For at least 2000 years woodland in Britain has been managed to greater or lesser extents (Rackham 2003, Smout 2003). Prior to the 20th century much of the ancient woodland, particularly in England, was managed as coppice or coppice with standards. Often the coppice system was abandoned during the latter part of the 19th century, but the First and Second World Wars led to heavy felling to meet the war demands (HMSO 1952, Richards 2003).
Sources of data
Data on the overall changes in the composition and structure of British broadleaved woodland since 1947 have been assembled primarily from the various forestry censuses (HMSO 1952, Locke 1970, 1987, Forestry Commission 2003a). There are some differences in the minimum size of woodland considered and the categories of woodland recorded in the different censuses, but these are not likely to significantly affect the broad comparisons made in this paper. The census results are supplemented by various long-term monitoring studies. Most of these are of individual sites, such as those at Lady Park Wood (Gloucestershire) or Clairinsh Island (Loch Lomond) (Peterken & Jones 1987, Backmeroff & Peterken 1989), but Kirby et al. (2005) provide data from a re-survey of 103 broadleaved woods spread across Britain. These were first recorded using a random sample of 16 200-m2 plots in each wood in 1971; between 1999 and 2002 (the ‘2001’ records) these plots were re-surveyed using the same methods. Data are also drawn from structural records from the Repeat Woodland Bird Survey (Amar et al. 2006) and the Countryside Survey (1990, 1998), the latter based on stratified random survey of plants and habitats in 1-km squares (Haines-Young et al. 2000).
Patterns of change since 1947
In 1947 70% of the broadleaved woodland was in England, particularly in the southeast, 20% in Scotland and 10% in Wales; proportions in 2002 were similar. This review is inevitably therefore dominated by the changes that have affected lowland England. Not all of these changes will be as relevant to the uplands, particularly to northwest Scotland, but the important ecological differences do not necessarily relate to political boundaries: oakwoods in Cumbria have more in common with those in Argyll or Gwynedd (and with those in Killarney) than they do with lime–hornbeam woods of East Anglia.
Felling of mature broadleaves continued after 1945, until c. 1980, as part of the post-war policies to increase home timber production through replacing broadleaved stands with conifers (NCC 1984, Roberts et al. 1992, Spencer & Kirby 1992). However in most woods that remained broadleaved, and particularly post-1980, relatively little management has taken place compared with previous centuries. For example, only about 30% of woodlands on private land in England (which makes up 60% of the British broadleaved resource) have a felling licence or are in some form of woodland grant scheme (Slee et al. 2006). In 68 of 103 woods visited in 2001, surveyors reported that there was no recent management activity (Kirby et al. 2005).
Many of the areas recorded in the forestry censuses as coppice in 1947 and 1967 have now grown up and are classed as high forest (Table 1). Coppice rotations were usually less than 30 years and hence, in worked coppice, many of the stands were either open or young growth (Evans 1984). Such stands are highly valued in biodiversity terms (Buckley 1992, Fuller & Warren 1993). The shift from coppice to a high forest regime, where rotations may be over 100 years (Evans 1984), inevitably reduces the proportion of the stands that are open space and young growth (Table 2) and hence may reduce associated woodland species.
Table 2. Contrasting expected age distributions for woods managed as coppice and high forest under a regular felling cycle.
|Age category||Percentage area in different age categories|
|Coppice with standards||High forest|
|Open space (< 5 years)||12|| 4|
|Young growth (6–30 years)||62||21|
|Thicket to early mature (31–100 years)||20||58|
|Mature (> 100 years)|| 6||17|
For example, of six butterfly species associated with clearings in woodlands, three have shown marked declines: a 77% decline since 1970–82 in the case of the High Brown Fritillary Argynnis adippe (Asher et al. 2001). The higher proportion of the early mature growth stages in high forest regimes may benefit some canopy species (Hambler & Speight 1995), but these forest stages tend to have a less distinctive flora and fauna than those at the beginning and end of the growth cycle (Warren & Key 1991). Generalist deadwood species should benefit from the increases in deadwood reported (Kirby et al. 2005, Amar et al. 2006). However, specialist saproxylic species tend to be poor colonists and so may not spread to sites where there has not been a continuity of appropriate conditions (Warren & Key 1991).
Intensive coppice regimes in Scotland appear to have been less common (e.g. Samsum 2005) and had already virtually disappeared by 1947. However, in Scotland, a much higher proportion of the woods was classed as scrub (half the total scrub area of Great Britain) – defined as ‘inferior growth unlikely to develop into a utilisable crop of coppice, poles or timber’ (HMSO 1952). Similar processes of growth and canopy expansion are likely to have taken place as with the coppice stands in England.
The shift to high forest (from scrub or coppice) and changes within the structure of high forest stands are supported in trends in the basal area, size distribution of stems, and canopy cover reported in detailed studies. In 103 broadleaved woods surveyed in 1971 and 2001, Kirby et al. (2005) found that basal area of woody stems had increased overall; small stems (5–20 cm diameter breast height, dbh) tended to have declined whereas larger size classes had increased. Using different methods, Amar et al. (2006) found no increase in mean basal area in their study of 406 woods, but did note an increase in tree height and of subcanopy cover (4–10 m), which would be consistent with a general growth of trees and shrubs in many woods. Similar patterns of increasing stem sizes and/or increasing basal area have also been shown in long-term monitoring studies at individual sites, for example in Wytham Woods (Oxfordshire) (Kirby & Thomas 2000), Monks Wood (Cambridgeshire) (Crampton et al. 1998), Dendles Wood (Devon) (Mountford et al. 2001) and Clairinsh Island (Backmeroff & Peterken 1989). At other sites, such trends may then be disrupted by major disturbances such as droughts or storm, for example Lady Park Wood (Peterken & Jones 1987) and The Mens (West Sussex) (Mountford 2004).
Vascular plant richness tends to increase under woodland gaps, whether caused for example by storms (Buckley et al. 1994), ride management (Buckley et al. 1997), coppice cuts (Barkham 1992) or clear-fells (Kirby 1990), and decline as woody species cover is re-established (Mitchell & Kirby 1989). In their national woodland survey Kirby et al. (2005) found a reduction in ground flora species richness associated with increasing basal area of trees and shrubs (basal area often being closely associated with canopy cover). Woodland specialists declined in frequency more often than generalist species while species increasing in abundance tended to be those associated with semi-shade as opposed to open habitats. Haines-Young et al. (2000) found similar evidence for increasing shade in the changes in broadleaved woodland flora recorded in the Countryside Survey 2000 for England and Wales.
These national analyses hide local and regional patterns of change in stand structure, for example the contrasting changes in the old and young growth stands in Lady Park Wood, which reflect past management (Peterken & Jones 1987, 1989). Natural events such as storms also impact at local and regional scales. Quine and Gardiner (2002) report the incidence of approximately one regionally significant storm every 2 years between 1945 and 2000. The severe storm in early 2005 mainly affected woods in northwest England (C. Quine pers. comm.), whereas that of 1987 had little effect outside southeast England (Kirby & Buckley 1994). Ten of the sites surveyed by Kirby et al. (2005) were in the track of the 1987 storm. In 2001 these were more likely to show an increase in plant species richness than sites that were outside the storm track, reflecting a tendency for storms to create gaps in otherwise shaded woods.
Changes in tree and shrub composition of broadleaved woods
Since 1947 the woody composition of broadleaved woods has shown some changes (Table 3). Small fluctuations between survey dates and differences in minor species may not mean much because there is a large area that is not differentiated to species in the reports. However, there does appear to be a decline in the relative contributions of Oak Quercus robur/petraea and Beech Fagus sylvatica (but less change in absolute area terms) and an increase in Ash Fraxinus excelsior and Sycamore Acer pseudoplatanus. In the 103 woods surveyed by Kirby et al. (2005) the frequency of occurrence of the main species showed little change (Table 4). This may indicate therefore that the increases in Ash and Sycamore in part reflect the abundance of these latter two species in new woodland rather than increases within existing sites.
Table 3. Changes in the composition of British broadleaved woodland based on Forestry Commission census data (HMSO 1952, Locke 1970, 1987, Forestry Commission 2003a).
| ||Area (ha) (% of total broadleaved area in parentheses)|
|Oak||210 069 (31.0)||205 182 (28.2)||190 044 (25.5)||20 6154 (22.7)|
|Ash|| 44 261 (6.5)|| 52 348 (7.2)|| 79 204 (10.6)||119 232 (13.1)|
|Beech|| 67 997 (10.0)|| 68 056 (9.3)|| 75 006 (10.1)|| 76 551 (8.5)|
|Birch||140 893 (20.8)||170 081 (23.4)||131 391 (17.6)||155 355 (17.1)|
|Sweet Chestnut|| 22 678 (3.3)|| 22 834 (3.1)|| 29 226 (3.9)|| 10 800 (1.2)|
|Sycamore|| 27 623 (4.1)|| 31 214 (4.3)|| 54 291 (7.3)|| 61 357 (6.8)|
|Alder|| 9388 (1.4)|| 17 327 (2.4)|| 8363 (1.1)|| |
|Hornbeam|| 5961 (0.9)|| 1498 (0.2)|| 3823 (0.5)|| |
|Poplar|| 1331 (0.2)|| 7854 (1.1)|| 13 590 (1.8)|| 10 418 (1.1)|
|Lime|| 719 (0.1)|| 729 (0.1)|| || |
|Elm|| 9711 (0.1)|| 8380 (1.1)|| 9514 (1.3)|| 3743 (0.4)|
|Willow|| 2426 (0.4)|| 526 (0.1)|| 4964 (0.7)|| |
|Norway Maple|| 89 (0.01)|| || || |
|Cherry|| 46 (0.01)|| || || |
|Hazel|| 60 305 (8.9)|| 38 421 (5.3)|| 11 656 (1.6)|| |
|High forest undifferentiated|| 373 (0.5)|| 25 141 (3.4)|| 94 211 (12.6)||237 111 (26.2)|
|Coppice undifferentiated|| 50 408 (7.4)|| 2672 (0.4)|| 9287 (1.3)|| 23 526 (2.6)|
|Scrub undifferentiated|| 24 055 (3.5)|| 75 384 (10.4)|| 30 600 (4.1)|| |
Table 4. Composition of broadleaved woodland 1971–2001 based on resurvey of 103 woods (Kirby et al. 2005).
| ||Frequency of occurrence in 103 sites (1648 plots)|
|% of plots||% of sites|
|Field Maple|| 9|| 8||28||31|
Sweet Chestnut Castanea sativa and Hazel Corylus avellana were the two most common coppice species in 1947 and their decline in the record since then is likely to be in part due to coppice moving into the high forest category and then being classified by the overstorey tree species rather than by the composition of the coppice.
Individual tree species have shown contrasting changes in their size distribution over recent decades both in a national survey (Kirby et al. 2005) and in various individual monitoring studies. Oak nationally showed net loss from all size classes smaller than 50 cm dbh and an increase in the larger size classes. Suckering Elm Ulmus spp. was hit by Dutch elm disease and declined in most size classes nationally; Wych Elm (U. glabra) showed a similar pattern at Lady Park Wood (Peterken & Mountford 1998). Large Beech declined in the national survey and at Lady Park Wood probably because of drought (Peterken & Mountford 1996), although on some sites it is becoming more frequent and its national area has increased slightly (Table 3). In the understorey, Hazel showed a marked decline in the number of stems in the 5–10 cm diameter class nationally and also in Monks Wood (Crampton et al. 1998), whereas the shade-tolerant Holly Ilex aquifolium increased nationally and in The Mens and on Clairinsh (Mountford 2004).