New Forest visitor survey report

Summary

This report presents the results of visitor survey work across the New Forest SPA/SAC/Ramsar in 2018/2019. The survey has been commissioned by Test Valley Borough Council on behalf of a partnership which also includes Eastleigh Borough Council, New Forest District Council, New Forest National Park Authority, Southampton City Council and Wiltshire Council. The work has been supported by Natural England and Forestry England.

Surveys took place at 60 locations, mostly car parks across the New Forest SPA/SAC/Ramsar. At each location, 40 hours of survey work were conducted, split to cover a weekday and a weekend day in the autumn/winter (October – November), a weekday and a weekend day in the spring (April – May) and a single day in the summer (late July – August, school holiday period). Surveys were broken into two-hour periods that were spread to cover different parts of the day (i.e. including early mornings and late evenings). Tally counts of people, dogs, bikes and horses were maintained for each two-hour survey period and a random selection of people seen were approached and interviewed.

Key findings included:

• The counts generated an hourly rate of 4.7 groups, 9.7 people (including minors – those thought to be under 18 years old), 3.7 dogs, 1.6 minors, 0.9 bikes and 0.1 horses entering/passing through the survey point, averaged across all locations and seasons.
• Typical group size from the tally counts was 2.1 people, accompanied by 0.8 dogs and including 0.3 minors.
• In total 5,236 interviews were conducted.
• 83% of interviewees were on a short visit directly from home that day. Those staying away from home on holiday accounted for 14% of interviewees and a further 2% were staying with friends or family.
• During the summer there were relatively more holiday makers (22%) and fewer day visitors (76%), compared to the spring (12% and 85% respectively) and the winter (11% and 86%).
• For most interviewees the main activity was given as either dog walking (55%) or walking (26%). No other single main activity was named by more than 5% of interviewees.
• Dog walking was very much focussed around the peripheral areas of the SPA/SAC, while walking (without a dog) was the most common main activity at the more central survey locations. Cyclists were interviewed at scattered locations but notably those around Brockenhurst and also at Burbush Hill
• Overall, 61% of interviewees were accompanied by at least one dog and the maximum number of dogs per interviewed group was 12. In total, 4,807 dogs were counted accompanying interviewees, giving an average of 0.9 dogs per interviewee.
• Dog walkers accounted for a slightly lower proportion of visitors in the summer compared to the other times of year. 60% of dogs were seen off lead by the surveyor during the interview.
• 26% of interviewees tended to visit the New Forest SPA/SAC/Ramsar site on a daily basis.
• Dog walkers were the most regular visitors, with 41% of dog walkers indicating they visited on a daily basis and a further 14% of dog walkers indicating they visited more than once per day.
• Typical visit duration for all interviewees was around 95 minutes. Those visiting to play golf and for Duke of Edinburgh tended to visit for longer and those dog walking and running typically had relatively short visits.
• 64% visited equally all year round and did not tend to visit at a particular time of year.
• 90% had arrived by car/van or other motor vehicle.
• Reasons given for the choice of specific location to visit that day included: close to home (or work or holiday accommodation) (25%), previous knowledge or familiarity (16%), quick & easy travel route (10%), scenery/variety of views (10%) and for a change/variety (10%). Some 2% had stopped at random and 1% had been deflected from other locations because they were full or because the car park was shut.
• Interviewee routes were mapped as part of the interview. Route lengths tended to be shorter in the summer (for dog walkers and cyclists at least) and cyclists tended to do much longer routes than the other activities.
• Across all seasons, the typical (median) dog walk was 2.7km, typically extending to 922m from the start point. For walkers the equivalent values were 3.2km and 1,004m and for cyclists 12km and 2,828m.
• Factors influencing the choice of route included: previous knowledge of the location (22%), time available (13%), weather conditions (such as shade or shelter etc.,12%), following a marked trail or the paths available (12%) and activity specific factors (such as where the buggy could go, golf course etc., also 12%).
• Maps were the most commonly cited type of information used to plan interviewee’s visits (15% of interviewees), followed by websites (8%) and recommendations from friends or family (7%).
• 67% of interviewees were aware of a wildlife habitat or species that could be affected by recreation and could give a named example. Breeding birds (including ‘ground-nesting birds’) were the most commonly named concern (40% of interviewees).
• For those interviewees on a short visit or day trip, travelling directly from home that day, 41% indicated that all their visits for their chosen activity took place within the New Forest SPA/SAC/Ramsar.
• A wide range of other, alternative locations were given. Those most frequently cited included Hengistbury Head (4%), Lepe Country Park (3%), Barton-on-sea/Barton-on-sea beach (2%), Purbeck (2%), Lymington Marshes (2%), Highcliffe/Highcliffe Beach (3%), Southampton Common (2%), South Downs (2%) and Bournemouth Beach (2%).
• The was little difference in the proportion of interviewees that would use a new Country Park or improved footpath network away from the New Forest SPA/SAC/Ramsar site, suggesting relatively little difference in these as mitigation approaches.
• 4,871 interviewees (91%) gave a full, valid UK postcode that could be geocoded using the national database.
• The Bournemouth/Poole conurbation was the single built-up area from which the most interviewees originated (12%), with the South Hampshire built-up area second (9%).
• 20% of interviewees on a short visit or day trip from home that day gave postcodes within the National Park boundary. A further 40% came from outside the National Park but within the New Forest District. Other local authorities accounted for relatively small proportions of the interviewees in comparison.
• 62% of interviewees lived within a 5km radius of the New Forest SPA/SAC/Ramsar site boundary. The median distance for all interviewees from their home postcode to the interview location was 7.75km and 75% originated from within 21.4km; for those on a short visit/day trip from home, the equivalent values were 6.1km and 13.8km.

Contents

Overview....................................................................................................................................... 57
By built-up area........................................................................................................................... 62
By local authority ........................................................................................................................ 63
Linear distance (‘as the crow flies’)............................................................................................ 67

Acknowledgements

This report was commissioned by Test Valley Borough Council on behalf of a number of different local authorities. We are grateful to Karen Eastley (Test Valley Borough Council) for overseeing the work and our thanks to other members of the steering group that included: Richard Burke (Forestry England), Adam Egglesfield (Hampshire County Council), Louise Evans (New Forest District Council), Dawn Heppell (Eastleigh Borough Council and Southampton City Council), David Illsley (New Forest National Park Authority), Louisa Kilgallen (Wiltshire Council), Nigel Matthews (New Forest National Park Authority), Graham Smith (Test Valley Borough Council) and Nick Squirrell (Natural England).

Fenella Lewin (Footprint Ecology) co-ordinated the survey team. Survey fieldwork was undertaken by Footprint Ecology staff and a range of sub-contracted surveyors. We are grateful to Forestry England, Hampshire County Council, the National Trust, New Forest District Council and Wellow Parish Council for permission to survey on their land and our thanks also to all those who stopped to be interviewed and shared information with us as part of the survey.

1. Introduction

Overview

This report, commissioned by a partnership of local authorities with funding from central government, is part of a series that relates to understanding the impacts of recreation (arising from new housing development) on the New Forest international nature conservation designations. The various studies are intended to inform necessary mitigation approaches.

In this report we present the results of visitor surveys carried out within the boundaries of the New Forest international nature conservation designations. The work provides detailed information on visitor profiles, access patterns and visitor origins.

Relevant legislation

The designation, protection and restoration of key wildlife sites is embedded in the Conservation of Habitats and Species Regulations 2017 (as amended), which are commonly referred to as the ‘Habitats Regulations.’ These Regulations are in place to transpose European legislation set out within the Habitats Directive (Council Directive 92/43/EEC), which affords protection to plants, animals and habitats that are rare or vulnerable in a European context, and the Birds Directive (Council Directive 2009/147/EC), which originally came into force in 1979, and which protects rare and vulnerable birds and their habitats. These key pieces of European legislation seek to protect, conserve and restore habitats and species that are of utmost conservation importance and concern across Europe. European sites include Special Areas of Conservation (SACs) designated under the Habitats Directive and Special Protection Areas (SPAs) classified under the Birds Directive. Ramsar sites, those wetlands of international importance that are listed in the Ramsar Convention are, through government policy, are also treated as European sites.

Public bodies, including local planning authorities, have specific duties in terms of avoiding deterioration of habitats and species for which sites are designated or classified, and stringent tests have to be met before plans and projects can be permitted. Importantly, the combined effects of individual plans or projects must be taken into account. For local planning authorities, this means that the combined effect of individual development proposals needs to be assessed collectively for their cumulative impact, as well as on an individual basis.

The New Forest

The New Forest is one of the largest tracts of semi natural vegetation in the country, and as such is one of our most important wildlife sites. The area hosts three international wildlife site designations and is closely located to other international wildlife sites such as the Solent and Southampton Water.

The New Forest is classified as a SPA for its breeding and overwintering bird species of European importance, in accordance with the European Birds Directive. The designation relates to internationally significant breeding populations of Dartford Warbler Sylvia undata, Nightjar Caprimulgus europaeus, Woodlark Lullula arborea, Honey Buzzard Pernis apivorus, Hobby Falco subbuteo and Wood Warbler Phylloscopus sibilatrix and over-wintering Hen Harrier Circus cyaneus.

The New Forest is also designated as a SAC for its habitats and non-avian species of European importance, in accordance with the European Habitats Directive. This designation reflects the unique mosaic of habitats across the New Forest, which includes eight Annex 1 heathland, grassland, woodland, wetland, bog and open water habitats, together with three Annex 2 species, Stag Beetle Lucanus cervus, and Southern Damselfly Coenagrion mercuriale, and Great Crested Newt Triturus cristatus.

Also relevant is the New Forest’s listing as a Ramsar site, under the Ramsar Convention. This recognises the international importance of the site as a wetland, supporting wetland flora and fauna of international importance, and adding to the global network of Ramsar listed wetlands.

Housing growth and recreation impacts

A challenging issue for UK nature conservation is how to respond to increasing demand for access without compromising the integrity of protected wildlife sites. Areas that are important for nature conservation are often important for a range of other services, including the provision of space for recreation for an increasing population. Such recreation space can be used for a wide variety of activities, ranging from the daily dog walks to competitive adventure and endurance sports.

There is now a strong body of evidence showing how increasing levels of access can have negative impacts on wildlife. Visits to the natural environment have shown a significant increase in England as a result of the increase in population and a trend to visit more (O’Neill, 2019). The issues are particularly acute in southern England, where population density is highest. Issues are varied and include disturbance, increased fire risk, contamination and damage (for general reviews see: Liley et al., 2010; Lowen, Liley, Underhill-Day, & Whitehouse, 2008; Ross et al., 2014; Underhill-Day, 2005).

The issues are not however straightforward. It is now increasingly recognised that access to the countryside is crucial to the long term success of nature conservation projects, for example through enforcing pro-environmental behaviours and a greater respect for the world around us (Richardson, Cormack, McRobert, & Underhill, 2016). Access also brings wider benefits to society that include benefits to mental/physical health (Keniger, Gaston, Irvine, & Fuller, 2013; Lee & Maheswaran, 2011; Pretty et al., 2005) and economic benefits (ICF GHK, 2013; ICRT, 2011; Keniger et al., 2013; The Land Trust, 2018). Nature conservation bodies are trying to encourage people to spend more time outside and government policy is also promoting countryside access in general (e.g. through enhancing coastal access).

There are two statutory purposes for national parks in England and Wales. The first is to conserve and enhance natural beauty, wildlife and cultural heritage and the second is to promote opportunities for the understanding and enjoyment of the special qualities of national parks by the public. This second purpose includes opportunities for open air recreation. However, if it appears that there is a conflict between the two National Park purposes, the Environment Act 1995 requires greater weight to be attached to the purpose of conserving and enhancing the natural beauty, wildlife and cultural heritage of the National Park (this is known as the Sandford Principle¹). When national parks carry out these purposes, they also have the duty to encourage the social and economic well-being of local communities within the national park.

There is therefore, a significant challenge: to avoid or mitigate potential negative impacts associated with recreation so as to comply with legislation without compromising the ability of people to be outside enjoying sites for recreation.

¹ Named after Lord Sandford, who chaired the 1974 National Parks Policy Review Committee.

Aims of this work

The New Forest has a particular draw for recreation, and it is unique in scale and the recreation opportunities it provides. Previous work has considered the recreation impacts and links to new development (Fearnley, Hoskin, Liley, White, & Lake, 2012; e.g. Sharp, Lowen, & Liley, 2008) and also highlighted the range of use that includes both use by local residents living in or near the New Forest as well as visitors from a wide area including tourists. Previous visitor survey work on the New Forest (Tourism South East Research Services & Geoff Broom Associates, 2005) extended to the whole National Park estimated 13.5 million visitor days to the New Forest. More recent work (RJS Associates Ltd., 2018) estimated that figure had increased to 15.2 million visitor days for recreation and leisure in 2017 and estimated it could rise to 17.6 million visitor days by 2037.

In order to better understand the relative draw of the New Forest and the links between housing and recreation use, the on-site visitor survey was commissioned to:

• To understand the types of recreational use and patterns of access by visitors to the heathland and woodland parts of the New Forest;
• To better understand where recreational visits to the New Forest originate from;
• To be able to establish the links between where people live and the patterns of recreational use.

Other reports

This report solely relates to on-site visitor work involving counts of people and interviews with visitors at locations across the heathland and woodland parts of the New Forest. The work forms part of a series of reports that relates to understanding the impacts of new development on the New Forest international nature conservation designations. The project as a whole involves visitor surveys combined with work to understand the impacts of recreation and relevant mitigation approaches. Other reports, produced in parallel with this one, include:

• Recreation use of the New Forest SAC/SPA/Ramsar: New Forest vehicle counts 2018/19 – results of vehicle counts across the New Forest SAC/SPA/Ramsar car parks, counting all parked vehicles on a range of different dates over a year;
• Recreation use of the New Forest SAC/SPA/Ramsar: Results of a telephone survey with people living within 25km - the results of a telephone survey with 2,000 residents living within a 25km radius of the woodland/heathland areas of the New Forest SAC/SPA/Ramsar;
• Recreation use of the New Forest SAC/SPA/Ramsar: Overview of visitor results and implications of housing change on visitor numbers - a summary of the visitor survey results, drawing the findings from the telephone survey, on-site survey and vehicle counts together and making predictions for change in recreation as a result of new housing.
• Recreation use of the New Forest SAC/SPA/Ramsar: Impacts of recreation and potential mitigation approaches – sets out the impacts of recreation and provides options for mitigation and avoidance.

This study is the first that has specifically considered visitors to the New Forest SAC/SPA/Ramsar. Previous work looking at visitor numbers and their activities and impacts in the New Forest includes A Survey of Recreation Visits to the New Forest National Park (Tourism South East Research Services & Geoff Broom Associates, 2005).

2. Methods

Visitor survey work involved interviews and counts of people at a sample of locations across the New Forest SPA/SAC. The counts provide an overview of visitor flows at each point and the visitor interviews, involving a random sample of people, provide data on visitor origins, visitor profile and factors that influence behaviour.

Selection of survey points

60 survey points were selected, based on the following criteria:

• They provided access onto the New Forest SPA/SAC;
• They provided a good geographic spread;
• Locations included those easily accessible from outside the National Park and on the main routes and arteries into the New Forest, i.e. directly accessible from settlements outside the New Forest;
• They broadly represented the different types of access points, including formal car parks, more informal parking and pedestrian access direct from housing;
• They worked well from a practical perspective, for example to safely and easily intercept visitors to interview.

An initial selection of 50 potential locations was made based on GIS data and using the following steps:

1. Car park data were combined in the GIS to give a single file of car parking locations within the National Park focussed around the New Forest SPA/SAC. This layer was derived from:
   • GIS data provided by the steering group (including 128 Forestry England car parks and 14 other car parks);
   • 134 informal parking locations (lay-bys etc.) within the New Forest SPA/SAC and identified from aerial photographs.
2. These car parks were then categorised according to which part of the New Forest they fell within, based on three broad sections (north of A31; south of the A31 and east of A337 (Lyndhurst–Brockenhurst) or south of the A31 and west of the A337). The number of locations chosen within each sector was directly in proportion to the number of parking locations identified in each, giving 15 locations in the north, 14 in the south and 21 in the south-west
3. Locations within each sector were identified based on the parking capacity, i.e. ranking by capacity and selecting locations based on a stratification according to the proportion of the total capacity within the sector.

This initial map was shared with the steering group and provided the foundation for the final selection, following advice and discussion. Locations that were removed included those with seasonal closures (such as the Reptile Centre). Points were added where there were clear gaps in the spatial coverage and in particular to include locations around the periphery and along major routes (i.e. accessible for people living outside the New Forest SPA/SAC). Two additional locations were added that were predominantly pedestrian access points, with the survey point close to housing and in a good location to intercept visitors and in addition Bolderford Bridge was included – this has no parking adjacent and is a river crossing point reasonably close to Brockenhurst. It is also on a cycle route and therefore a good location to intercept people passing, potentially including residents from Brockenhurst.

The final selection of locations are shown in Map 1 and Table 1.

Table 1: Survey locations

Survey approach

At each survey point, surveyors maintained a tally of visitors using the site during the survey periods, recording numbers of groups, people, minors², horses, cycles and dogs. Counts were typically split into those entering, leaving or passing through, but these varied with the location. Most locations were car parks, where someone ‘entering’ would be leaving the car park to access the surrounding countryside and someone ‘leaving’ would be returning to their vehicle. Those who cycled or walked through the car park were recorded as ‘passing through’.

A record was also kept of the number of refusals, i.e. where someone was approached and declined to be interviewed.

Surveyors conducted face to face interviews with a random selection of visitors (by selecting the next person they see after completing the previous interview). Only one person per group were interviewed, and no unaccompanied minors were.

Interviews were conducted on tablets hosting SNAP survey software. The questionnaire (Appendix 1) included questions relating to access patterns and behaviour, visitor profile, home postcode and the route taken. At the end of the interview the surveyor also recorded – from observation – information relating to the group size, gender of interviewee, number of dogs in group and whether dogs were seen off lead.

Routes taken by respondents (or planned to be taken if they were just setting off) were recorded by drawing the visitor’s route on a paper map linked by a unique reference number to the SNAP questionnaire. The routes were plotted interactively, with the surveyor checking with the interviewee regarding landmarks and features on the ground. Different scale maps were available allowing long routes (e.g. cycling etc.) to be plotted as well as very short walks. Routes were later digitised to give a polyline in GIS as a single polyline for each interview.

² Those thought to be under 18 years old

Survey timings

Each survey point was surveyed for:

• A weekday and a weekend day in the autumn/winter (October – November)
• A weekday and a weekend day in the spring (April – May)
• A single day in the summer (late July – August, school holiday period), with 30 survey points surveyed on weekdays and 30 survey points on weekend days.

Survey effort was spread across the survey dates set out above, ensuring the risk of bad weather is minimised. As far as possible survey visits were rescheduled to avoid particularly adverse weather conditions.

Each ‘day’ of fieldwork involved 8 hours of survey work, split into two-hour sessions. These reflected daylight hours:

• Autumn: 0700-0900; 0930-1130; 1230-1430; 1500-1700;
• Spring: 0700-0900; 1000-1200; 1300-1500; 1600-1800;
• Summer: 0700-0900; 1000-1200; 1400-1600; 1700-1900.

This level of survey effort meant that each location was surveyed for a total of 40 hours, with comparable weekend/weekday effort and timings for the winter and spring.

Horseshoe Bottom (location 5) was closed during the spring surveys and therefore the nearest alternative car park (Longslade Bottom) was surveyed instead. Throughout the results, where we summarise or present data by survey point, we do not include Longslade Bottom as a separate location and instead “Horseshoe Bottom” results include the Longslade Bottom data.

Data presentation and analysis

Tally data were collated and summarised as people/groups/dogs/bikes/horses entering, leaving or passing through. In order to give hourly rates of people going through the access point we combined people passing and people entering, broadly representing the number of people spreading out from the survey point into the woodland/heathland.

Activity-type was recorded in the questionnaire based on the response of the interviewee, rather than any interpretation of the surveyor (i.e. no attempt was made to ensure all those with binoculars were logged as wildlife watching or those with dogs logged as dog walking).

Route data were digitised as polylines with a GIS (QGIS version 3.4.5). For each polyline we calculated the overall length, the length that was within the New Forest SPA/SAC and also the ‘penetration’ distance, which was calculated as the distance between the interview location and the mid-point of the polyline. This broadly represents (for those people doing a circular route from a car park or other entry point) how far people ranged from the entry point before turning back, or for those doing a linear route from one location to another, the mid-point of that route.

Interviewees who indicated they visited other sites (in Q17, see Appendix 1) were subsequently asked to name another location they visit for the given activity. Responses were recorded as free text and subsequently checked and standardised to correct spelling and ensure consistency. Where the same location was clearly referred to under different names (e.g. ‘Moors Valley’, ‘Moors Valley Country Park’) a single, consistent name was used. Where there was some ambiguity and it was not clear the same location was being referred to, different names were retained (e.g. ‘Bournemouth’ and ‘Bournemouth Beach’). Lymington Marshes was however used as a consistent name to encompass 'Lymington sea wall', 'Lymington salterns', 'Lymington coast' and 'Lymington seafront'. Pennington Marshes and Keyhaven Marshes were kept separate, although it's likely there will be some overlap between all three. We used Hurst Spit to include both Hurst Castle and Hurst Beach and similarly Lepe Country Park was used for responses such as ‘Lepe’, ‘Lepe Beach’ and ‘Lepe Country Park’. We also grouped responses according to the following broad categories of site:

• Coast: only if the interviewee's answer included 'beach', 'coast' or 'cliffs', unless it is obvious from the site name that it is by the coast e.g. Hengistbury Head.
• Other NP: if the interviewee mentioned another UK national park, or a site e.g. Old Winchester Hill that is known to be within another national park
• Country Park: flags those sites that are country parks e.g. Moors Valley
• Other parts of New Forest: only those sites that are within the New Forest SPA/SAC boundary e.g. Hatchet Pond, Deerleap
• Long distance path: national trail or long distance footpath e.g. South West Coast Path, Itchen Way
• Vague/unspecified: not enough information to identify a single site e.g. local streets, local beaches. Also includes large areas such as 'Dorset', 'Purbeck', 'Jurassic Coast'.

Home postcodes given by interviewees were georeferenced against a national database, allowing the visitor origins to be mapped. Where we have summarised postcode data by settlement we have used the built-up area data available from the Office of National Statistics³.

³ https://geoportal.statistics.gov.uk/datasets/built-up-areas-december-2011-boundaries-v2

All data analysed with statistical tests were not normally distributed (usually positively skewed, with a small number of very high outlier values), and therefore we used non-parametric tests and typically give median values. Box plots are often used to explore these data and show median values (horizontal lines), interquartile range of 25-75% of the data values (boxes) and the upper and lower limits of the data range (whiskers). Mean values have been overlaid, shown as crosses, and all data points as circles.

The tally data were analysed using generalised linear models (GLMs) in the computer program ‘r’⁴. A global model was produced, combining a large combination of initial parameters, which was then interrogated using the ‘dredge’ function in the MuMIn package⁵. Dredging explores all potential models which can be constructed from varying combinations of the parameters included in the global model. Each model’s goodness-of-fit was then assessed using the Akaike Information Criterion adjusted for small sample sizes (AICc), and the model with the lowest AICc value selected. Significant parameters within the final model were also identified.

The global model included the following parameters, in addition to terms describing interactions between them: season; day type (weekday/weekend); survey location name; time period (early/late morning and early/late afternoon); rainfall level; cloud cover, and temperature (cool, mild, warm and hot). In order to account for the different number of parking spaces available at each interview location, an offset of log(capacity) was applied. Weights were also applied to the data used in the model to account for the greater number of surveys carried out in the winter and spring than in the summer months.

In the final analysis section, we have use multiple metrics from the visitor data to identify different types or groups of visitors. Visitor data were summarised by survey point and then by distance bands. 10 distance bands were used, with the band width determined by sample size to give an equal count i.e. the same number of interviewees in each band. The data were summarised as the count of interviewees in each group within the metric examined. Values were examined as a correlation matrix which allowed identification of metrics which were highly related to each other. Visitor metrics were also analysed using a Principal Component Analysis (PCA). This approach transforms a large number of variables into a smaller set of variables or principal components. For the summary by survey point, where samples sizes (interviewees) were not equal, we used the number of samples as a weighting. The PCA creates a series of principal component variables with combined ever-increasing overall explanation of the variation. The first two variables are used as the two main axes in PCA plots, (the percentage of variability for each axis is given on the axis label). Variables used in the PCA are shown on the graph as red arrows, which show the direction of the variable within the two axes. Informed by the results of the PCA, we categorised visitors in set groups and discuss the visitor profiles created.

⁴ R Core Team (2018). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria.

⁵ Kamil Barton (2019). MuMIn: Multi-Model Inference. R package version 1.43.6.

3. Results: tally counts

In total 9,593 groups, comprising 19,713 people, were counted entering during the survey work, across all seasons (Table 2). The number of dogs counted entering was 8,335 and there were also 3,207 minors, 1,109 bikes and 76 horse riders. Counts were also made of people leaving and passing through (see Table 2).

These count data are derived from 2,400 hours of fieldwork. Using the combined data for entering and passing through gives an hourly rate of 4.7 groups, 9.7 people (including minors), 3.7 dogs, 1.6 minors, 0.9 bikes and 0.1 horses per survey hour, averaged across all locations and seasons. These figures indicate a typical group size of 2.1 people, accompanied by 0.8 dogs and including 0.3 minors.

Table 2: Summary of tally data by season and type of day

Note that the data for winter and spring are directly comparable (same level of survey effort at each location) whereas during the summer the survey effort was different. WD = weekday; WE = weekend day.

Tally data are summarised by survey point in Map 2. The top map shows relative proportions of counts for the winter and spring (both with 16 hours of survey) and the summer (8 hours survey effort), with the size of the circles reflecting the total numbers of people counted (including minors) both entering and passing through. There is relatively little evidence of particular clusters with similar patterns, however the central survey points around Brockenhurst and Lyndhurst all have relatively high counts for the summer surveys. The lower map shows the number of groups, dogs, minors, cyclists and horse riders counted at each location. It can be seen that the peripheral locations, particularly those towards Southampton Water on the eastern side of the Forest and those near Ringwood, have high numbers of dogs, with some locations averaging more than 1 dog per group. Locations with high counts of cyclists were typically around Brockenhurst while there were no locations with relatively high numbers of horse riders.

Modelling of the combined numbers of individuals entering and passing each location identified a number of significant relationships, with the model output statistics provided in Appendix 2. It should be noted that 2 of the interview survey locations (Bolderford Bridge and Heath Roundabout Pegasus Crossing) were excluded from the global model, as it was not possible to estimate a capacity value for these ‘roving’ locations.

The final model indicated that larger numbers of individuals are likely to use the interview locations on weekends, and on dry, warm or hot, days with limited cloud cover rather than on wet, cold, and cloudy days. Furthermore, more individuals will use the locations in the late morning, and fewer will do so in the early morning or late afternoon.

The retention of individual locations in the model was usually associated with a positive relationship with the number of individuals using the locality. Nevertheless, various interaction terms were also retained, which are often more difficult to interpret. These are generally however indicative of increased use of locations during the summer months, and reduced usage during the winter.

4. Results: interview data

Overview of interview data

In total 5,236 interviews were conducted. 2,215 (42%) interviews were during the spring, 1,954 (37%) during the winter and 1,157 (22%) during the summer (when survey effort was half that of the other times of year).

The number of interviews per survey location (Figure 1) ranged from 19 (at Minstead Road) to 170 (at Marchwood Inclosure). Data are summarised by survey point in Appendix 3, which gives the totals of the number of interviews at each location (by season) and the typical group sizes at each location.

For those interviewees where the gender was recorded by the interviewee, slightly more (51%) of interviewees were female compared to male (49%).

Group size (i.e. the number in the interviewed party) ranged from 1 – 59. The most commonly recorded group size was one (i.e. people on their own) (2164 interviewees, 41%). Groups with more than twenty people were mostly organised events and included a fungal foray, a church group, a works barbeque, a family get together, a WI group and a Duke of Edinburgh Group (the latter being the group of 59). In total, 50 interviewees (1%) were noted as appearing to be part of an organised group.

The average group size for the interviewees was 2.1 people, with an average of 0.6 minors included. In total 906 (17%) of interviewed groups included minors.

3,234 interviewees (61%) were accompanied by at least one dog. Around two-thirds (2,164, 67%) of those interviewees with a dog just had one dog with them. The maximum number of dogs per interviewed group was 12. In total, 4,807 dogs were counted accompanying interviewees, giving an average of 0.9 dogs per interviewee. 2,874 (60%) of dogs were seen off lead by the surveyor during the interview.

The questionnaire typically took 8 minutes (the median value) to complete.

There were a total of 1,293 refusals, where a potential interviewee was approached and declined to be interviewed, for example because they were in a hurry. This was broadly equivalent to 1 refusal per 2 hour session (there were 1,200 sessions). The number of refusals per two-hour session ranged from 0-13, with just 3 sessions with 10 or more refusals. In all cases these involved a group/organised event (for example a running event taking place at Fritham). A total of 560 sessions (43%) had no refusals at all.

A total of 639 people were approached and had already been interviewed. These were not re-interviewed. The locations with the highest totals for people already interviewed were Racecourse View (37 people), Marchwood Inclosure (36 people) and by the phone box on Woodlands Road (34 people).

As far as possible, surveys were rescheduled to avoid particularly bad or inclement weather. Overall 948 (79%) of the two-hour sessions were without any rainfall and 184 sessions (15%) involved some rainfall but not continuous during the whole two-hour period. A total of 68 sessions (6%) involved continuous rain. For the winter period 338 sessions (70%) were without any rain, during the spring there were 412 (86%) without rain, and during the summer 240 (83%) were without any rain. No single survey location had all survey sessions in a single season with continuous rain.

Figure 1: Number of interviews per survey location

Type of visit (Q1)

The majority (83%) of interviewees were on a short visit directly from home that day. Those staying away from home on holiday accounted for 14% of interviewees and a further 2% were staying with friends or family. ‘Other’ visit types (37 interviewees in total, 1%) were varied and included those visiting as part of work or during work (10 interviewees) and 2 interviewees who lived in their campervans.

The relative proportions of different visit types were significantly different across the seasons (Χ²₆=91.69, p<0.001). In particular, during the summer there were more holiday makers (22%) and fewer day visitors (76%), compared to the spring (12% and 85% respectively) and the winter (11% and 86% respectively) (Figure 2).

Locations where tourists (those staying away from home in a second home, mobile home, camping or on holiday) accounted for a high proportion (above a fifth) of interviewees included Beachern Wood (20% of interviewees were tourists), Ocknell Pond (21%), Brock Hill (25%), Burley Cricket (31%), Minstead Road (32%), Blackwater (33%), Whitefield Moor (33%), Tilery Road (36%), Clayhill Heath (39%), Bolderwood (41%), Knightwood Oak (47%), Balmer Lawn (46%) and Bolderford Bridge (62%).

Figure 2: Visit type by season (from Q1).

Activity (Q2)

The survey involved asking interviewees for a single, main activity they were undertaking that day and then also for any other activities being undertaken by the interviewee or others in their group that day. A wide range of activities were recorded, however for most interviewees the main activity was either dog walking (55%) or walking (26%) (Table 3 and Figure 3). Activities were assigned to predetermined categories within the questionnaire or attributed as ‘other’ with additional details recorded. These ‘other’ activities were reviewed and as relevant additional categories were created, these are included in Table 3.

Table 3: Number (%) of interviewees by main activity and season

⁶ ‘Foraging’ includes those people on a fungi foray and also those collecting sweet chestnuts, wood or similar.

Figure 3: Breakdown of activity types by season

The relative proportion of those whose main activity was dog walking, walking, cycling (off-road and road) or all other activities combined was significantly different between the three seasons (Χ²₆=80.05, p<0.001). The differences were primarily in the summer, when dog walkers made up a lower proportion of interviewees while those cycling and undertaking all other activities were more frequent.

The variation in activity types across survey points are summarised in Map 3, with interviewees grouped by main activity (dog walking, walking, cycling (off-road and road) or all other activities combined). The map shows that dog walking was very much focussed around the peripheral areas of the SPA/SAC, while walking (without a dog) was the most common main activity at the more central survey locations. Cyclists were interviewed at scattered locations but notably those around Brockenhurst and also at Burbush Hill. Roundhill was the only location where an activity beside dog walking or walking was the most frequently given main activity, and here it was golf that was the main activity for many (47%) interviewees.

All activities recorded are summarised in Figure 4, which shows both the % of interviewees who gave a particular activity as either a single ‘main’ activity or as a secondary activity. The inclusion of the secondary activities little alters the relative frequency with which different activities were recorded. However, there were five activities that were more frequently given as secondary activities and therefore are perhaps more ancillary, these were: enjoying the view/picnic, bird/wildlife watching, photography, visiting café/pub and meeting up with friends.

Those who selected a secondary activity were most commonly those visiting café/pub (75% giving a secondary activity), followed by those meeting up with friends (53%). Golfers were the only interviewed group in which no interviewees gave a secondary activity.

Figure 4: Main and secondary activities (from Q2)

Frequency of visit (Q3)

Around a quarter (26%) of interviewees indicated that over the previous year they had tended to visit the New Forest woodland and heathland on a daily basis, with 1 to 3 times per week the next most common visit frequency (given by 18% of interviewees) (Table 4). Dog walkers were notable in being the most regular visitors, with 41% of dog walkers interviewed indicating they visited on a daily basis and a further 14% of dog walkers indicating they visited more than once per day. Over half (57%) of the commercial dog walkers visited more than once per day.

Aside from those walking dogs, for many activities 1 to 3 times per week was the most common visit frequency (18% of all interviewees). Frequency data are shown in Map 4. It can be seen that the interview locations around the periphery of the New Forest SPA/SAC tend to have the highest proportion of regular visitors.

We assigned an approximate number of visits to the categories in the questionnaire – so for example we estimated someone who visits more than once a day might tend to make around 400 visits over a year. Using these values we could estimate roughly how many annual visits were typically made by interviewees. Across all interviewees we estimate the typical number of visits per year to be 166. The number of visits per year for each activity are summarised in the final column in Table 4 and graphically in Figure 5. It can be seen that commercial dog walking, dog walking, horse riding, running and golf are the main activities that stand out as involving regular or frequent access.

Figure 5: Approximate number of annual visits per interviewee per year, by main activity type

Table 4: Numbers of interviewees (%) and frequency of visit (from Q3) by main activity type

Visit duration (Q4)

Visit duration is summarised by activity in Table 5. Visits typically lasted between 30 minutes and an hour (35% of interviewees) and between 1 and 2 hours (34%).

We assigned an approximate duration to the categories in the questionnaire – so for example we estimated someone who visited for 1 – 2 hours to have a visit of 90 minutes. Using these values, we estimated an approximate typical visit duration. Across all interviewees we estimated this to be 95 minutes. Visit duration values for each activity are summarised in the final column in Table 5 and graphically in Figure 6. It can be seen that those undertaking Duke of Edinburgh and golf clearly have the longest visits while dog walking and running are activities which involved interviewees having relatively short visits.

Figure 6: Approximate average visit duration, by main activity type

Table 5: Numbers (%) of interviewees and visit duration (from Q4) by main activity type

Time of year (Q5)

The majority (64%) of interviewees stated they visited equally all year round and did not tend to visit at a particular time of year (Table 6). Visiting equally all year round, was the most common response across most activity types, however where interviewees did indicate a season, summer was the most common (13% of interviewees tended to visit more in the summer). Photography and foraging were two activities where a notable proportion visited outside the summer, with 23% of interviewed photographers and 71% of foragers tending to visit more in the autumn.

Table 6: Numbers (%) of interviewees by main activity and time of year they tend to visit (from Q5)

Note that interviewees could give multiple responses and therefore the percentages do not necessarily add to 100.

Relatively few (19%) interviewees in the summer stated they tended to visit equally all year round compared to interviews at other times of year (39% visiting all year round in the winter, 42% in the spring).

Transport (Q6)

Virtually all interviewees (90%) had arrived by car/van or other motor vehicle⁷ (Table 7). Dog walkers, walkers and runners were the only activity types where any notable proportion of interviewees had arrived on foot. Some cyclists had travelled by bicycle to the New Forest that day, but otherwise most (70% of mountain bikers and 76% of road cyclists) had travelled by car/van and brought their bicycle with them. Around a fifth (27%) of horse riders had arrived at the interview location on horseback.

Table 7: Number (%) of interviewees by main activity and mode of transport (from Q6)

Note that interviewees could give multiple responses and therefore the percentages do not necessarily add to 100. ‘Other’ transport types were horse and mobility scooter.

⁷ All motor vehicles were included in this category – including horse boxes, motor homes, campervans, motorbikes etc. The only exception was mobility scooters which are included under ‘other’

Reasons for site choice (Q7)

Interviewees gave a wide range of reasons as to why they had chosen to visit the specific location where interviewed, as opposed to other parts of the New Forest. Interviewees typically cited a range of reasons, for example sites could be favourite locations within easy access from home and with good parking. Responses were coded as part of the survey and free text responses also recorded. Free text responses were reviewed and categorised and key reasons are summarised in Figure 7. It can be seen that close to home (or work or holiday accommodation) was the most common response, given by a quarter of interviewees (25%). The second most frequent response (16% of interviewees) related to previous knowledge or familiarity (for example it was the interviewee’s ‘favourite’ walk, ‘preferred’ route or there was a sentimental connection with the location). Travel routing was also important, with many (10%) citing that they had stopped because the location was en route (to the shops, to friends, to the garage etc.) or even just stopping at random (2% interviewees). A few interviewees (1%) had been deflected from other locations because they were full or because the car park was shut.

Figure 7: Reasons for site choice (from Q7) by activity. Percentage figures on the graph indicate overall total (all interviewees).

Visitor Routes (Q8-9)

Most (57%) of interviewees indicated their route on the day interviewed reflected their usual route. In addition, 10% of interviewees didn’t have a typical visit and a further 10% were on their first visit. Out of all interviewees, 19% indicated their route was shorter than normal and 2% indicated it was longer than normal.

Route data were plotted for 5,136 interviews⁸, and in total involved 210,000 nodes (i.e. ‘clicks of the mouse’) to plot. Data are shown for all interviews in Map 5. In Map 6 route data are shown as a heat map on a grid (100m cells) with different maps for dog walking, walking and all cycling (the three main activity groups). The grid is clipped to the SPA/SAC and therefore shows the relative intensity of use (by the interviewed visitors) within the European site. It is important to note when viewing the maps that the interview data relate to specific access points and not all access points were surveyed, as such the data as plotted do not show use across the New Forest woods and heaths as a whole – just the use from the surveyed locations. Nonetheless it is clear that the visitors at the surveyed locations cover a wide area and that the main activities plotted all overlap in space. The map for cyclists tends to show the most concentration along main tracks and routes, and for example around 63% of the total route lengths mapped for those undertaking mountain biking/off-road cycling was on the promoted cycle routes.

It can be seen that there are some differences in Map 6 between the different activities. Dog walking tends to be concentrated in space more than the other activities with relatively high concentrations of use at many car parks but with use discrete to particular survey points. For walkers there are a few key routes that are clearly well trodden, in particular the Tall Trees route between Brock Hill and Blackwater and the east-west route from Abbott’s Well to Fritham. The map for cyclists shows use focussed more along particular routes (e.g. the cycle track from Beachern Wood to Bank, and the old railway line between Burbush and Wilverley) than the other activities. In the ‘Other activities’ map the high concentration near Brook relates to golfers at Roundhill.

There were significant differences in total route lengths between the seasons (Kruskall-Wallis H = 16.92, 2 d.f., p<0.001) with route lengths tending to be shorter in the summer. There were also marked differences between activities (main activity either dog walking, cycling (all cycling), walking or other, Kruskall-Wallis H = 713.28, 3 d.f., p<0.001), with cyclists tending to go much further than the other groups. Looking across season and activity, it can be seen that dog walkers (and to some extent cyclists), interviewed over the summer, tended to have shorter routes (Figure 8).

⁸ Those interviews without a route map were predominantly either interviewees in too much of a hurry or were people who did not know where they were going or where they had been.

Figure 8: Route lengths by activity and season. Simplified activity types based on main activity and all cyclists grouped together.

Route data are summarised in Table 8, which gives summary statistics for overall route lengths (as mapped), route lengths solely within the New Forest SPA/SAC and also for penetration distances. Data are summarised by season and by a selection of different main activity types. Taking all interviewees, the typical overall distance covered on their ride/walk was just under 3km (2969m), virtually all of which was within the New Forest SPA/SAC. This kind of distance tended to extend a little under 1km (943m) from the start point (i.e. the penetration distance, how far the route goes from the start point before turning back).

Table 8: Summary of route length metrics by season and selected main activities

Factors that influenced choice of route (Q9)

Interviewees gave a variety of different reasons as to what had influenced their route that day. Responses were coded as part of the survey and free text responses/further details were also recorded. Free text responses were reviewed and categorised and key factors are summarised in Figure 7. Previous knowledge of the location was the most common response (22% of interviewees), for example people following a favourite route, habitual route etc. Other factors included time available (13%), weather conditions (such as shade or shelter etc.,12%), following a marked trail or the paths available (12%) and activity specific factors (such as where the buggy could go, golf course etc., also 12%). For 3% of interviewees the route was random while a further 3% were influenced by the dog, for example following the dog.

Differences between seasons were slight. However, avoiding muddy paths/tracks was a factor for just 2% of those interviewed in the summer (compared to 7% in the spring and 8% in the winter) while the a viewpoint or feature (such as the Knightwood Oak) was more frequently cited in the summer (9% of interviewees) compared to the winter and spring (5% of interviewees for both).

Figure 9: Factors that influenced choice of route (from Q9) by activity. Percentage figures on the graph indicate overall total (all interviewees).

Information used to plan visit (Q10–Q15)

Maps were the most commonly cited type of information used to plan interviewee’s visits (Table 9), with 15% of interviewees using maps. In addition, many interviewees cited map-based websites and apps as important sources. Looking across activities, few dog walkers used information sources to plan their visit compared to other visitors and none of the commercial dog walkers interviewed had used any information sources in their visit planning. Of the websites specifically mentioned, we checked Strava as this can generate similar route data to the data shown in Map 6. Among those off road cycling/ mountain biking just 3% (9 interviewees) stated they used the Strava app and none of the 17 road cyclists that were interviewed had used it. Strava was slightly higher among runners, 6% (6 interviewees).

Table 9: Number (%) of interviewees and information sources used to plan visit (from Q10)

Awareness of species vulnerable to recreation impacts (Q16)

Overall around a fifth of interviewees (22%) indicated that they were not aware of a wildlife habitat or species that could be affected by recreation while around two-thirds (67%) of interviewees were both aware and could name a species or habitat (Figure 10).

There was no significant difference in the proportion of interviewees who were aware of habitats/species affected by recreation for dog walkers when comparing those with (18% aware) or without their dogs on-lead (21% aware) during the interview (χ²₂=2.3, p=0.13).

Species or habitats named were coded by the surveyor according to pre-determined categories, for example recording whether breeding birds were mentioned or whether a specific species or species group was mentioned. These were not mutually exclusive, with the potential to log multiple responses against individual interviewees. Responses are summarised Figure 11. Breeding birds were the most commonly named concern (40% of interviewees) and this category included responses such as ‘ground-nesting birds’ or nesting birds in general. Around 2% of all interviewees specifically named heathland breeding birds – Nightjar, Woodlark or Dartford Warbler and around 4% indicated breeding waders (either naming a particular wader species such as Curlew or Lapwing or simply indicating breeding waders in general). Around a fifth (24%) of interviewees cited deer as vulnerable to recreation impacts and 30% highlighted livestock (either livestock in general or specific mention of cows, pigs, ponies etc.). Other responses (13% of all interviewees), which didn’t fit in the pre-determined categories included reptiles (i.e. ‘snakes’, ‘lizards’, ‘sand lizard’ etc; 4% of all interviewees) and mushrooms (1%).

Visits to other sites (Q17–Q18)

Most (4,434 interviewees, 83%) stated that they were on a short visit or day trip, travelling directly from home that day. These visitors are essentially those who are not on holiday and therefore more likely to live relatively close to the New Forest. This subset of interviewees were asked to indicate an approximate percentage of their weekly visits (for the given activity) they undertook within the New Forest woodland and heathland. Most (41%) of these interviewees indicated that all their visits for their chosen activity took place within the New Forest heathland and woodland, while a further 27% indicated at least 75% of their visits were to the New Forest woodland and heathland. In total therefore, around two-thirds (68%) of day visitors to the New Forest choose the New Forest above other locations virtually all of the time.

Data are summarised by activity in Figure 12. It can be seen that horse riders use the New Forest almost exclusively, with 73% indicating they visited no-where else, while family outing was the activity least associated solely with the New Forest, with only 10% of interviewees indicating they only visited the New Forest for that activity.

There were 2,604 interviewees who indicated they visited other sites in Q17 and these were subsequently asked to name another location they visit for the given activity. Responses were recorded as free text and subsequently checked and standardised to correct spelling and ensure consistency.

Of these 2,604 interviewees:

• 31% of these named a specific coastal site or indicated they visited the coast in general (e.g. ‘local beaches’);
• 7% named a country park;
• 4% named another National Park in the UK;
• 3% named another part of the New Forest SPA/SAC (for example ‘Hatchet Pond);
• 1% referred to a long distance path/route;
• 19% indicated a broad location rather than a specific site (e.g. ‘local beach’, ‘Dorset’, ‘Purbeck’).

In total 531 different sites were named by interviewees, reflecting a wide range of locations. Sites (named by at least 5 interviewees) are listed in Appendix 4. The top ten named locations were:

• Hengistbury Head (96 interviewees, 4%)
• Lepe Country Park (86, 3%)
• Purbeck (61, 2%)
• Lymington Marshes (54, 2%)
• Highcliffe beach (52, 2%, note that a further 38 interviewees also stated ‘Highcliffe’)
• Southampton Common (45, 2%)
• South Downs (44, 2%)
• Bournemouth beach (40, 2%)
• Barton-on-sea (39, 1%, note that a further 23 also stated Barton-on-sea ‘beach’)
• Mudeford (38, 1%)
• Highcliffe (38 ,1%)

Other locations (besides the top ten in bullets above) are shown in Figure 13

Off-site management interventions (Q19–Q21)

In order to understand what interventions could be done away from the New Forest woodland and heathland to draw recreational use, interviewees were asked if they would use footpaths and rights of way (away from the New Forest woodland and heathland) if they were improved. In a similar vein, interviewees were also asked whether they might use a new Country Park (outside the woodland and heathland of the New Forest), if such a park were created.

Responses are summarised in Figure 14, which is split between visit type. The responses would suggest that both approaches (Country Park and improved footpath network) would be used by interviewees to a similar extent and there is relatively little difference in the numbers of interviewees selecting each.

Using the data for all interviewees we checked the number of interviewees who showed a distinct preference, i.e. answered that they would use one type of measure (improved path network or Country Park) and not the other. This showed relatively even proportions showing a preference for a single measure, with 14% of interviewees saying they would use a new Country Park and not an improved path network while 16% of interviewees said they would use an improved path network and not a new Country Park.

Comparing between activities (Table 10) there were some slight but significant differences, with those undertaking other activities showing a preference for the Country Park and over a fifth (21%) of cyclists preferring the improved path network (Χ²₆=28.99, p<0.001). Comparing between visit frequency suggested frequent visitors showed greater preference for improved footpaths over a country park, but the opposite pattern for the infrequent visitors (see Table 11).

Table 10: Number (%) of interviewees answering ‘yes’ that they would use either a new Country Park (Q20) or an improved path network (Q19), but not the other, categorised by simplified main activity

Table 11: Number (%) of interviewees answering ‘yes’ that they would use either a new Country Park (Q20) or an improved path network (Q19), but not the other, categorised by visit frequency

Interviewees suggested a range of facilities for any potential new Country Park (Figure 15 and Figure 16). Responses were grouped based on the predetermined categories within the questionnaire and further categories were created where there were frequently cited additional suggestions. Some kind of café (or restaurant, pub or bar) was the most popular answer (18% of interviewees that made a suggestion). Extensive or good walking routes were also frequently cited (17%), as was the importance of the site having a wild feel, natural or with wildlife (16%). Interviewees clearly had very divergent opinions, however. Some were clear that they thought no Country Park was necessary as the New Forest was ‘perfect’ or ‘already a haven for dog walking’ and there was no need for more commercialisation or facilities. Some felt they would not use a Country Park themselves but wanted it to draw other users, so the New Forest areas remained less busy. Those that undertook specific activities often wanted very specific infrastructure, for example one of the people flying model aircraft suggested a runway while golfers suggested a new golf course and a good clubhouse. Some simply suggested that a new Park should be similar to an existing one – with Moors Valley the most common suggestion (2%).

Figure 16 is drawn from the free text responses/further details. The figure is based on those responses which were different (or provided further information on) the categories used in Figure 15. For example, a response such as ‘seating areas’ was omitted from the word cloud as it is already listed in Figure 15, however a response such as ‘benches along paths to provide places for elderly to sit’ was included. The word cloud highlights how important natural, quiet and wildlife-rich spaces were for many interviewees. The prominence of ‘bins’ reflects comments relating to both rubbish bins and dog bins.

Figure 15: Features interviewees suggested they would like to see at a new Country Park (from Q21).

Figure 16: Word cloud summarising further details relating to features interviewees wished to see at a new Country Park (from Q21).

Visitor origins (home postcodes, Q23)

Overview

In total, 4,871 interviewees (91%) gave a full, valid UK postcode that could be geocoded using the national database. Postcode data are summarised in the following maps:

• Map 7 shows all interviewee postcodes, which range from the Isle of Wight in the south to Aberdeenshire in northern Scotland;
• Map 8 repeats the information in Map 7, but just shows more local postcodes and also shows 5km bands out to 25km around the New Forest SPA/SAC;
• Map 9 shows the same geographic area as Map 8, however only those who were on a short visit/day trip from home are shown and the shading reflects the different activity types;
• Map 10 shows the same data as Map 9, however the shading reflects frequency of visit, with frequent visitors shown in the red colours with the darker red reflecting the most frequent.

These data reflect a general pattern whereby visitors on a short visit/day trip from home live relatively close whereby those on holiday come from a wide geographic spread (and potentially further afield than the UK as we have not tried to map those holiday makers who were visiting from abroad).

By built-up area

Table 12 lists the built-up areas with the most interviewees, showing all those with at least 20 interviewees. In total around two-thirds (67%) of interviewees came from these settlements. The Bournemouth/Poole conurbation⁹ was the single built-up area from which the most interviewees originated (12% of all interviewees), with the South Hampshire¹⁰ built-up area second (9%).

⁹ Built-up areas are defined by the Office of National Statistics, representing discrete settlements; Bournemouth/Poole does also include Christchurch and extends also to Barton on Sea and New Milton
¹⁰ The built-up area for; South Hampshire extends from Southampton to Havant.

Table 12: Built-up areas (as defined by Office of National Statistics) and the number of interviewees from each, by visit type

Built-up areas ranked by total number of interviewees, all those with at least 20 interviewees shown

By local authority

Visitor origins by local authority are shown in Figure 17 and Table 13. These summaries give the totals for the New Forest National Park and then for the other local authorities within 25km, excluding those areas within the National Park. We have only listed those authorities which fall at least partly within 25km of the SPA/SAC and we have given totals for the original Dorset local authorities (within the 25km) which were merged while the survey was taking place.

In total, 20% of interviewees on a short visit or day trip from home that day gave postcodes within the National Park boundary. A further 40% came from outside the National Park but within the New Forest District. Other local authorities accounted for relatively small proportions of the interviewees by comparison.

Figure 17: Proportion of interviewee geocoded postcodes (those travelling from home on short visit/day trip only) by local authority. Data matches Table 13. (numbered labels are relate to: 1=Purbeck, 2= N. Dorset, 3=W. Dorset, 4=Weymouth, 5=Fareham, 6=Gosport, 7=Portsmouth, and IOW below Portsmouth)

Table 13: Numbers of interviewee geocoded postcodes (%) by local authority

Only local authorities that fall partly or wholly within 25km of the New Forest SPA/SAC are included. *Figures for Dorset were extracted using the original boundaries and therefore the total given relates to the five old authorities listed. Data matches Figure 17.

Visitor origins at different survey points are shown in Map 11. Here the pie-charts show the numbers of interviewees from different broad zones, defined primarily using local authority boundaries. The inset shows how the zones have been defined. The map clearly shows that people visiting from Wiltshire tend to visit the northern parts of the New Forest SPA/SAC (north of the A31). People from Bournemouth, Christchurch and Poole tend to visit the south-western part while those living between the SPA/SAC and Southampton Water tend to visit the eastern side of the SPA/SAC. The survey points towards the middle of the SPA/SAC tended to have higher proportions of very local people (e.g. from Brockenhurst, Lyndhurst etc.) and those from much further afield.

Linear distance (‘as the crow flies’)

The median distance for all interviewees from their home postcode to the interview location was 7.75km and 75% originated from within 21.4km (Table 14). For those on a short visit/day trip from home, the median was 6.1km with 75% living within 13.8km of the interview location. For those visiting directly from home on a short visit, the median distance in the summer (5.83km) is the same as the winter distance and very close to the spring, suggesting little annual change. For all visitors however, the summer median distance is higher (11.28km), indicating the higher proportion of holiday-makers at this time of year.

Table 14: Summary statistics for the distance (km) between home postcode and survey location for different groups of interviewees