02brainard (ds)

20 downloads 0 Views 1MB Size Report
purpose UK forestry policies using just six regions for the whole of Great .... Blidworth Tower‡. 37596. Estimate .... ation blocks, and represented each by a single.
02brainard (ds)

5/11/01

2:14 pm

Page 423

Modelling demand for recreation in English woodlands JULII BRAINARD, IAN BATEMAN AND ANDREW LOVETT Centre for Social and Economic Research on the Global Environment, School of Environmental Sciences, University of East Anglia, Norwich NR4 7TJ, England

Summary Previous models to describe the desire for recreation at English forest sites have tended to use fairly crude and regional measures. This study demonstrates how forest recreation demand can be modelled quite locally and using just site-specific characteristics or simple measures of available population as input. A field survey of 33 Forestry Commission sites was made in order to collect data on attractive features at each site. These data were supplemented with variables to indicate the availability of competing woodlands and population totals within set travel distances. The outputs were simple but robust stand-alone functions to describe visits across many sites.

Introduction The impetus to create amenity woodlands in the UK has never been stronger. This follows years of increasing recognition of the many unpriced amenities that forests can provide, in addition to their market timber value (Matthews, 1994; Selman, 1997). The importance of recreation has been the subject of particular scrutiny, as demonstrated in an extensive body of relevant research by Willis and Benson (1989), Willis (1991), Willis and Garrod (1991, 1992), Garrod and Willis (1993) and Bateman et al. (1996). It is even arguable, at least in lowland Britain, that consideration of the possible ecological, aesthetic and recreational benefits of forests now tends to outweigh their commercial uses (see discussions by Pitt, 1992; Innes, 1993; Jones, 1994; Matthews, 1994; Selman, 1997). A change in bias towards multi-purpose forestry is also demonstrated by the establishment of a new ‘National Forest’ in the English Midlands (Countryside Commission, © Institute of Chartered Foresters, 2001

1993), initiatives to foster a series of Community Forests throughout England and Wales, and in discussion documents issued jointly by the Forestry Commission and Countryside Commission (FC/CC, 1996, 1997). And, although support for multi-purpose forestry is not universal (see Bobiec et al., 2000; Calder et al., 2000), it seems likely that the push to plant amenity woodlands will continue to be strong in the near future. However, it is noteworthy that decisions on where to encourage amenity woodland creation have largely been made at a regional scale, using somewhat crude assessments. For instance, Selman (1997) assessed the efficacy of multipurpose UK forestry policies using just six regions for the whole of Great Britain. Similarly, the FC/CC 1996 discussion document mapped several factors relevant to forest creation – including the ratio of population to current woodland cover, incidence of derelict land, agricultural quality and percentage tree cover – all on a county Forestry, Vol. 74, No. 5, 2001

02brainard (ds)

5/11/01

424

2:14 pm

Page 424

F O R E S T RY

basis. These four variables were overlaid and a summary, five-category map produced to indicate the opportunity for new woodland sites. That this map was only preliminary the report admitted at the time, and moreover in their follow-up paper, the FC/CC reported a general response that this approach ‘needs to be adapted and refined at the regional or sub-regional level. This would involve looking at a greater level of detail and incorporating a considerably wider range of information, potentially using GIS [geographic information systems]’ (FC/CC, 1997, p. 22) A GIS is a specialized set of computer programs for the storage and processing of spatially referenced data. We used the ARC/INFO package in this research (ESRI, 1996), which made manipulation of the many digital data sources both tenable and practical. Using a GIS made it realistic to identify amenity woodland needs much more locally. In particular, this study demonstrates how recreation demand at a specific location can be modelled using just site characteristics as input, or combining these with simple indicators of local population access. Note that we use the word ‘demand’ in the sense of want or desire, rather than as an aspect of economics, forceful request or other meaning. The simplicity of the models that will be presented here is one of their main merits. Previous and similar work to describe recreation demand in Britain (see Willis, 1991; Willis and Garrod, 1991, 1992; Bateman et al., 1996) tended to concentrate on measuring the distance decay relationship between the travel costs for potential visitors and sites of interest. Considerable scepticism has been expressed about the validity of these economics-based assessments (Grove-White, 1999). Moreover, all of these studies depended upon relatively expensive and time-consuming collection and processing of visitor survey information at each site of interest. The research presented here, however, shows how less complex and yet still fairly robust modelling might be undertaken to calculate and perhaps even predict site visits at unsurveyed woodlands. Several datasets of relevance were collected and are described at length. A field survey of 33 woodlands for which vehicle counts were available was undertaken in order to collect data on amenity features at these locations. These details were supplemented by calculation of two indices to

represent the availability of competing woodland, along with some basic measures of local population access. Initially, a number of models to predict visit numbers at multiple locations were generated from the site-specific and competing woodland variables only. These functions have the advantage of being easy to implement, and are relatively successful. When measures of market population around each woodland were subsequently introduced, the resulting models yielded still more effective descriptions of recreation usage.

Data Several types of information were needed in this analysis. Actual party arrivals statistics at existing woodlands were used as a starting point. Detailed inventories of amenity features at each location were compiled, and the extent of competing forest area along with some basic measures of possible market population estimated. Vehicle counts at 33 English woodlands Traffic count data at 33 English Forestry Commission (FC) sites form the basis for this work, and were taken from two sources. Guest and Simpson (1994) presented results for a vehiclemonitoring scheme at 28 locations in northern and eastern England during the period from 21 June to just prior to Christmas, 1993. Unfortunately, the authors described numerous problems with the counter devices that could be expected to introduce errors in the data, including a high incidence of batteries failing, and multiple recordings for a single vehicle. Nevertheless, by examining seasonal trends and weekday/weekend use patterns, Guest and Simpson felt able to adjust their raw data to produce annual estimates of vehicle arrivals for each site. These researchers also expected problems with the traffic counters to be resolved over time, making counts from later dates more reliable – a perception echoed informally by other FC staff. A second source of vehicle numbers at 15 English woodlands, in 1996, was provided by the FC office in Edinburgh. Ten sites occurred in both the 1993 and 1996 datasets (listed together in Table 1). In such cases, because of the general

02brainard (ds)

5/11/01

2:14 pm

Page 425

M O D E L L I N G D E M A N D F O R R E C R E AT I O N

Figure 1. Study sites.

425

02brainard (ds)

5/11/01

2:14 pm

Page 426

426

F O R E S T RY

Table 1: Estimated annual vehicle counts at Forestry Commission sites in 1996 and 1993, with comments from Guest and Simpson (1994) Forest District

Name

East Anglia

Dunwich High Lodge Lynford Arboretum Lynford Stag Two Mile Bottom Bakethin Bull Crag Castle Forest Drive Lewisburn Warksburn Bogle Crag Grizedale Main/Hall* Noble Knott Whinlatter Centre

Kielder

Lakes

New Forest

Blackwater Bolderwood Dorset Moors Valley Northants Bucknell Salcey Wakerley N. York Moors Dalby† Rothbury Chopwell Hamsterley

Simonside Sherwood/Lincs Blidworth Bottom‡ Blidworth Lane‡ Blidworth Tower‡ Chambers Farm Goyt/The Street Normans Hill§ Thieves Wood§ Visitor Centre

1996

7101

9533 24243 31641 3794

7543 55797

1993

1993 Comments

18980 14940 6346 42010 22936 5379 – 21896 – 8746 4900 14924 85181 12054 69283

Estimate from weeks working Estimate from weeks working Estimate by adding week data Estimate by adding week data Estimate by adding week data Incorrect week data altered – Estimate by adding week data – Estimate by adding week data Estimate by adding week data Estimate from weeks working Estimate by adding in month data Estimate from weeks working Estimate by adding in month data and from weeks working

39338 58628 165552

130151 42298 76796

12430

21360 77650 51490 143626 39316 17724

12688 54547 52754 37596 23605 84279 30936 72276 38919

Estimate from weeks working Estimate from weeks working Estimate from weeks working Estimate by adding month data Estimate from weeks working Estimate from weeks working; only ran for the last months of year and likely an underestimate Estimate from weeks working Estimate from weeks working Estimate from weeks working Estimate from weeks working Estimate from weeks working Estimate from weeks working; likely an overestimate due to seasonal reading Estimate from weeks working Estimate from weeks working Estimate from weeks working

* In the original Guest and Simpson report Grizedale Main and Grizedale Hall are listed separately. However, entrances to the two car parks are