The extent and intensity of management burning ... - Wiley Online Library

Journal of Applied Ecology 2006 43, 1138–1148

The extent and intensity of management burning in the English uplands Blackwell Publishing Ltd

A. R. YALLOP,* J. I. THACKER,* G. THOMAS,† M. STEPHENS,* B. CLUTTERBUCK,* T. BREWER and C. A. D. SANNIER* *Institute of Water and Environment, Cranfield University, Silsoe, MK45 4DT, UK; and †Queen Mary College, University of London, London, E1 4NS, UK

Summary 1. World-wide, the controlled use of fire is an important ecological management tool and is essential for the continuance of many communities. It is used extensively in upland regions of England to maintain dwarf shrub habitats for game-bird rearing. Inappropriate burning, however, is now cited as the second most important reason for the poor condition of conservation sites in these areas. Despite this there are few data on the extent and frequency of its use to help judge its potential impact on biodiversity. 2. This study, using aerial photography of a 2% sample (208 km2) of the English uplands, surveyed the national scale of fire management for the first time, and used historical photography to identify medium-term trends in its use. 3. Management burning in the English uplands is now widespread on ericaceousdominated moorland; in the year 2000 17% of the area of this habitat had been burned within the previous 4 years, equivalent to 114 km2 year−1. The present median burn repeat time of consistently managed sites is approximately 20 years. 4. Within most of the English national parks there has been a significant increase in the extent of new burns (from 15·1% to 29·7%) over this period, indicating an intensification of burning regimes in some areas. 5. Synthesis and applications. The extent and frequency of burning, and the habitats in which this management occurs, are contentious issues. Reconciling the differing objectives of conservation, game rearing and agricultural stakeholders to allow the development of both strategic and local management planning to address these issues requires information on the extent and history of burning practices. This study provides a much needed first national estimate of burning practices in England and serves as a baseline against which changes in management regimes and their impacts on habitats can be judged. Key-words: bog, Calluna, fire, grouse management, heather burning, management, moorland, uplands Journal of Applied Ecology (2006) 43, 1138–1148 doi: 10.1111/j.1365-2664.2006.01222.x

Introduction Controlled burning in the English uplands has a history of hundreds, perhaps thousands, of years (Pearsall 1950; Birks 1988; Simmons 2003) and, in conjunction with livestock grazing, is probably the most extensive and ecologically important management activity undertaken in these areas of the British Isles. As a natural absence © 2006 The Authors. Journal compilation © 2006 British Ecological Society

Correspondence. A. R. Yallop, Institute of Water and Environment, Cranfield University, Silsoe, MK45 4DT, UK (e-mail [email protected]).

of tree cover in England is rare (Dimbleby 1952; Simmons 2003), without this management, following Mesolithic/ Neolithic clearances of upland woodland (Brown 1997), the extent of the Ericaceae-, Cyperaceae- and Poaceaedominated moorlands would be far smaller than is seen today. Within the contemporary upland landscape there are two major drivers for the continuation of management by burning: sheep Ovis aries L. grazing and red grouse Lagopus lagopus (L.) game shooting. Burning of grass swards, often dominated by Molinia caerulea (L.) Moench, is primarily undertaken to improve grazing

1139 Burning in the English uplands

© 2006 The Authors. Journal compilation © 2006 British Ecological Society, Journal of Applied Ecology, 43, 1138–1148

access to new growth early in the season by removal of accumulated necromass. On ericaceous-dominated moorland, fire is used to reverse the phasal growth process of heather Calluna vulgaris (L.) Hull to encourage the development of new succulent shoots, as well as reducing overall vegetation height. Where such burning is undertaken to improve sheep grazing it is usually executed in the form of infrequent but large burns, whereas management for grouse rearing tends to be more intensive, with numerous small-scale fires producing a mosaic of different aged stands of heather. Such management also favours animal species dependent on open habitats, so moorlands can support a wide diversity of terrestrial (Usher & Thompson 1993) and aquatic (Petersen et al. 2004) invertebrates. Many bird species also benefit, including curlew Numenius arquata (L.), golden plover Pluvialis apricaria (L.) and lapwing Vanellus vanellus (L.) (Whittingham, Percival, & Brown 2000; Tharme et al. 2001). However, although heather moorland, and to an extent other upland communities, is only widespread because of this management, it cannot be considered axiomatic that all burning regimes are beneficial, in either a species or landscape conservation sense. It is an ecological paradigm that management considered essential for many species and communities, and responsible for the landscape we see today, is, or has been, detrimental to many others. Adverse effects on meadow pipit Anthus pratensis (L.) (Smith et al. 2001) and other bird species (Tharme et al. 2001) arise from burning, and Tucker (2003) highlights concerns about negative impacts on vegetation, invertebrates, soil structure and hydrology, water quality and carbon storage. Today 32% of the area of upland sites of special scientific interest (SSSI) in England is considered to be in poor condition because of inappropriate burning (English Nature 2006), although this figure may be inflated by defining the entirety of a SSSI to be in poor condition when only part may be inappropriately burnt. The effects of burning regimes on the ecological dynamics of individual moorlands are clearly complex and inextricably linked with those of grazing (Miles 1988; Thompson et al. 1995). However, a general pattern is apparent; burning for grouse management since the middle of the 19th century, combined with grazing, drainage and air pollution, has resulted in the replacement of extensive and diverse peat-forming bog communities with relatively species-poor cotton grass Eriophorum spp.-dominated degraded bog and Calluna moorland landscapes (Ratcliffe & Thompson 1988). Anecdotal records indicate that in many areas a history of drainage and low grazing intensity, combined with the deliberate use of intense fires during drier conditions, often called ‘hot-burning’, have helped to produce near monoculture stands of Calluna. Elsewhere, the intolerance of heather to heavy grazing (Palmer & Hester 2000) and too-frequent burning can lead to its replacement by, for instance, M. caerulea swards (Stevenson

et al. 1996), an end-point that may require considerable intervention to correct (Ross, Adamson & Moon 2003; Marrs et al. 2004). Although available data on the effect of burning on bogs are often sparse and contradictory (Stewart, Coles & Pullin 2004 and references therein), where burning occurs on the remaining areas of blanket bog it does raise particular concerns for biodiversity (Shaw et al. 1996; Tucker 2003; Stewart, Coles & Pullin 2004). There is also historical evidence from peat cores that increased frequency of wildfires reduces or halts peat production in mires (Kuhry 1994; Pitkänen, Turunen & Tolonen 1999) and rotational burning has been shown to reduce carbon sequestration in peat bog (Garnett, Ineson & Stevenson 2000). These effects are presumably mediated through adverse effects on peat-forming species, although the consistent presence of evidence of bog mosses Sphagnum spp. above charcoal layers in peat cores shows that re-establishment occurs if conditions are appropriate (Kuhry 1994; Pitkänen, Turunen & Tolonen 1999). Much of the upland landscape therefore appears dependent upon intermediate levels of burn management, with complete removal of burning (and grazing) presumably resulting in the development of scrub and woodland communities, but with excessive burning frequency clearly linked to such loss of biodiversity interest that SSSI status is compromised. However, despite this inextricable link between human activity, both beneficial and adverse, with the upland landscape, there is at present little information about either the extent or frequency of burning nationally to help judge its impacts. Perhaps more importantly, we have little or no evidence regarding changes in practice that may be occurring. This study surveyed, for the first time, the scale of burning in the English uplands, to provide a baseline against which future practices can be assessed, and examined the evidence for changes that have occurred over the past 40–50 years.

Materials and methods      2000 Blanket bog, dry and wet heath, bracken, acid grass and montane habitats were identified as target classes likely to contain examples of management by controlled burning. To identify the distribution of these in the English uplands, an initial sampling frame was derived from the Countryside Survey 2000 Environmental Zone 3-Uplands (Haines-Young et al. 2000). This was stratified using Land Cover Map 2000 1-km2 habitat distribution data for these classes (Fuller et al. 2002), which defined a total study area of 10 360 km2 on which the national estimates presented here were based. Within this area a 2% sample of 1-km2 squares was selected using a random systematic design that provided even coverage and, with the use of two samples per block, gave an unbiased estimate of error (Webster & Oliver 2000). A sample frequency of 2% has been shown to

1140 A. R. Yallop et al.

produce good results for estimating areas of landscape features (Hanuschak et al. 1979) and preliminary work had shown that 1-km2 samples provided a good balance between the amount of information each site contained and the number of sites that could be surveyed. Sampling was conducted by dividing the study area into 10 × 10-km blocks (i.e. 100 1-km2 sample squares). Within the first block two sample squares were chosen randomly. The systematic sample was then created across the study area by using the two sample points at the equivalent locations in all blocks. This resulted in the selection of 208 1-km2 sample sites. At the time this work was undertaken the most recent source of photography completely covering the English uplands was orthocorrected and geo-referenced digital aerial photography from GetMapping Plc. (Hampshire, UK) acquired around the year 2000, and this was therefore used for this project.

      Because of restrictions in either availability or quality of historical images from the original 208 sample squares, a complete national assessment of change was not possible. However, an archive of 1970s photography, held by Cranfield University, Silsoe, UK, and covering the majority of the national parks in England (excluding Dartmoor and Cumbria), was used, with the addition of some imagery obtained from the National Monument Records Centre (Swindon, UK), to provide information about possible changes in management practice. Those of the original 208 1-km2 sample sites described above that occurred within park boundaries were selected and further refined by only using those with more than 30% presence of the target habitat classes. Three of these were rejected because of excessive cloud cover in the 1970s images, resulting in 23 sites for analysis. Overall changes in the area of recently burnt areas in this data set were measured using paired t-tests. Before analysis, proportion data were arcsin square-root transformed and tested for normality. The direction and significance of changes in repeat times for these sites were quantified using a Wilcoxon signed-rank test. Two squares were deleted from this latter analysis because there was no evidence of burning in them in the 1970s imagery; no repeat times could be calculated for them as the calculation involves dividing by the proportion of a sample square that has been burnt, which cannot be done if that proportion is zero. This left 21 paired samples. Because of the spatial bias in sampling, these data were not directly interpretable in terms of national statistics. © 2006 The Authors. Journal compilation © 2006 British Ecological Society, Journal of Applied Ecology, 43, 1138–1148

    The upland landscape is usually categorized into a mosaic of essentially intergrading habitat types. The criteria and definitions for some of these can be difficult

to judge in the field and from aerial imagery. For reliability, therefore, areas of moorland examined were divided into two readily identifiable generic upland moorland classes: an Ericaceae-dominated dwarf shrub heath (DSH) and Poaceae–Cyperaceae-dominated moor. Note that, as used here, DSH is defined as a continuum from almost pure stands of C. vulgaris to more diverse mixtures with Erica spp. and Eriophorum spp. present. It covered all upland moorland where ericaceous shrub was visible in the imagery and encompassed definitions such as wet and dry heath and some areas of bog and degraded bog. The Poaceae– Cyperaceae-dominated moorland class consisted of Molinia swards and Eriophorum- and Deschampsiadominated communities, and encompassed all areas of upland heath, moor and bog with few shrub species visible. This category did not include enclosed and improved areas of grazing. Preliminary work showed that, despite considerable variation in colour and texture arising from the age of Calluna/DSH burned and/or the presence of faster recovering vegetation such as Eriophorum and Vaccinium species, four classes of regrowth related to burning could generally be identified in aerial imagery. These were defined as: Class 1, new burn ‘scar’ with no visible regrowth of dwarf shrub/Calluna; Class 2, recent burn with only partial canopy of regenerating dwarf shrub/ Calluna; Class 3, a visually smooth, dark and relatively complete dwarf shrub/Calluna canopy within a burn scar; Class 4, no visible remains of burn scar, visibly pale with ‘lumpy’ texture of degenerating heather canopy. While Classes 1–3 represented visible evidence of burn management, Class 4 represented areas that had not been burnt for a considerable period (Fig. 1). The accuracy of the classification process was checked by two independent interpreters, each of whom classified 1000 polygons with an overall areal concordance of 94%. Differences in interpretation of recently burnt polygons (i.e. Class 1 and Class 2) and other classes were small (0·6% of the average area of Classes 1 and 2) and recognition of recently burnt areas was generally consistent. While assignment of burn signatures into these categories was consistent, the actual age or time since burn that each class represented was ‘fuzzier’ and varied according to the rapidity of Calluna regeneration. Although a generalization of a range of responses, Gimingham’s (1959) representation of the phasal growth of heather allows a possible framework for interpreting these classes in terms of the age of Calluna regrowth (Fig. 2). It should be noted, however, that in Gimingham’s (1959) analysis plant height and coverage were linked, as in a dense stand, whereas the colour and textures being interpreted in the aerial photographic interpretation (API) resulted primarily from the density of visible Calluna and not its height. Class 1 represented the commencement of Calluna pioneer stage before plant size and density made it readily visible within the image. It was frequently possible to distinguish within this class examples of very new burns (1–2 years old),


Fig. 1. Examples of appearance of burn classes in 25-cm aerial imagery


Fig. 2. Illustrative relationship between API of growth following burning in Calluna vulgaris within the growth phase contexts of Gimingham (1959). Note there is not a perfect correspondence between growth phases and API interpretable classes. It should be noted that under certain circumstances, particularly on wetter areas of bog, this simple interpretation may not hold. See note in text.

which looked dark as a result of the presence of charcoal. The remains of charcoal weather over time, and burns can appear much lighter at 2– 4 years old. Class 2 represented a period between early pioneer and early

building phases, although it could represent patchy regeneration of older plants. Class 3 probably represented the early to late building phase and into early maturity, and was distinguished primarily by a smooth


texture. Class 3 was terminated by the change in appearance as stem length reaches the stage where central gaps begin to appear in the foliage of each plant. This textural change was readily apparent in 25-cm resolution imagery and defined Class 4. This interpretation of Calluna regrowth is simplified in that it does not take account of regeneration by layering, which occurs particularly in wetter areas (Macdonald et al. 1995). This may have some implications for API of Calluna growth as it may not make the transition to a visually ‘leggy’ (API class 4) form but remain as a low-growing, sparse shrub. However, this did not influence the interpretation of either burn extent or change for areas of recent burn, i.e. Classes 1 and 2.

     


The duration of API Classes 1 and 2 was estimated with reference to burn parcels whose age was known. A sample of 180 burn parcels, for which 2005 field data were available from another project, were allocated to an API Class (1, 2 or 3) from 2005 photography. As API classes were interpretive and no absolute definition existed for them, accuracy measures based on a comparison with ‘truth’ were not appropriate in this case. The age of the classified burn parcel was checked against dated photography. While it was not possible to follow individual burns through their total life span, the year in which a burn appeared could be used to establish its age. Gaps in photography meant that exact dating was possible for 88 burn scars of the 180 available. For the remaining burn scars the range of possible ages was refined with reference to quadrat field data on heather height and a known heather height/age relationship, derived from the 88 exactly dated burn scars [height (cm) = 3·8 × season – 0·033; d.f. = 87, F = 432·236, P < 0·001]. For each quadrat, the measured height of the heather was divided by the slope of the heather height–age relationship to produce an age prediction. The average predicted age of all quadrats within a burn, rounded up to the next highest growth season, was taken to be the age of that burn, unless this fell outside the known possible age range of the burn, in which case the predicted age was taken to be the nearest known possible value. ‘Survivorship’ curves for Class 1 and 2 burns were produced by plotting the proportion of burn scars still classified in those classes against growth season (Fig. 3). Regressions fitted to the approximately linear portion of these curves (as the curves were roughly logistic in form, trailing 1s and 0s at either end of this curve carried no information and were excluded) were used to predict the duration of each class, defined as the length of time for half of the burn scars to move from Class 1 to Class 2 or from Class 2 to Class 3. The regression lines for Class 1 and Class 2 crossed 0·5 at 3·82 years and 7·71 years, respectively (proportion in Class 1 = 1·233 − 0·192 × season; proportion in Class 2 = 1·903 − 0·182

Fig. 3. ‘Survivorship’ curves for Class 1 and Class 2 burn scars, showing the proportion of burn scars remaining in each class in each season of growth according to two independent observers. (a) Proportion of burn signatures remaining in Class 1. (b) Proportion of burn signatures remaining in Class 2. The values at which the regression lines cross 0·5 are taken to be the median durations of Class 1 and Class 2 burn scars. (circles, observer 1; crosses, observer 2).

× season). These figures were taken to be the median end points of these classes in subsequent analyses.

    If the proportion of the area of a sample square in Class 1 and Class 2 is C1 and C2, respectively, and their durations are D1 and D2, then the time for the entire square to be burnt (R) is simply: R=

D1 + D2 C1 + C 2

Thus, if Classes 1 and 2 cover 10% of a sample square and if the total duration of these classes is estimated to be 10 years, then the repeat time is 10/0·1, or 100 years. Where consistent management was not seen in the sample squares it was possible to estimate a value of R


Table 1. Distribution of target classes (%) within the sample sites. Estimates are from CEH CS2000 Land Cover map data (Fuller et al. 2002) Class

Segments Minimum Mean Maximum SD

Bog 71 (34) Heath 165 (79) Bracken 88 (42) Acid grass 159 (76) Montane 0 (0) All target 208 (100)

0·2 0·1 0·1 0·1 – 5·1

29·5 23·7 16·1 23·1 – 53·4

98·9 100·0 78·8 100·0 – 100·0

1·9 1·7 1·4 1·5

Results          A full representation of target classes in the 208 sample squares is summarized in Table 1. The most frequent classes recorded within the sample squares were dwarf shrub heath and acid grassland, occurring in 79% and 76% of samples, respectively (Table 1).

   based on the presence of Class 1 or Class 2 parcels only. This is of interest for sites where management has only just commenced (i.e. within the duration of Class 1) and is likely to continue. This equation assumes an even spread of burns in each year class. An estimation of R based on class 1 alone (D1/C1) may yield the same result whether the value of C1 is derived from a single large burn in the previous year or several smaller burns over the length of D1.

Table 2. Estimates of extent of managed burning in the English uplands. Baseline for year 2000 Interpreted class

Area (km2) CI 95% (±) Percentage

No DSH

DSH

Class 1

Class 2

Class 3

Class 4

7834 451 75·6

2526 451 24·4

437 122 4·2

328 100 3·2

250 79 2·4

1510 310 14·6

The values shown represent results aggregated for the complete study area of 10 360 km2. DSH/Calluna covered 24% of the study area, with burning management equating to 7% and the presence of Calluna without visible evidence of management equating to 17% of the study area.

Table 3. Estimates of extent of management burning in upland DSH/Callunadominated communities in England. Baseline for year 2000 Interpreted class, as area

% of total DSH CI 95% (±)

Class 1

Class 2

Class 3

Class 4

17·3 4·3

12·9 3·6

9·9 2·8

59·8 9·3

Proportion burned by time Last 4 years

Last 8 years

Last 20 years

NVM

% of total DSH 17·3 30·3 40·2 59·8 © CI2006 95%The (±) Authors. 4·3 6·6 8·5 9·3 Journal compilation © 2006 shown Britishare the estimates for the extent of burning within DSH /Calluna Values Ecological dominated Society, communities in upland England. Estimates for amount burned in preceding Journal of Applied years based on cumulative class periods and show typical age ranges observed from field Ecology , 43,of API class. verification 1138–1148 NVM, no visible (burning) management.

No unambiguous burn management was identified on any grass/sedge-dominated moorland. To confirm that the resolution of the photography or the interpretation process were not preventing identification of burn events in these habitats, additional data were acquired from the Exmoor National Park Authority comprising field mapping of burn parcels for the years 1997–2000 inclusive. One map, produced in 1997, occurred within a sample segment. Closer inspection of this image confirmed that no identifiable signature of the burn remained in the photograph (dated 18 June 2000). The transient nature of the grassland burn signature, and the low frequency of burning relative to sampling density, rather than imagery constraints, was therefore responsible for the failure to detect burning in these habitats in this study.

     Visible evidence of burning was found consistently within non-grass/sedge moorland. Of the 208 sample squares, 106 contained moorland with visible presence of ericaceous dwarf shrubs (DSH), representing 24·4% of the sampled area. Visible evidence of burning was seen in 75 (71%) of these. The proportions of interpreted classes and derived estimates of the extent of management burning within the complete study area are summarized in Table 2. Interpretation of these figures as proportions of ericaceous-dominated moorland showed that class 1 burns (estimated as a maximum of 3·8 years old) and class 2 burns (aged from 3·8 to 7·7 years) occupied 17% (95 ± 4·3% CL) and 13% (95 ± 3·6% CL) of the total area of these habitats in upland England, respectively. The full results for all classes are presented in Table 3. Of the 106 1-km2 sample sites with dwarf shrub heath/Calluna present, 55 contained sufficient evidence of current management for a repeat time to be estimated. Estimated repeat times for the level of management observed in 2000 for all sites showing recent evidence of burning are summarized in Table 4. The repeat times for sites assessed as consistently managed lay between 14 and 25 years (overall median 20·1 years; Fig. 4), a period that is probably not too intensive for conservation purposes on continuously managed Calluna stands on shallow peat or mineral soil. Indeed, it is longer than


   

Fig. 4. Distribution of estimated burn repeat times for all consistently managed sites (n = 55).

Table 4. Median burn repeat times in years for managed sites within the English uplands

History of management

Lower 95% CI

Central estimate

Upper 95% CI

Consistent (n = 55) Recently introduced (n = 71)

14·0 16·8

20·1 19·4

24·9 29·4

Medians of estimated repeat times for consistently managed sites (where Class 1 and 2 heather present) and sites that show only recent (Class 1) burns. Confidence intervals of medians from binomial distribution. Length of Classes 1 and 2, 3·82 and 3·89 years, respectively.

that sometimes recommended for heather moor (Tucker 2003). However, it is not clear what proportion of the data came from areas of bog that should not be burnt at this frequency (Shaw et al. 1996). Estimation of repeat times based on the presence of Class 1 parcels alone (n = 71) yielded an estimated median repeat time of 19·4 years. Scaling the figure of 437 km2 Class 1 dwarf shrub heath in Table 2 by the 3·82-year average duration of this class showed that, for the period 1996–2000, an average of c. 114 km2 of moorland was burnt in the English uplands each year.

Table 5. Summary of changes in burning recorded in paired 1-km2 samples between 1970 and 2000 Area percentage

1970s mean (SE)

2000 mean (SE)

DSH 62·7 (5·7) 62·6 (5·3) Class 1 15·1 (2·3) 29·7** (3·8) Class 2 18·9 (2·8) 18·9 (3·2) © 2006 The Authors. Classes 1 and 2 combined 34·0 (4·1) 48·6* (4·8) Journal Class 3 compilation 25·7 (5·4) 15·8* (3·3) © 20064 British Class 40·3 (6·4) 35·7 (5·6) Ecological Society, Journal of Applied *P < 0·05; **P < 0·001; paired t-test, data arcsin square-root transformed prior to Ecology, analysis; 43, d.f. = 22. Because of spatial bias in sampling these cannot be used for national 1138–1148 estimates.

Considerable changes in the extent of different burn classes were apparent in the studied squares between the 1970s and the year 2000 (Table 5). Of the two ‘new burn’ classes, the area of Class 1 almost doubled, from 15·1% to 29·7% (t = 5·31, d.f. = 22, P < 0·001, paired t-test), with no change in Class 2. Within historical monochrome photography there is some potential for error in assigning burn signatures between Classes 1 and 2 as both may appear very bright. To check that the observed increase was not an artefact of photographic misinterpretation, the percentage of both classes was pooled for each year and re-analysed. This produced a significant increase for the combined classes from 34·0% to 48·6% (t = 3·21, d.f. = 22, P < 0·01, paired t-test) and a corresponding decrease in the classes most likely to be burnt, i.e. classes 3 and 4 decreased from 25·7% to 15·8% and 40·3% to 35·7%, respectively. This increase in the area of new burn resulted in a statistically significant reduction in estimated repeat time for the sample squares, from a median of 19·5 (n = 21) to 16·1 (n = 23) years using the pooled values for Classes 1 and 2 and their total average duration (Wilcoxon signed ranks test, n = 21, T+ = 54, P < 0·05). Estimating the repeat time of sites using Class 1 area and duration only yielded figures of 27·8 (n = 20) and 18·0 (n = 23) years, respectively (not tested). It is now nearly impossible to find any extensive or continuous areas of ericaceous-dominated moorland in the uplands of England that have not been burned within the past 10 years. At many places the changes in intensity of burn management over the last 20 years are remarkable (Fig. 5). The example shown lies within the Peak District National Park and was declared an SSSI in 1992. This increase has been driven by management for red grouse.

Discussion Burning as a management practice within the uplands is predominantly executed in habitats with a significant frequency of dwarf shrubs. Grassland burning, which may be locally important in the south-west of England, is a much less significant feature of the landscape nationally, although this is partly because of the short duration of burn signatures on grassland. For a typical area of DSH/Calluna at year 2000, the area of moor showing visible evidence of burning (Classes 1, 2 and 3) was approximately 40%. The area of the newest burns (Class 1), a measure of current activity, was 17·3%, equating to c. 114 km2 of the English uplands being burnt annually. This figure is not just for areas of dry heath but includes all forms of wet heath and bog with visible presence of dwarf shrubs. This study is the first large-scale analysis of current burning practices in England, hence there are few other estimates of the extent or intensity of burning with which to compare the results. The only comparable


Table 6. Changes in estimated median burn repeat times (years) for paired 1-km2 samples between 1970 and 2000 Median repeat time

Sites with Classes 1 and 2 present Sites with Class 1 present

1970

2000

19·5 25·4

16·1* 16·5†

*P < 0·05, Wilcoxon signed ranks test, n = 21, T+ = 58. †Not tested. Estimate derived using age estimate for Class 1 as 0–3·82 years and Class 2 as 3·83–7·71 years. Because of spatial bias in sampling these cannot be used for national estimates.

Fig. 5. Example of change in burn practice at the landscape scale. Sequence shows the startling transition of a large area of wet heath/bog to intensive commercial shoot. This is far from unusual and the final pattern is typical of the majority of upland moors in England today. For reference this area was notified as a SSSI in 1992.


study was undertaken in Scotland covering the period 1940–90 (Hester & Sydes 1992). Hester & Sydes’ (1992) concern was that burning was in decline (because of increased human resource costs), which might lead to the eventual loss of dwarf shrub heath, although no evidence of long-term changes was in fact found. A repeat time of the order of 50–100 years was reported, however, representing 1–2% DSH burnt year−1. This is a much less intense burning management regime than that found in this study (repeat time of c. 20 years on managed sites, with more than 4% of the area of ericaceous moorland burnt year−1 overall). Bardgett, Marsden & Howard (1995) studied the distribution of heather in English moorlands and, although their definition of moorlands is not directly comparable to that used here and they did not present figures for the extent of burning, noted that areas of DSH managed for grouse seemed to be of better average condition than other (generally more southerly) areas. Although the headline figures reported here for median repeat times are informative at a national level,

they should be interpreted cautiously. Repeat times were calculated from the total area of ericaceous moorland within each kilometre square that exhibited areas of burn. They did not therefore include samples that showed no evidence of burning. Hence the figures for repeat times reported are not for the whole of the area of ericaceous moorland but the area of ericaceous moorland that is under burning management. It is also possible that some areas within an image may never be burnt, perhaps because they are too small and isolated, on steep slopes or in excessively ‘hagged’ (eroded) areas. In such situations, the reported repeat time will be an overestimate. Controversy exists regarding the correct burning management protocol for the English uplands (Tucker 2003; Glaves & Haycock 2005; Select Committee on Environment, Food & Rural Affairs 2005). This controversy is driven by both a lack of scientific evidence and the different aims of various groups (principally grouse rearing and nature conservation). Indeed, the Select Committee on Environment, Food & Rural Affairs (2005) states: ‘It is important that the various parties involved can reach a common science-based understanding of the impacts of burning on moorland’. One problem is our imprecise knowledge of the interaction of burning with other ecological processes, particularly grazing, which is the subject of much current research (Freckleton 2004; Vandvik et al. 2005). The effect of local environmental variation on the results of such interactions may be complex (Palmer & Hester 2000; Hulme et al. 2002; Pakeman et al. 2003; Fuhlendorf & Engle 2004) and the spatial effect of burning management may influence the behaviour of important herbivores, so that it is difficult to predict the outcome of burning regimes without detailed study (Palmer & Hester 2000; Fuhlendorf & Engle 2004). It has been widely stated in the literature that too-frequent burning of dwarf shrub heath leads to an increase in Graminae and Cyperaceae at the expense of dwarf shrubs, principally Calluna (Hobbs & Gimingham 1984; Hobbs 1984; Miles 1988; Shaw et al. 1996; Marrs et al. 2004), although other processes may be important, particularly in bogs, for example overgrazing (Anderson & Yalden 1981; Pakeman et al. 2003), nitrogen deposition (Tomassen et al. 2004) and draining (Stewart & Lance



1991). Similarly, too-infrequent burning of moorland may lead to Graminae becoming dominant (Hester & Sydes 1992) via the poor regeneration of over-mature heather. Such changes are difficult to reverse (Pakeman et al. 2003; Marrs et al. 2004). Miles (1988) indicated that burning at 3–6-year intervals favours the grasses Deschampsia flexuosa (L.) Trin. on well-drained soils and M. caerulea on poorly drained soils. Only three sites (of 71 recently managed) have an estimated repeat time this short, but Rawes & Hobbs (1979) describe how 10-year rotations can favour hare’s-tail cotton grass Eriophorum vaginatum L. over Calluna on blanket mire. Burning management of any kind may be detrimental to blanket bog (Shaw et al. 1996; Stewart, Coles & Pullin 2004) but it may be that the most severe damage occurs when uncontrolled fires occur at long intervals (Imeson 1971; Tallis 1987; Maltby, Legg & Proctor 1990), which some sort of regular burning regime might prevent (Mackay & Tallis 1996). Conversely, many sites (31 of 106) had no visible management, which in practice means that no burning has occurred there for 20 years or more. The results presented here show that, at least in some English national parks, the overall values for proportion burned arise in part from a highly significant increase in intensity of burning between the 1970s and 2000, with a concomitant decrease in repeat times from 19·2 years to 15·8 years. This general trend includes areas that show a pronounced intensification in burning, possibly at the expense of many nature conservation objectives, a concern supported by the high proportion of SSSI currently failing condition assessments because of inappropriate burn regimes. In terms of understanding the significance of this change, we emphasize that no attempt was made in this study to distinguish between differing ericaceous moorland types. These are overall figures and therefore include not only areas of dry upland heath on mineral soils and thin peat, where such burning regimes are sustainable (Tucker 2003), but also areas of blanket bog, where they are not (Stewart, Coles & Pullin 2004). It was not the primary objective of this study to either assess the ecological impact of burning or pass judgement regarding the appropriateness or otherwise of observed levels of activity. While it could be argued that the new burning regimes identified here may appear to be more intense than they will ultimately be, both the extent of new burns and the degree of change observed in many places over the past few decades are striking. We are concerned about the possible adverse ecological impacts of this level of management. We also note that this process has occurred in areas that have some of the greatest conservation protection available in England, i.e. most are SSSI lying inside national parks. Note also that, despite the rate of change observed here, the estimate for year 2000 does not appear to represent a peak in activity and, as Fig. 5 clearly shows, in some areas the level of burning activity observed in 2000 is small in comparison with the activity observed in 2005.

A problem in addressing the Select Committee on Environment, Food & Rural Affairs (2005) concern regarding the lack of scientific consensus on the impacts of burning has been the lack of any comprehensive information on the extent of this activity. It is surprising that there is no national system of licensing, consenting and monitoring burn practices, even on SSSI. Much of the frequently acrimonious debate about burning in the uplands has been allowed to continue in a void of factual information regarding the true extent of its practice, or any assessment of changes over the recent past. This study goes some way to addressing this issue and will, hopefully, allow land managers and all concerned in this debate to reach a considered agreement. If, as the data presented here suggest, there has been an increase nationally in the intensity of burning regimes in the English uplands, what could be the driving forces behind such changes? One possibility is the introduction of the Phase II Environmentally Sensitive Areas programme in 1988 (Defra 2002). Where this programme relates to moorland (Tier 1c), it aimed to address perceived long-term management failings on moorland relating to overgrazing and reduced burning. The first 9 years of this programme led to a fourfold increase in burning in the North Peak area (ADAS 1997) and, although this was felt to be principally on dry heath rather than ‘dry bog’ concern was raised that the amount of burning on ‘dry bog’ was too high. There is an obvious need for further data on the extent and distribution of burning management, particularly the extent to which bogs, even if already badly degraded, are subject to burning. Subsidies exist elsewhere to promote controlled burning in England (the entry-level stewardship scheme; Defra 2005). For example, the North York Moors National Park Authority grantaided 21·6% of its SSSI moorland to be burnt in the four winters from 1995 to 1998 (Moorland Regeneration Programme 2001). Anecdotal reports by land agents, managers and gamekeepers during this project also indicated a trend to ‘commercialize’ more of the shooting activities on many moors. Incomes from such practices can be high and what has traditionally been considered a traditional country sport may in places have become more of an ‘industrial’ activity.

Acknowledgements This project was managed by Alistair Crowle at English Nature and we are grateful to him for his comments and guidance throughout this study. The help of Richard Pollitt at English Nature is also gratefully acknowledged. We thank David Lloyd and Alison Cox of the Exmoor National Park, and Suzanne Goodfellow and David Partridge of the Dartmoor National Park Authority. We would particularly like to thank numerous land managers and gamekeepers, including Ben Hayes of the Bolton Abbey Estate. We thank G. B. Stewart and two anonymous referees for their helpful comments on an earlier draft of the manuscript.


References

ADAS (1997) Environmental Monitoring in the North Peak ESA 1988–96. ADAS Report to the Ministry of Agriculture, Fisheries and Food. Defra, London, UK. Anderson, P. & Yalden, D.W. (1981) Increased sheep numbers and the loss of heather moorland in the Peak District, England. Biological Conservation, 20, 195 –213. Bardgett, R.D., Marsden, J.H. & Howard, D.C. (1995) The extent and condition of heather on moorland in the uplands of England and Wales. Biological Conservation, 71, 155 – 161. Birks, H.J.B. (1988) Long-term ecological change in the British uplands. Ecological Change in the Uplands (eds M. B. Usher & D.B.A. Thompson), pp. 37 –56. Blackwell Scientific Publications, Oxford, UK. Brown, T. (1997) Clearances and clearings. Deforestation in Mesolithic/Neolithic Britain. Oxford Journal of Archaeology, 16, 133 –146. Defra (2002) Environmentally Sensitive Areas. http:// www.defra.gov.uk/erdp/schemes/esas/default.htm (accessed 17/12/2005). Defra, London, UK. Defra (2005) Entry Level Stewardship Scheme. http:// www.defra.gov.uk/erdp/schemes/els/handbook/chapter3l.htm #el6 (accessed 19/12/2005). Defra, London, UK. Dimbleby, G.W. (1952) The historical status of moorland in north-east Yorkshire. New Phytologist, 51, 349 –354. English Nature (2006) Reasons for Adverse Condition Summary. http://www.englishnature.org.uk/special/sssi/reportAction. cfm?Report=sdrt17&Category=N&Reference=0 (accessed 19/4/2006). English Nature, Peterborough, UK. Freckleton, R.P. (2004) The problems of prediction and scale in applied ecology: the example of fire as a management tool. Journal of Applied Ecology, 41, 599– 603. Fuhlendorf, S.D. & Engle, D.M. (2004) Application of the fire–grazing interaction to restore a shifting mosaic on tallgrass prairie. Journal of Applied Ecology, 41, 604– 614. Fuller, R.M., Smith, G.M., Sanderson, J.M., Hill, R.A., Thomson, A.G., Cox, R., Brown, N.J. & Gerard, F.F. (2002) Countryside Survey 2000 Module 7: Land Cover Map 2000. Final Report, Cslcm/Final. Unpublished report to Defra. Centre for Ecology & Hydrology (CEH), Monks Wood, Huntingdon. Garnett, M.H., Ineson, P. & Stevenson, A.C. (2000) Effects of burning and grazing on carbon sequestration in a Pennine blanket bog, UK. The Holocene, 10, 729 –736. Gimingham, C.H. (1959) The Maintenance of Good Heather. Enquiry into the Decline of Red Grouse. 5th Progress Report. Scottish Landowners Federation, Edinburgh, UK. Glaves, D.J. & Haycock, N.E. (2005) Science Panel Assessment of the Effects of Burning on Biodiversity, Soils and Hydrology. Report to Defra Conservation, Uplands and Rural Europe Division, Upland Management Branch. Haines-Young, R.H., Barr, C.J., Black, H.I.J., Briggs, D.J., Bunce, R.G.H., Clarke, R.T., Cooper, A., Dawson, F.H., Firbank, L.G., Fuller, R.M., Furse, M.T., Gillespie, M.K., Hill, R., Hornung, M., Howard, D.C., McCann, T., Morecroft, M.D., Petit, S., Sier, A.R.J., Smart, M.S., Smith, G.M., Stott, A.P., Stuart, R.C. & Watkins, J.W. (2000) Accounting for Nature: Assessing Habitats in the UK Countryside. Centre for Ecology and Hydrology and © 2006 The Authors. Department of Environment Transport and the Regions, UK. Journal compilation Hanuschak, G., Sigman, R., Craig, M., Ozga, M., Luebbe, R., © 2006 British Cook, P., Kleweno, D. & Miller, C. (1979) Obtaining Timely Ecological Society, Crop Area Estimates Using Ground-Gathered and Landsat Data. Journal of Applied Technical Bulletin No. 1609. United States Department of Ecology, 43, Agriculture, Economics, Statistics and Cooperatives 1138–1148 Service, Washington D.C.

Hester, A.J. & Sydes, C. (1992) Changes in burning of Scottish heather moorland since the 1940s from aerial photographs. Biological Conservation, 60, 25 –30. Hobbs, R.J. (1984) Length of burning rotation and community composition in high-level Calluna–Eriophorum bog in N. England. Vegetatio, 57, 129 –136. Hobbs, R.J. & Gimingham, C.H. (1984) Studies on fire in Scottish heathland communities. II. Post-fire vegetation development. Journal of Ecology, 72, 585 –610. Hulme, P.D., Merrell, B.G., Torvell, L., Fisher, J.M., Small, J.L. & Pakeman, R.J. (2002) Rehabilitation of degraded Calluna vulgaris (L.) Hull-dominated wet heath by controlled sheep grazing. Biological Conservation, 107, 351–363. Imeson, A.C. (1971) Heather burning and soil erosion on the North Yorkshire moors. Journal of Applied Ecology, 8, 537– 542. Kuhry, P. (1994) The role of fire in the development of Sphagnumdominated peatlands in western boreal Canada. Journal of Ecology, 82, 899–910. Macdonald, A.J., Kirkpatrick, A.H., Hester, A.J. & Sydes, C. (1995) Regeneration by natural layering of heather (Calluna–Vulgaris): frequency and characteristics in upland Britain. Journal of Applied Ecology, 32, 85 –99. Mackay, A.W. & Tallis, J.H. (1996) Summit-type blanket mire erosion in the Forest of Bowland, Lancashire, UK: predisposing factors and implications for conservation. Biological Conservation, 76, 31– 44. Maltby, E., Legg, C.J. & Proctor, M.C.F. (1990) The ecology of severe moorland fire on the North York Moors: effects of the 1976 fires, and subsequent surface and vegetation development. Journal of Ecology, 78, 490 –518. Marrs, R.H., Phillips, J.D.P., Todd, P.A., Ghorbani, J. & Le Duc, M.G. (2004) Control of Molinia caerulea on upland moors. Journal of Applied Ecology, 41, 398 – 411. Miles, J. (1988) Vegetation and soil change in the uplands. Ecological Change in the Uplands (eds M.B. Usher & D.B.A. Thompson), pp. 57–70. Blackwell Scientific Publications, Oxford, UK. Moorland Regeneration Programme (2001) Final Report: October 1995–March 2001. North York Moors National Park Authority, Helmsley, UK. Pakeman, R.J., Hulme, P.D., Torvell, L. & Fisher, J.M. (2003) Rehabilitation of degraded dry heather [Calluna vulgaris (L.) Hull] moorland by controlled sheep grazing. Biological Conservation, 114, 389 – 400. Palmer, S.C.F. & Hester, A.J. (2000) Predicting spatial variation in heather utilization by sheep and red deer within heather/ grass mosaics. Journal of Applied Ecology, 37, 616–631. Pearsall, W.H. (1950) Mountains and Moorlands. Collins, London, UK. Petersen, I., Masters, Z., Hildrew, A.G. & Ormerod, S.J. (2004) Dispersal of adult aquatic insects in catchments of differing land use. Journal of Applied Ecology, 41, 934–950. Pitkänen, A., Turunen, J. & Tolonen, K. (1999) The role of fire in the carbon dynamics of a mire, eastern Finland. The Holocene, 9, 453 – 462. Ratcliffe, D.A. & Thompson, D.B.A. (1988) The British uplands: their ecological character and international significance. Ecological Change in the Uplands (eds M.B. Usher & D.B.A. Thompson), pp. 9–36. Blackwell Scientific Publications, Oxford, UK. Rawes, M. & Hobbs, R. (1979) Management of semi-natural blanket bog in the northern Pennines. Journal of Ecology, 67, 789 – 807. Ross, S., Adamson, H. & Moon, A. (2003) Evaluating management techniques for controlling Molinia caerulea and enhancing Calluna vulgaris on upland wet heathland in northern England, UK. Agriculture, Ecosystems and Environment, 97, 39– 49. Select Committee on Environment, Food and Rural Affairs (2005) Fourteenth Report. http://www.publications.parliament.uk/



pa/cm200304/cmselect/cmenvfru/475/47506.htm (accessed 19/12/2005). Shaw, S.C., Wheeler, B.D., Kirby, P., Phillipson, P. & Edmunds, R. (1996) Literature review of the historical effects of burning and grazing of blanket bog and upland wet heath. English Nature Research Report No. 172. English Nature, Peterborough, UK. Simmons, I.G. (2003) The Moorlands of England and Wales: An Environmental History 8000 BC–AD 2000. Edinburgh University Press, Edinburgh, UK. Smith, A.A., Redpath, S.M., Campbell, S.T. & Thirgood, S.J. (2001) Meadow pipits, red grouse and the habitat characteristics of managed grouse moors. Journal of Applied Ecology, 38, 390 – 400. Stevenson, A.C., Rhodes, A.N., Kirkpatrick, A.H. & Macdonald, A.J. (1996) The Determination of fire Histories and an Assessment of their Effects on Moorland Soils and Vegetation. Survey and Monitoring Report No. 16. Scottish Natural Heritage Research, Edinburgh, UK. Stewart, G.B., Coles, C.F. & Pullin, A.S. (2004) Does Burning Degrade Blanket Bog? Systematic Review No. 1. Centre for Evidence-Based Conservation, Birmingham, UK. Stewart, A.J.A. & Lance, A.N. (1991) Effects of moor draining on the hydrology and vegetation of northern Pennine blanket bog. Journal of Applied Ecology, 28, 1105 –1117. Tallis, J.H. (1987) Fire and flood at Holme Moss: erosion processes in an upland blanket mire. Journal of Ecology, 75, 1099 –1129. Tharme, A.P., Green, R.E., Baines, D., Bainbridge, I.P. & O’Brien, M. (2001) The effect of management for red grouse shooting on the population density of breeding

birds on heather-dominated moorland. Journal of Applied Ecology, 38, 439 – 457. Thompson, D.B.A., MacDonald, A.J., Marsden, J.H. & Galbraith, C.H. (1995) Upland heather moorland in Great Britain: a review of international importance, vegetation change and some objectives for nature conservation. Biological Conservation, 71, 163 –178. Tomassen, H.B.M., Smolders, A.J.P., Limpens, J., Lamers, L.P.M. & Roelofs, J.G.M. (2004) Expansion of invasive species on ombrotrophic bogs: desiccation or high N deposition? Journal of Applied Ecology, 41, 139 –150. Tucker, G. (2003) Review of the Impacts of Heather and Grassland Burning in the Uplands on Soils, Hydrology and Biodiversity. Research Report No. 550. English Nature, Peterborough, UK. Usher, M.B. & Thompson, D.B.A. (1993) Variation in the upland heathlands of Great Britain: conservation importance. Biological Conservation, 66, 69 – 81. Vandvik, V., Heegaard, E., Maren, I.E. & Aarestad, P.A. (2005) Managing heterogeneity: the importance of grazing and environmental variation on post-fire succession in heathlands. Journal of Applied Ecology, 42, 139 –149. Webster, R. & Oliver, M.O. (2000) Geostatistics for Environmental Scientists. Wiley, Chichester, UK. Whittingham, M.J., Percival, S.M. & Brown, A.F. (2000) Time budgets and foraging of breeding golden plover Pluvialis apricaria. Journal of Applied Ecology, 37, 632– 646. Received 6 January 2006; final copy received 9 June 2006 Editor: Phil Hulme