC surveyor) scallops bramble edges, to provide sheltered basking areas for emerging adders, and maximise edge habitat for biodiversity in general, including carder bees (J. Cranfield, pers. comm.). In Hampshire, Martin Down NNR has been cited as an example of a sensible chalk grassland grazing regime that still supports good adder populations. Grazing of heathland at low densities at Eelmoor Marsh allows adders to co-exist (P. Edgar, pers. comm.). In Dorset, an urban heathland has undergone a subtle transformation over the last ten years, owing to low-intensity grazing by ponies. Continuous heather blocks have been broken into smaller areas, beneficial to basking adders; but Molinia tussocks have been destroyed. Seasonal cattle grazing was introduced a few years ago on the same site, with bad timing in early spring; but following advice to the owner, the adder emergence period is avoided. A similar outcome has since been achieved on another coastal grassland site, where cattle are grazed over the winter, but removed in February before the adders emerge from hibernation. In contrast, large areas of the New Forest in Hampshire, and some Dorset heathland sites, are heavily grazed. Grass sward is greatly reduced (almost to a bowling green in parts of the New Forest), and scrub blocks are isolated. Reptiles are severely compromised in such a landscape, and exist only in isolated islands or inaccessible strips beyond fences Examples of bad management practice In Hampshire, a nature reserve formerly supporting a good population of adders recently underwent over-zealous scrub clearance and damage to adder hibernation areas. Overgrazing also destroyed Molinia tussocks used as hibernacula by adders. Adder numbers have reduced from peak counts of over twenty, to zero in The removal of bramble and scrub meant CGO Ecology Ltd Adder Status Project (England) Mar

37 adders were displaced onto marginal areas and neighbouring sites, and there is hope for recolonisation (J. Cranfield, pers. comm.). In Essex, scrub clearance and scarification left adders in isolated islands of habitat. No surveys had taken place prior to the works, but displaced adders were seen in local gardens afterwards (J. Cranfield, pers. comm.). In Essex, some local stakeholders have expressed concerns over mowing of seawalls that led to a decline in adder numbers (J. Cranfield, pers. comm.). At a well-known forest site in Essex, a dispute over heathland management led to declines in adder numbers of over 75%. Adders are an interest feature of the SSSI (J. Cranfield, pers. comm.). In East Yorkshire, heathland restoration works in March 2012 are currently causing a furore on social networking sites and internet blogs. The main adder hibernation area on the site proved to be unaffected by this work, despite rumours to the contrary. Nevertheless, previous advice from a reptile expert (as part of a proactive campaign) was apparently overlooked, and mechanical clearance works were more severe (C. Monk, pers. comm.). In Oxfordshire, where adders are perhaps down to one site, it has been suggested that the management history of other sites has been complicit in the species demise (R. d Ayala, pers. comm.). Similarly in Warwickshire, neglect and unmitigated habitat damage are thought to be among the reasons for adder extinction in that county. In Nottinghamshire, the last known adder population was on rides in a forestry plantation. Inevitably, conifer and scrub growth extirpated them. Ironically there were extensive heathland blocks within the same forest, but the adder population was too small or simply unable to colonise them naturally. In Dorset, there have been many examples of heathland management coming into conflict with reptiles over the last ten years. Insensitive gorse clearance at several sites was damaging to several adder sites monitored for many years by Tony Phelps (pers. comm.). Controlled burning and mechanical turf-stripping have also been employed on heathland (without mitigation); the latter conducted for the benefit of woodlark, on statutory advice. A peri-urban coastal grassland and scrub site in Dorset was badly damaged by mechanical gorse removal and scarification in early 2011 (Gleed-Owen, unpublished). Photographs have featured in several public presentations and workshops. Two adult female corpses were found mangled among gorse chip, and one juvenile adder was moved to intact habitat nearby. Numerous common lizards and slow-worms were inevitably killed too. The site had been monitored for MTAC since 2005, with permission of the landowner. The maximum count is six, and the population is thought to number fewer than ten adults. The owner had previously removed scrub by hand on a neighbouring area in 2007, without consultation. Most of the key gorse bushes used by hibernating adders had been removed, and one dead male adder was found. Remedial action in the form of carefully-positioned brash piles remedied the situation somewhat, and offered protection to adders when they emerged soon afterwards. Advice was given to the owner (a local authority) after the 2007 clearance, but this was evidently ineffective in preventing a more severe recurrence four years later. It is difficult to know what else to do, other than keep plugging away persuasively at land managers. One distinct paradox in Dorset and Hampshire is the differing treatment that controlled burning and arson/accidental fires receive. In the Southeast Dorset Conurbation (Bournemouth, Poole, Christchurch), a multi-million-pound partnership project involves the police, fire brigade, landowners and other stakeholders to prevent heathland fires, and educate people about the value of urban heaths. Arson and accidental fires are a major problem on urban heaths. A rapid CGO Ecology Ltd Adder Status Project (England) Mar

38 response from the Dorset Fire & Rescue Service means that fires are often extinguished quickly, but very large fires (tens of hectares) still occur every few years. Ironically, in other parts of Dorset, and over the border in Hampshire, controlled heathland burning is used as a management tool. (Athough arguably, some heathland sites have suffered a little through lack of burning, as small burns helped to maintain a mosaic of habitats). ARC manages almost eighty nature reserves where reptiles are one of the primary interests. Their in-house winter management team always uses hand-cutting of invasive scrub, except in situations where reptile harm would be minimal if mechanical means (external contractors) were used. During May they carry out bare sand management, using tractor or rotovator, which carries acknowledged risks; but the timing is the best to avoid hibernating animals, mating, and sand lizard egg-laying. Their bare sand provision is praised for its benefits to solitary bees and wasps. Adders are treated as one of the three most sensitive reptile species on ARC reserves; the other two being the rare species, sand lizard and smooth snake. Hence hibernation areas are identified where possible, and incorporated into management. ARC also carries out heather mowing of sinuous strips (to create firebreaks and a mosaic of heights), which has inherent risks but results in an immediate improvement through provision of edge habitats. As an organisation with extensive advocacy experience, ARC encourages third parties to follow best practice, and has published the Reptile Habitat Management Handbook (Edgar et al., 2010). It provides extensive advice on all aspects of habitat management where reptiles are involved, including sections on damaging activities and conflict resolution. It lists mechanised treatments, reducing scrub to below 5% cover, and intensive grazing as high-impact practices with regard to reptiles. Paradoxically, it treats flailing of vegetation to ground level as only a moderate impact on areas used for spring basking. The handbook lists a host of factors that can be the source of conflict between opposing views on the best way to manage habitat; from scale and methods, to timing and purpose; and offers avoidance mechanisms. ARC also ran three training courses on reptile habitat management for practitioners in 2011, aimed at spreading the message of the Reptile Habitat Management Handbook to a wider audience. CGO Ecology Ltd Adder Status Project (England) Mar

39 5. Recommendations for conservation policy 5.1. Overview A lot has been learnt from this exercise, but a great deal of latent knowledge is also in unexploited, among adder experts such as survey volunteers. Some of the key issues facing current adder conservation policy are discussed below. These should set the scene for further consultation, perhaps via a workshop, to gather latent knowledge and use it to inform future conservation policy Recommendations for research Further research into adder population ecology would be beneficial. We know too little about the relatedness of neighbouring populations, possible genetic effects, movement between populations, and viable population sizes. It would also be valuable to learn more about the correlation of adder numbers with habitat quality, to explore the value of habitat quality and extent as a proxy for conservation status. Ongoing genetic and habitat-based projects described earlier ought to be encouraged, and built upon where usefully possible Developing systems for monitoring adder status Following on from this phase of the project should be a period of consultation and systems development. It has not yet been possible to reach the stage where knowledge of adder status is sufficient to design monitoring systems; but progress is sufficient to inform further iterations of the process, and guide it in the right direction. The short timescale for this work prevented a formal workshop or meeting between stakeholders. As directed by the brief, key players were included, and consultation went as widely as possible in the time available. However, with more time and resources, it would be beneficial to gather relevant experts and stakeholders together in a workshop environment, to assist the design of further research and development. The current coordination structure has worked well, with overall direction from Natural England, administration via ARC, and delivery by CGO Ecology Ltd with assistance from Natural Acuity Ltd. A second phase of ASP is recommended, starting as soon as possible, to explore the abovementioned aspects of status assessment Developing volunteer capacity The perennial limitation of large-scale field-based projects is the availability of surveyors. Projects such as NARRS have achieved initial take-up of up to two thousand volunteers, but relatively few remain involved for the long-term. It is also typical for a few of the keenest CGO Ecology Ltd Adder Status Project (England) Mar

40 volunteers always to be responsible for a disproportionate amount of data. Although, in Derbyshire at least (see below), the number of sites surveyed seems to track the number of surveyors. For a monitoring project such as MTAC, it seems unrealistic to aim for more than one or two hundred regular, reliable surveyors. There are undoubtedly many more people capable of participating, but unless bitten by the bug or somehow persuaded to take part, they are liable to drop out eventually, which undermines the data. There are currently 171 surveyors on the MTAC contact list, but fewer than a hundred are regular surveyors at present. (J. Baker, pers. comm.). Evidence from the Surrey ARG reptile survey programme shows that in order to maintain good numbers of active surveyors, a continuous training programme must be in place. This not only supports recruitment, replacing surveyors who become inactive, but is also essential for maintaining high quality recording. SARG currently has around 50 active reptile surveyors who deliver approximately 300 formal reptile surveys per annum. The feasibility of attracting surveyors to participate in a new monitoring project, or an extension of an existing one, depends on what commitment and skill level it involves. At present, MTAC and NARRS require some training and familiarity with adders, and especially their hibernation areas for the former. Another problem is the tight seasonal window of adder count monitoring. It restricts fieldbased surveys of adder presence to the spring. If there were aspects of adder status that could be monitored in other seasons, or by proxy, this would help spread limited volunteer time more widely. Academic students would be an alternative source of intensive field labour, especially in a localised geographical area, and if the methodology is not tightly dependent on early spring emergence. Post-graduate research degrees typically have a strong research element that can be field-based. Many ecological professionals can be persuaded to partake in voluntary surveys, provided clear infrastructure and direction are provided, and there is a sense that their effort contributes towards tangible conservation benefits. Being professionals, the quality of data provided from such surveyors tends to be high. Some professionals are attracted by the career development opportunities that are provided by the training and experience that a formal survey programme can deliver. CGO Ecology Ltd Adder Status Project (England) Mar

41 It might also be possible to gather habitat data more easily than actual adder data. For example, the hibernation characterisation exercise discussed earlier could be extended and tested in the field, perhaps as part of an HSI-piloting exercise. A random sample of sites could be assessed for likely suitability for adders, and then fieldsurveyed for actual adder presence Recommendations for Natural England policy There are no obvious quick-fixes, and few low-hanging fruit in terms of future adder conservation policy. Evidently the way the countryside is managed at present has a net detrimental effect on to adder s status. At the national policy level, continued efforts should be made to minimise the effects of controlled burning. The hard-won reductions in burn sizes in recent years should be maintained, and the potential for adders to re-colonise moorland landscapes should be examined, alongside local mapping exercises to elucidate the spatial relationship between adders and controlled burning. This is of equal importance to upland grouse moor burning, and to lowland heath burning, such as in the New Forest. Neither practice has been properly evaluated in terms of reptile status. Research projects should be pursued and supported wherever possible, and guided as appropriate by the current ARC-coordinated pilot project. This is examining the effects of burning on reptile status in Dorset, delivered by Chris Reading of the Institute of Terrestrial Ecology. A workshop would also be worthwhile, to explore ways of communicating and influencing land managers in a measurably-effective manner. Too many people have stories describing years of communication, often thought to have been friendly and productive, only to discover the advice had fallen on deaf ears, or the land manager had wilfully ignored it. It seems that, despite the best will in the world, land managers can sometimes be inexplicably unresponsive to friendly, cost-free advice with cost-free implications. This is even the case (and possibly more so?) with some of the major conservation NGOs. ARC s advocacy efforts in Dorset have been the source of frustration more than success; and a subtle softly-softly approach seems to be the most effective approach. Despite being the natural reaction upon seeing a devastated adder site, the robust confrontational approach offends land managers, and makes them close ranks. It burns bridges and cuts off communication pathways. Calling the police is an option openly discussed regularly, where perceived as justified (e.g. an avoidable and potentially criminal act). Clearly this would not be advisable if a good ongoing relationship with a land manager is desired. Gary Powell, ARC s Dorset Reserves Manager, has extensive experience in local reserves management politics. He says: If we are to change people s approaches to adder management, I think it has to be in a non-confrontational way, employing a bit of tact and guile, and understanding of their position. All ARC reserves in Higher Level Stewardship (HLS) have a Farm Environment Plan (needed for the actual application), which is in essence a management plan. However they are different from any plan that ARC might create independently, as by necessity they focus on Natural England s indicators of success (percentage coverage of gorse, tree density, heather age structure etc). Within the HLS, ARC is able to tweak the maintenance and manage sites for reptiles, but there is little herp-specific action (other than points awarded for rare species). The concern over the last few years is that there isn t anywhere near enough weight given to adders in management planning and field operations. ARC is now beginning to map adder CGO Ecology Ltd Adder Status Project (England) Mar

42 hibernation and emergence sites, both known and potential, in a way similar to their existing sand lizard foci maps. Field staff are being educated where necessary, about sensitive habitat management for adders, adder ecology, vegetation structure, seasonal movements, and implications of management. New management plans are being drafted, and centred on actual species requirements, as HLS (which is good for overall heathland management), has no explicit species focus. ARC would then be keen to bring some of the local authorities on board, to take similar measures, for example setting up grazing monitoring programmes like ARC is carrying out (G. Powell, pers. comm.). Gary sums up what a lot of adder watchers have been feeling: I can t bear to see any more hibernation sites over-managed and affected negatively by conservationists. CGO Ecology Ltd Adder Status Project (England) Mar

43 6. Acknowledgements ASP has been a short-lived exercise that will hopefully be a precursor to a great deal of further work, exploring other avenues, testing systems, and developing modelling and monitoring methods in the future. Funding was provided by Natural England, via a Memorandum of Agreement with Amphibian and Reptile Conservation. CGO Ecology Ltd coordinated the project, and Natural Acuity Ltd was project partner for the delivery of the data analysis, modelling and GIS elements of ASP. Paul Edgar (Natural England) was instrumental in the initiation and guidance of this project. Tony Gent (Amphibian and Reptile Conservation) was invaluable, and other ARC assisted with various aspects of ASP: Andy Arnell, Chris Dresh, Nick Moulton, Gary Powell, Angela Reynolds, John Wilkinson, and Helen Wright. A large amount of data collation was involved in this project, despite the short timescale available. Many data providers kindly assisted by extracting and supplying data at very short notice. Most suppliers required compliance with their in-house data supply policies and agreements. We have endeavoured to comply with all of them, particularly with regard to sensitivity of data, and data protection issues. We can assure data providers that we will not share data with third parties, or use data beyond the short life of this project, without receiving permission to do so. We also appreciate The greatest acknowledgement should go to the thousands of recorders who have gathered reptile records over the last century or more, and upon which this project relies entirely. Unfortunately the data sources are too numerous and idiosyncratic, and the data protection issues too complex, to list all recorders here. Amphibian and Reptile Conservation provided datasets from the following projects: Rare Species Database (incorporating rare species monitoring programmes), Add an Adder, Sliding Scales, and NARRS Widespread Species Surveys. John Baker (independent consultant) and Jon Cranfield (Herpetologic Ltd) provided access to data from the Make The Adder Count scheme, which operates in partnership with ARC and ARGUK. Louise Cranfield (Herpetologic Ltd) provided further analysis on the MTAC data. The following assisted with verification of Add an Adder data for their respective counties: John Baker of Suffolk Amphibian & Reptile Group, Ian Bradley (Lincolnshire & Northumberland records), Jon Cranfield of Essex Amphibian & Reptile Group (Essex & Hampshire records), Andy Harmer of Cheshire Active Naturalists Amphibian & Reptile Group, Chris Monk of Derbyshire Amphibian & Reptile Group, Andy Ryder of Avon Reptile & Amphibian Group (Somerset records). Staff of the National Biodiversity Network (Graham French, Paula Lightfoot) kindly assisted at short notice, by providing a custom download of reptile data for the whole of England, and processing onward requests to every Local Records Centre in England. The following Local Records Centres and other data holders (and their staff) provided improved access or resolution to their datasets: Dorset Environmental Records Centre (Jon Corkill, Carolyn Steele, Alison Stewart), Essex Field Club (Peter Harvey), Lincolnshire Environmental Records Centre (Charlie Barnes), North East Environmental Records CGO Ecology Ltd Adder Status Project (England) Mar

44 Information Centre (Katherine Pinnock), Nottinghamshire Biological & Geological Records Centre (Rob Johnson), Somerset Environmental Records Centre (Agni-Louiza Arampoglou), Staffordshire Ecological Record (Craig Slawson), Sussex Biodiversity Record Centre (Penny Green), Warwickshire Biological Records Centre (Chris Hill; unfortunately data not included in the analysis), Wiltshire & Swindon Biological Records Centre (Purgle Linham). The following local Amphibian & Reptile Group members provided additional data to the project: Lee Brady (Kent Reptile & Amphibian Group), Nic Butler & Janet Proctor (Devon Reptile & Amphibian Group, Jon Cranfield (Essex Amphibian & Reptile Group), Rod d Ayala & Angela Julian (Oxfordshire Amphibian & Reptile Group), Gemma Harding & Neil Pullen (Wiltshire Amphibian & Reptile Group), Andy Harmer (Cheshire Active Naturalists Amphibian & Reptile Group), Andrew Heaton (Leicestershire Amphibian & Reptile Group), Steve Langham (Surrey Amphibian & Reptile Group), Oli Robins & Chris Roughley (Dorset Amphibian & Reptile Network). Other third parties who supplied data: Gordon Barker (National Trust), Giles Alder (Berkshire Buckinghamshire & Oxfordshire Wildlife Trust), Jon Coote (British Herpetological Society). Some records were also derived from a database of preserved museum specimens created by Simon Maddock (Natural History Museum, London) as part of a separate ARC project. The largest datasets were from the BRC (accessed via the NBN Gateway) and ARC, with over 50,000 records each. Smaller datasets (typically up to 5,000 records each) were provided by LRCs and others. Only a fraction of the LRCs and ARGs in England were able to respond within the short timescale available, and this is acknowledged as one of the project s biggest constraint. The following participated in a questionnaire survey of adder hibernation sites: Lee Brady, Jon Cranfield, Chris Monk, Kevin Palmer. For use of anecdotes from Add an Adder: Lois Kemp, Frederick Facey, and Joy Douglas. Angela Julian (Secretary of ARGUK ably assisted the process of consultation among the ARGs). Consultation with various people added insights and substance to this project, including: John Baker, Lee Brady, Dave Carter, Chris Cathrine, Jan Clemons, Jon Cranfield, Chris Dresh, Paul Edgar, Trent Garner, Tony Gent, Richard Griffiths, Nigel Hand, Steve Langham, Chris Monk, John Newton, Gary Powell, David Sewell. CGO Ecology Ltd Adder Status Project (England) Mar

45 7. References Arnold, H.R. (1973). A preliminary atlas of amphibians and reptiles in Britain and Ireland. Huntingdon, BRC Monks Wood. Arnold, H.R. (1995). Atlas of amphibians and reptiles in Britain. London, ITE research publication 10, HMSO. Baker, J., Suckling, J. & Carey, R. (2004). Status of the adder Vipera berus and slow-worm Anguis fragilis in England. Peterborough, English Nature Research Report 546, English Nature. Barker, M. & Elliott, M. (2001). Sussex Amphibian & Reptile Group Millennium Report. A Great Leap Forward. Sussex Amphibian & Reptile Group. Beebee, T.J.C. & Griffiths, R.A Amphibians and reptiles. A natural history of the British herpetofauna. HarperCollins, London. Clemons, J A Provisional Atlas of the Amphibians and Reptiles of Warwickshire. Warwickshire Museum Service, Warwick. Cooke, A.S. & Arnold, H.R. (1982). National changes in status of the commoner British amphibians and reptiles before British Journal of Herpetology, 6, Cooke, A.S. & Scorgie, H.R.A. (1983). The Status of the Commoner Amphibians and Reptiles in Britain. Peterborough, Nature Conservancy Council Report 3, Nature Conservancy Council. Edgar, P., Foster, J. & Baker, J. (2010). Reptile Habitat Management Handbook. Amphibian and Reptile Conservation, Bournemouth. Frazer, D. (1983). Reptiles and Amphibians in Britain. Collins New Naturalist Series, London. Hilton-Brown, D. and Oldham, R.S. (1991). The status of the widespread amphibians and reptiles in Britain, 1990, and changes during the 1980 s, Peterborough, Nature Conservancy Council Report 131, Nature Conservancy Council. Inns, H. (2009). Britain s Reptiles and Amphibians. Wild Guides Ltd, Old Basing. Philp, E.G Provisional atlas of the amphibians and reptiles of Kent. Transactions of the Kent Field Club, 15(2): Prestt, I., Cooke, A.S. & Corbett, K.F British Amphibians and Reptiles. In: Hawkesworth, D.L. (ed.) The Changing Flora and Fauna of Britain. Academic Press, London, Reading, C.J., Buckland, S.T., McGowan, G.M., Jayasinghe, G., Gorzula, S. and Balharry, D. (1996). The distribution and status of the adder (Vipera berus L.) in Scotland determined from questionnaire surveys. Journal of Biogeography, 23 (5), CGO Ecology Ltd Adder Status Project (England) Mar

46 Sheldon, S Wyre Forest Adder Census and Report 1999, and Adder Population Trends Through the 1990s Correlated with Meteorological Data. Wyre Forest Study Group, pp Swan, M.J.S. and Oldham, R.S. (1993). Herptile sites volume 2: national common reptile survey final report. Peterborough, English Nature Research Report 38, English Nature. Taylor, R.H.R. (1948). The Distribution of Reptiles and Amphibia in the British Isles, with notes on Species recently introduced. British Journal of Herpetology, 1, Taylor, R.H.R. (1963). The Distribution of Amphibians and Reptiles in England, Wales, Scotland and Ireland and the Channel Islands: a Revised Survey. British Journal of Herpetology, 3 (5), Wright, S., Osborne, J., & Pursglove, C Frogs and Friends. The Distribution and Conservation of Amphibians and Reptiles in Nottinghamshire. Nottingham Natural History Museum, Nottingham. Wycherley, J. & Anstis, R Amphibians and Reptiles of Surrey. Surrey Wildlife trust, Woking. CGO Ecology Ltd Adder Status Project (England) Mar

47 8. Appendix 8.1. Excerpt from Memorandum of Agreement between Amphibian and Reptile Conservation and Natural England Delivery Profile Table Describing Themes, Actions, Outcomes and SMART Outputs for the MoA between ARC and Natural England (Developing Amphibian and Reptile Conservation in England, ). Species or Theme Adder Actions Outcomes SMART outputs Milestones and Responsibilities 1. Undertake a research project to evaluate the national status of the adder and develop monitoring metrics and protocols Collation of existing data and data sources Further analysis of exitig data and data from on-going projects Develop dialogue, perhaps via meetings/ workshops involving a small number of key players Map of actual distribution plus link to NCA Predictive mapping exploring different models to define current status Developed methodology for describing adder status at local and national level Provisional assessment of status of the adder in England and in National Character Areas. Further development of a monitoring network of recorders and increased coordination of existing monitoring programmes Report completed by end of March 2012 including: Distribution maps Liaison meetings, workshop or other communiation event by end February 2012 Description of monitoring protocols and approaches to address needs of all key data users & relating to conservation needs Recommendations for future recording and monitoring work Milestone Report end March 2012 Report Produced - ARC March 2012 ARC (TG) CGO Ecology Ltd Adder Status Project (England) Mar

48 8.2. Condition Assessment Model (CASM) developed by Natural Acuity Ltd By Steve Langham (Natural Acuity Ltd, Farnham) CASM Purpose The Conservation Condition Assessment Model (CASM) is an information management, analysis and exploitation toolset. It is used to determine the likely conservation condition for a species, using the assumption that available survey and distribution data will never be comprehensive. Conservation condition is a combination of the population dynamics, available habitat, and range of a species. CASM also calculates the likely trend in conservation condition for a given species across time. CASM 2.0 Development CASM 2.0 is a development from the CASM 1.0 model used for the assessment of the conservation condition assessment for all native reptile and amphibian species found within the county of Surrey. CASM Overview CASM uses a federation of Microsoft Access databases as the backbone for its information management system, although any database platform could be used. A database is a collection of information laid out in structured tables; queries that assist in the structuring of the data; forms for selecting variables such as geographic extent and species of interest; reports for outputting analysis; and code modules for the logical sequencing of analytical steps. The CASM 2.0 federation of databases are currently co-located on a single computer. However, the architecture of the model allows for remote compilation of data sources, such that thirdparties could be tasked with ongoing information management, and CASM 2.0 could remotely access this data, provided that sufficient connectivity was available via the internet. The diagram to the right illustrates the various databases that comprise the CASM 2.0 model. CGO Ecology Ltd Adder Status Project (England) Mar

49 CASM 2.0 Modules Species Distributions The Species Distributions module contains a variety of input dataset tables. By maintaining the format of source data, errors can be more easily traced, and the donor organisation appraised. Code modules validate the data in each source table, perform geotransformations to ensure commonality of coordinate reference system, and collate the source tables into a master Species Distribution table. The key fields within this master table are the sighting date, the grid reference, the source dataset, and the foreign record key Land Cover The Land Cover module contains various land cover datasets covering the British Isles Currently these tables comprise the CORINE 2000 Land Cover and CORINE 2006 Land Cover datasets. For each monad across the UK, it gives the percentage area for each of 44 land cover types. Only the cover for England was utilised in the current exercise. The CASM model currently only covers the UK (using the OSGB36 CRS grid), but could be extended to include the entire European Union. Geology The Geology module holds data from the British Geological Survey. For each monad across the United Kingdom, both bedrock ( solid ) and superficial ( drift ) geologies are recorded. This geology database is a proxy for soil types in most cases (to be added to CASM in future developments). Where superficial drift exists, it forms the surface geology, and determines the nature of the overlying habitats. Where superficial drift is absent, the bedrock is the surface geology, and determines the habitats. Area Templates The Area Template database holds a variety of spatial templates, such as the monad list for the country of England, monad lists for the Watsonian Vice Counties (VCs), and England s National Character Areas (NCAs). Templates can be drawn from this CASM module to define the area of analysis, or to categorise results by sub-areas (NCAs and VCs). Elevation and Aspect Although not enabled for the Adder Status Project (ASP), Elevation using Digital Terrain Elevation Data (DTED) is available, as is mean aspect (based on DTED analysis). The application of aspect to likely occupancy is not useful at monad resolution, but is exceptionally useful at hectare resolution. Climate Again, not enabled for the ASP study, this module holds climatic data from Météorologique Aviation Régulière (METAR) and climate means for the UK. As the adder s range extends across all of Great Britain, correlation to climate is not particularly relevant (unless climate CGO Ecology Ltd Adder Status Project (England) Mar

50 change is being assessed). However, such a module would be useful for assessing the conservation condition of smooth snake, sand lizard or grass snake. Species Analysis This is the core analysis module for CASM. Relevant tables are imported from the federation of databases, and the Species Analysis code performs a variety of ANOVA and weighted normalisation comparisons of characteristic data against known species distributions. Statistical significance tests are performed to ensure that characteristic data correlates to species distribution, and not to spurious factors. CASM Outputs CASM outputs are in two forms: Species Analysis Report This is the direct model output, in the form of a run-log file. It calculates the number of monads known to be occupied in each reporting epoch; the number of monads predicted to be occupied; and correlations to land cover types and geology. GeoFile Generation Many statistical analysis approaches can be used to generate an estimate of monad occupancy across an area. The benefit of CASM is that it defines exactly those monads (or other area units) that are likely to be occupied. The model automatically generates Keyhole Markup Language (KML) files, which can be viewed in Google Earth, Google Maps, Bing Maps, or other open-source GIS viewers. No bespoke software is required to view the precise geo-distributions generated by the model. CASM Strengths CASM is explicit in that it defines precise distributions, and generates easily-viewed geographic information. Unlike more generic approaches, the outputs are tightly strapped to the source data. As data quality improves, condition assessment also improves. Due to the statistical significance tests and correlation algorithms used, datasets for the same characteristics (e.g. Land Cover) do not need to be consistent in units nor type. The model adapts to whatever data format is presented. CASM Weaknesses Currently, there are insufficient land cover datasets to accurately depict trends over time. For instance, land cover data is not available prior to 1990; hence analysis for historical periods must rely upon recent land cover data. As semi-natural habitats have generally declined historically, the likely effect is an underestimation of the predicted baseline (historical) distribution. As with all models, the quality of output relies heavily on the quality of input data. Although great effort has been applied to obtaining and verifying reptile distribution datasets, the ASP study has been conducted under very challenging time constraints. CGO Ecology Ltd Adder Status Project (England) Mar

51 The model could be improved if more time were available to obtain more comprehensive input datasets. CGO Ecology Ltd Adder Status Project (England) Mar

52 8.3. Map of known current adder distribution in England CGO Ecology Ltd Adder Status Project (England) Mar

53 8.4. Map of known historic adder distribution in England CGO Ecology Ltd Adder Status Project (England) Mar

54 8.5. Map of predicted current adder distribution in England CGO Ecology Ltd Adder Status Project (England) Mar

55 8.6. Map of predicted historic adder distribution in England CGO Ecology Ltd Adder Status Project (England) Mar

56 8.7. Map of known current and historic adder distribution in England CGO Ecology Ltd Adder Status Project (England) Mar

57 8.8. Map of predicted current and historic adder distribution in England CGO Ecology Ltd Adder Status Project (England) Mar

58 8.9. Map of known current and historic adder distribution in Northern England CGO Ecology Ltd Adder Status Project (England) Mar

59 8.10. Map of predicted current and historic adder distribution in Northern England CGO Ecology Ltd Adder Status Project (England) Mar

60 8.11. Map of known current and historic adder distribution in Central England Map of predicted current and historic adder distribution in Central England CGO Ecology Ltd Adder Status Project (England) Mar

61 CGO Ecology Ltd Adder Status Project (England) Mar

62 8.13. Map of known current and historic adder distribution in Southwest England Map of predicted current and historic adder distribution in Southwest England CGO Ecology Ltd Adder Status Project (England) Mar

63 CGO Ecology Ltd Adder Status Project (England) Mar

64 8.15. Map of known current and historic adder distribution in Southeast England Map of predicted current and historic adder distribution in Southeast England CGO Ecology Ltd Adder Status Project (England) Mar

65 CGO Ecology Ltd Adder Status Project (England) Mar

66 8.17. Map showing paucity of adder records (predicted current and historic distribution) in the southern Pennines Map showing suspicious predictive patterns ( buffer syndrome ) in Herefordshire area CGO Ecology Ltd Adder Status Project (England) Mar

67 8.19. Map showing forestry plantations in the Breckland area of East Anglia, illustrating apparent flaws in the predictive model CGO Ecology Ltd Adder Status Project (England) Mar

68 8.20. Map showing predictive model working well in Dorset and the New Forest CGO Ecology Ltd Adder Status Project (England) Mar

69 8.21. Table of CORINE Land Cover Types GRID_CLC_COLABEL1 LABEL2 LABEL Artificial surfaces Urban fabric Continuous urban fabric Artificial surfaces Urban fabric Discontinuous urban fabric Artificial surfaces Industrial, commercial and transport units Industrial or commercial units Artificial surfaces Industrial, commercial and transport units Road and rail networks and associated land Artificial surfaces Industrial, commercial and transport units Port areas Artificial surfaces Industrial, commercial and transport units Airports Artificial surfaces Mine, dump and construction sites Mineral extraction sites Artificial surfaces Mine, dump and construction sites Dump sites Artificial surfaces Mine, dump and construction sites Construction sites Artificial surfaces Artificial, non-agricultural vegetated areas Green urban areas Artificial surfaces Artificial, non-agricultural vegetated areas Sport and leisure facilities Agricultural areas Arable land Non-irrigated arable land Agricultural areas Arable land Permanently irrigated land Agricultural areas Arable land Rice fields Agricultural areas Permanent crops Vineyards Agricultural areas Permanent crops Fruit trees and berry plantations Agricultural areas Permanent crops Olive groves Agricultural areas Pastures Pastures Agricultural areas Heterogeneous agricultural areas Annual crops associated with permanent crops Agricultural areas Heterogeneous agricultural areas Complex cultivation patterns Agricultural areas Heterogeneous agricultural areas Land principally occupied by agriculture, with significant areas of natural vegetation Agricultural areas Heterogeneous agricultural areas Agro-forestry areas Forest and semi natural areas Forests Broad-leaved forest Forest and semi natural areas Forests Coniferous forest Forest and semi natural areas Forests Mixed forest Forest and semi natural areas Scrub and/or herbaceous vegetation associations Natural grasslands Forest and semi natural areas Scrub and/or herbaceous vegetation associations Moors and heathland Forest and semi natural areas Scrub and/or herbaceous vegetation associations Sclerophyllous vegetation Forest and semi natural areas Scrub and/or herbaceous vegetation associations Transitional woodland-shrub Forest and semi natural areas Open spaces with little or no vegetation Beaches, dunes, sands Forest and semi natural areas Open spaces with little or no vegetation Bare rocks Forest and semi natural areas Open spaces with little or no vegetation Sparsely vegetated areas Forest and semi natural areas Open spaces with little or no vegetation Burnt areas Forest and semi natural areas Open spaces with little or no vegetation Glaciers and perpetual snow Wetlands Inland wetlands Inland marshes Wetlands Inland wetlands Peat bogs Wetlands Maritime wetlands Salt marshes Wetlands Maritime wetlands Salines Wetlands Maritime wetlands Intertidal flats Water bodies Inland waters Water courses Water bodies Inland waters Water bodies Water bodies Marine waters Coastal lagoons Water bodies Marine waters Estuaries Water bodies Marine waters Sea and ocean NODATA NODATA NODATA UNCLASSIFIED UNCLASSIFIED LAND SURFACE UNCLASSIFIED LAND SURFACE UNCLASSIFIED UNCLASSIFIED WATER BODIES UNCLASSIFIED WATER BODIES UNCLASSIFIED UNCLASSIFIED UNCLASSIFIED Table of component datasets included in current analysis CGO Ecology Ltd Adder Status Project (England) Mar

70 Dataset included in current analysis No. of records Notes Amphibian & Reptile Conservation - Add an Adder (adder data only) 406 verified Only eight English counties verified Amphibian & Reptile Conservation - Add an Adder (snake/slow-worm data) 4137 unverified Whole British dataset, including verified below Amphibian & Reptile Conservation - NARRS reptile survey sites 271 verified All reptile survey sites in Britain Amphibian & Reptile Conservation - Rare Species Database up to verified Mainly English rare species monitoring Amphibian & Reptile Conservation - Sliding Scales (snake/slow-worm data) 947 unverified Britain Andy Harmer - adders from current Cheshire herpetofauna atlas project 54 verified ARC/ARG UK - Make The Adder Count adder monitoring sites 223 verified Britain Chris Gleed-Owen/CGO Ecology Ltd - reptile data not yet held elsewhere 251 verified England Dorset Environmental Record Centre - adders only 2574 verified Environmental Record Information Centre for the North East - all reptile data 956 verified Essex Field Club - all reptile data 825 verified Jonathan Cranfield/Herpetologic Ltd - all reptile data 782 verified England Lincolnshire Environmental Records Centre - all reptile data 1642 verified Many tetrad coords; used SW monad National Biodiversity Network Gateway - all reptile data verified England only. Contains LRCs, BRC, ARC datasets National Trust - all reptile data from NT properties 476 verified Improved some 10km to 1km (England) Natural History Museum, London - preserved specimens with provenance 30 verified Only English ones with provenance Natural History Museum, London - preserved specimens with provenance 40 verified Only English ones with provenance Nottinghamshire Biological & Geological Records Centre - all reptile data 584 verified Somerset Environmental Record Centre - all reptile data 1813 verified Improved hundreds from 10km to 1km Staffordshire Ecological Record - all reptile data 1233 verified Sussex Biological Records Centre - all reptile data 9356 verified Tim Sandles -'Lore of the Adder' webpage 5 verified Dartmoor expert; places he has seen adders Various individuals - reptile data 26 verified England Total Including many duplicates CGO Ecology Ltd Adder Status Project (England) Mar

71 8.23. Line graph showing Make The Adder Count data for 16 sites monitored from CGO Ecology Ltd Adder Status Project (England) Mar

72 8.24. Questionnaire survey results: plant species associated with adder hibernation sites Questionnaire survey results: site features associated with adder hibernation CGO Ecology Ltd Adder Status Project (England) Mar

73 8.26. Questionnaire survey results: other reptile species associated with adders Questionnaire survey results: incorporation of adders into site management plans CGO Ecology Ltd Adder Status Project (England) Mar

74 CGO Ecology Ltd Adder Status Project (England) Mar

75 8.28. Map showing the predicted current ( ) range of the adder in England Key to map Area in scope (England) Occupied monads CGO Ecology Ltd Adder Status Project (England) Mar

76 8.29. Comparison between condition assessment based known survey data and predicted condition using the CASM 2.0 model to compensate for inadequacy of survey effort CGO Ecology Ltd Adder Status Project (England) Mar

77 8.30. Table showing monad loss between historic and current epochs per Vice County Historic (CASM 2.0) Current (CASM 2.0) Population Occupied 15,154 Monads Population 363,700 Estimate Habitat Suitable Monads Unknown Range Occupied Vice Counties 58 (all) Population Occupied 9,237 Monads Population 221,700 Estimate Habitat Suitable Monads 31,894 Range Occupied Vice Counties 58 (all) Change in conservation condition for the adder in England Historic Current Delta Absolute Delta % Occupied Monads 15,154 9,237-5,917-39% Population Estimate 363, , ,000-39% Range (VCs) * 0 0 * Overestimated due to spurious data. Adders are thought to be currently extinct in between 3-6 VCs. CGO Ecology Ltd Adder Status Project (England) Mar