ELSEVIER 0006-3207(95)00087-9 Biological Conservation 76 (1996) 31~4 Copyright © 1996 Elsevier Science Limited Printed in Great Britain. All rights reserved 0006-3207/96/$15.00+.00 SUMMIT-TYPE BLANKET MIRE EROSION IN THE FOREST OF BOWLAND, LANCASHIRE, UK: PREDISPOSING FACTORS AND IMPLICATIONS FOR CONSERVATION Anson W. Mackay* & John H. Tallis School of Biological Sciences, University of Manchester, Oxford Road, Manchester M13 9PL, UK (Received 6 December 1994; accepted 7 June 1995) Abstract The incidence of summit-type mire erosion has been investigated in a mire complex in the Forest of Bowland, Lancashire, using a combination of palaeoecological and radiometric techniques in conjunction with documentary evidence. Macrofossil analyses suggest that Sphagna have been a dominant component of the vegetation on Fairsnape Fell over the last 2000 years. However, over the last 100 years, Sphagnum species have disappeared fi'om many areas. We conclude here that this decline was caused by the onset of summit-type erosion, which in turn was initiated through a combination of unusual factors." a period of below-average rainfall in the region in the early 1900s, resulting in lowered water tables in the peat; exceptional summer drought in 1921; and a decline in management standards because of shortage of gamekeepers after the First Worm War. We believe that these precipi- tated a catastrophic burn (probably in 1921), which may well have been accidental in origin. High levels of atmo- spheric pollution did not cause the loss of Sphagnum, but may have influenced their ecology. The heather moorland and blanket mires of Bowland are internationally impor- tant habitats for many species of bird of prey, as well as being economically important for grouse shooting. Consequently, catastrophic fires that cause large-scale degradation of upland peat landscapes have important implications for the conservation of these ecosystems and of the scientific archive preserved in peat ecosystems. It is suggested that high sheep-stocking levels on Bowland may prevent recolonisation of bare peat surfaces, thereby allowing peat erosion to continue. Keywords." accidental fires, atmospheric pollution, moorland management, peatland conservation, Sphagnum remains. INTRODUCTION Blanket mire erosion is widespread throughout Great Britain: for example, in northern Scotland and the *Present address: Environmental Change Research Centre, University College London, 26 Bedford Way, London, WC 1H 0AP. Correspondence to: A.W. Mackay Fax 0171-380-7565; e-mail amackay@geog.ucl.ac.uk 31 Outer Hebrides (Osvald, 1949; Walker & Walker, 1961; Lindsay et al., 1988), south-west Scotland (Stevenson et al., 1990), north-west England (this study), northern England (Bower, 1959; Radley, 1962; Tallis, 1965, 1985a, b, 1987), south-west England (Taylor, 1983), Wales (Bostock, 1980; Robinson & Newson, 1986; Francis, 1990). Indeed, blanket mire erosion is so widespread as to have important implications for con- servation and land-use policies throughout the British uplands. First, erosion poses a direct threat to the very exis- tence of these habitats: ecologically, they are interna- tionally important in terms of animal biodiversity (Thompson et al., 1995), whereas economically they are important as grazing land for sheep and red grouse, as reservoir catchment areas, and as recreation and tourist attractions. Secondly, the incidence of blanket mire erosion may have indirect but important implications for global climate change (Heathwaite, 1993) through the general lowering of water tables (Bell & Tallis, 1974; Mackay, 1993). These changes in hydrological conditions will almost certainly produce an increase in the fluxes of important greenhouses gases such CO2 and NzO (Moore & Knowles, 1989; Freeman et al., 1993). Controversy still prevails, however, as to the actual causes of blanket peat erosion. Various theories have been proposed: climate change (Conway, 1954; Bower, 1961); biotic factors such as burning (Pearsall, 1950; Radley, 1965; Stevenson et al., 1990), grazing (Shimwell, 1974), atmospheric pollution (Tallis, 1964b; Ferguson & Lee, 1983) and forest clearance (Mackay, 1993; Tallis, 1995); and as a natural endpoint to blanket peat development (Lewis, 1906; Johnson, 1957). Knowledge of the time of onset of erosion is clearly important in assessing the possible contribution of these various fac- tors. Recent studies have emphasized that erosion is a long-standing feature of many blanket mire systems, perhaps pre-dating intensive upland land-use and pol- lution (Tallis, 1985a, 1987; Bradshaw & McGee, 1988; Livett & Tallis, 1989; McGee & Bradshaw, 1990; Stevenson et al., 1990). Dating of the inception of ero- sion may usually only be achieved by indirect methods, as definite documentary evidence is rarely available. Techniques employed include estimations of contempo- rary rates of erosion (thereby allowing the calculation