Information and Advisory Note Number 115                                            Back to menu

The effects of mammalian herbivores on natural regeneration of upland, native woodland

1. Introduction

1.1 Semi-natural woodland and scrub cover about 4% of the Scottish landscape. Grazing and browsing are major factors preventing, or restricting, the regeneration of woody species on many of these areas on unenclosed hill land.

1.2 The two herbivore species that are considered to have the greatest impact across Scotland are red deer and sheep, which have upland populations of about 350,000 and 2,350,000 respectively. Several other domestic and wild species such as cattle, roe and sika deer, horses and ponies, goats, hares, rabbits and voles can also affect rates of tree regeneration.

1.3 Under natural conditions, herbivore numbers would have fluctuated with weather conditions and with predator numbers. Wild boar would also have scarified the ground, creating regeneration niches. Conditions suitable for tree regeneration were therefore likely to have occurred often enough to maintain woodland cover across large parts of Scotland. Winter feeding of sheep and red deer, land improvement and a lack of predators have all led to herbivore numbers being kept continuously at artificially high levels.

2. Requirements for successful tree regeneration and establishment

2.1 Successful natural tree regeneration requires
• a source of viable seed or suckers;
• suitable temperatures, and adequate rainfall and light, to allow seedlings and saplings to grow1;
• the presence of mycorrhizal fungi for some tree species;
• a suitable seed bed where roots can reach the mineral soil, nutrient levels and soil moisture are sufficiently high and there is not too much competition from other vegetation; and
• seed predation and browsing levels low enough to allow sufficient numbers of saplings to reach a height where they are no longer susceptible to browsing.

Grazing and browsing directly affect the last two of these conditions.

2.2 All Scotland's dominant native tree species require relatively high light conditions to establish (see 3.3) hence successful regeneration and establishment happens either at the woodland edge or in gaps created by the death of adults, or by felling. In theory, the correct combination of conditions for regeneration and establishment need occur only once during the lifetime of each adult tree for it to be replaced and for a woodland to survive. Normal maximum lifespans for several native tree species in Scotland are given below. Little is known about how the reproductive potential of different tree species is affected by age and growing conditions. For this reason, it is hard to predict how frequently successful regeneration and establishment events must occur to ensure the survival of a particular woodland.

¹ There is no consistent distinction in the literature between a seedling and a sapling. In this note a sapling is taken to be a young tree taller than about 25 cm and a seedling is one less than about 25 cm.

2.3 If grazing and browsing are very heavy, regeneration will be completely suppressed but if less heavy, the density and species composition of surviving saplings depends on

• the pre-browsed density and growth rate of seedlings and saplings;
• the severity of browsing on seedlings and saplings; and
• the responses of seedlings and saplings to being browsed.

An understanding of the factors affecting each of these will help in planning appropriate grazing management for woodland regeneration.

3. The pre-browsed density and growth rate of seedlings and saplings

3.1 These are affected by the first four factors listed in section 2.1.

3.2 Availability of viable seeds depends on the abundance of seeds produced by local seed sources, the distance to those sources, the presence of the agent of dispersal and the viability of the seed. Abundance and viability of seed produced by a tree can vary greatly from year to year depending particularly on weather conditions. Both production and viability of seed often decline with altitude. For wind-dispersed species, such as birch, pine, willow, alder and ash, most seed falls within a distance of about twice the height of the parent tree and there is little regeneration at more than 1 km, though small numbers of birch seeds can travel to at least 3 km and those of pine to at least 7 km. Oak, rowan, juniper, hazel and holly require an appropriate animal vector to be dispersed beyond the canopy of the parent tree. In Britain, aspen reproduces largely by suckering. The seeds of most tree species do not remain viable in the soil for more than a year but birch seeds can survive for two to three years and ash seeds survive for up to six years. Ash and rowan seeds take at least 18 months to germinate.

3.3 Seedling germination and survival may be affected by lack of light, not only under an existing tree canopy but also where there is dense heather, bracken or bramble cover. Pine, birch, alder and willow seedlings are all light demanding. Oak seedlings can survive beneath a canopy for up to four years and ash will also tolerate shade for the first few years but then becomes light demanding. Rowan is tolerant of limited shade and can regenerate under a birch canopy whereas birch itself cannot. Holly seedlings are shade-tolerant and holly can reach 5-10 m in height underneath a woodland canopy. The high palatability of rowan and holly probably accounts for the rarity of woodlands dominated by these species.

3.4 Competition for nutrients and light from ground layer vegetation e.g. a dense carpet of dog's mercury or wavy hair grass, will limit seedling survival. Grazing helps to reduce the cover of ground layer vegetation.

3.5 Mycorrhizal fungi are found associated with almost all trees in long standing woodlands. Mycorrhizae seem to improve the ability of trees to absorb nutrients, especially phosphate, make them more resistant to infection and may also improve water economy. They thus allow seedlings to grow more quickly and seem to be essential where site conditions are particularly demanding e.g. at the tree line. Mycorrhizal fungi may be absent from sites that have been without tree cover for many years but they should quickly colonize any soil other than peats or podzols.

3.6 A deep raw humus layer, or dense ground layer vegetation, can increase the chances of seedlings drying out by preventing their roots gaining access to the mineral soil.

3.7 Where conditions are good, and growth rates high, there is a greater chance that saplings will get their leading shoot out of reach of large herbivores before they are browsed. However, where soil and tree nutrient levels are high, it is possible that there will be an increased chance of browsing (see 4.2.3).

4. The severity of browsing on seedlings and saplings

4.1 This depends on

• the species and condition of seedlings and saplings present;
• the nature of the ground layer vegetation
• the density and height of seedlings and saplings;
• the herbivore species; and
• the seasonality, intensity and spatial distribution of grazing and browsing.

4.2 Species and condition of seedlings and saplings present

4.2.1 Herbivores are selective feeders, their preferences being influenced by several attributes of the foliage of different tree and shrub species. These are thought to include digestibility, shoot biomass, fibre content, nutrient concentrations, level of toxic secondary compounds and spininess.

4.2.2 In general, the relative preferences of red deer and sheep for the foliage of different native tree species are given below.

4.2.3 All the attributes of a tree's foliage that determine its preference ranking can vary throughout the year, with the age of the foliage and with growing conditions. Digestibility tends to be higher, and fibre content lower, in new, spring growth. Birch is seldom eaten by red deer when it is not in leaf unless little other food is available. It is often heavily browsed at bud burst. Nutrient concentrations in foliage can be higher if the tree is growing on more fertile soils and there is evidence that deer and sheep selectively browse trees growing on more nutrient-rich soils.

4.3 The nature of the ground layer vegetation

4.3.1 The preference shown by a herbivore for a particular tree species is crucially dependent on what else is on offer. Pine and juniper are not favoured in summer when other food is usually abundant because the leaves contain terpenes, which are distasteful. Deer will, however, browse pine shoots around the time of bud burst. Pine and juniper are eaten in winter when deciduous trees and shrubs are bare of foliage and when other vegetation is of low digestibility or is relatively inaccessible under snow. Leafless shoots and twigs are normally not eaten unless there is a scarcity of other vegetation.

4.3.2 When seedlings are about the same height as the surrounding ground vegetation, they will often be eaten along with it. Seedlings therefore have a greater chance of survival if they are growing amongst a non-preferred vegetation type. Thus seedlings growing amongst tall heather are less likely to be browsed than those growing amongst short, grassy vegetation.

4.3.3 Most preferences are over-ridden by hunger. Red deer will browse pine in summer or birch in winter if there is no alternative and, when pressed, all herbivores may eat foliage or twigs or strip bark from species of trees which they would otherwise avoid.

4.3.4 Heavy grazing on ground layer vegetation can reduce the availability of good quality forage and lead to an increase in browsing rates on trees. When grazing pressures are low, ground cover may be high and seedlings will be less easily detected by herbivores until they emerge above the level of the surrounding vegetation. Even then, if there is an abundant source of alternative food, saplings will only be readily browsed by herbivores which have a high preference for browse (goats, roe deer and, to some extent, red deer).

4.3.5 When ground layer vegetation is tall and abundant, field vole numbers are likely to increase and they can also damage seedlings and saplings (see 5.1.5).

4.4 The density and height of seedlings and saplings

4.4.1 For a given herbivore density, the chances of achieving successful rates of sapling establishment are likely to increase as the initial, pre-browsed, density of saplings increases. However the density of established saplings may not increase in direct proportion to any increase in the initial density of saplings. Little work has been carried out on the form of this relationship so it is currently not possible to predict the likely effect of changing the initial abundance of saplings on final sapling density.

4.4.2 Seedlings and saplings which project above the level of the surrounding vegetation are more likely to be browsed until they grow taller than the mouth height of any herbivores present. At Beinn Eighe, browsing by red deer was found to be most severe on trees of 1-2 m in height. Browsing by sheep is likely to be most severe on shorter saplings. Maximum browse heights for different herbivore species are given below.

5.1 The herbivore species

5.1.1 The main impact of browsing by sheep and red deer is at the seedling and small sapling stage, when they clip leading, and lateral, shoots. Sheep of most breeds, and cattle, show less of a tendency to browse than do red deer. However, the more primitive breeds, such as Hebrideans, browse more readily and are used for scrub control on some grassland and heathland nature reserves. Sheep are also reputed to displace red deer from favoured areas hence a reduction in sheep numbers may lead to an increase in deer browsing on tree seedlings and saplings in these areas. Red and sika deer can strip bark from saplings and roe deer fray them with their antlers. Bark stripping is more likely than fraying to remove bark in a ring around the tree and therefore to kill it. Otherwise, within woodland, roe and sika deer can have similar impacts to those of red deer.

5.1.2 Cattle and ponies can cause damage by eating seedlings and by trampling. Ponies have a greater tendency to browse than do cattle. They also strip bark from trees of ail sizes.

5.1.3 Feral goat populations are restricted to a small number of sites in Scotland but, where they occur, they can have a major effect on tree regeneration because of their high propensity to browse. They tend to browse taller saplings than do sheep and can affect regeneration on cliff ledges or boulder fields that are inaccessible to other herbivore species.

5.1.4 Mountain hares browse and bark strip saplings, as do rabbits, and both can be particularly harmful when herbs and grasses are under snow cover. Rabbit damage can occur throughout the year but tends to increase in winter and early spring. Bark stripping is commonest in hard winters and late springs when snow persists.

5.1.5 Field Voles can be significant browsers of tree seedlings and, where they occur in high numbers, will often ring-bark saplings. One study found that voles killed 73% of seedlings regenerating on open moor by eating the main shoots. Wood mice and squirrels eat tree seeds and bank voles may eat small tree seeds.

5.2 The seasonality, spatial distribution and intensity of grazing and browsing

5.2.1 Because of the seasonal changes in the attractiveness of different species of seedling and sapling and in the abundance and digestibility of the ground layer vegetation, the same density of herbivores can have a very different effect on regeneration at different times of year. In summer, when ground layer vegetation is usually more abundant, saplings are less likely to be browsed than in winter, when food is often in short supply. On the other hand, deciduous saplings are more attractive in summer, when they have leaves.

5.2.2 Herbivores do not normally distribute themselves evenly across a landscape and young trees will suffer a heavier browsing pressure when they grow within, or adjacent to, heavily used areas. Forage quality and abundance and, in poor weather conditions, shelter, are the major factors affecting the distribution of large herbivores (See SNH l&A Note 47, The grazing behaviour of large herbivores in the uplands). Thus, areas of nutritious grassland on good soil will always be preferred and red deer tend to spend more time in at lower altitudes in winter than in summer. The latter can lead to lower browsing pressures, and more successful regeneration of young trees, at higher altitudes. For rabbits, hares and voles, cover, such as long grass or scrub, which gives protection from predators, is also a major factor affecting their distribution in the landscape.

5.2.3 Grazing and browsing pressure obviously have a large effect on regeneration success but numbers of successfully regenerating trees do not necessarily keep increasing as herbivore numbers decline to low levels. This is because, at low grazing pressures, the ground layer vegetation builds up resulting in both fewer regeneration niches and increased competition for soil nutrients and light.

6. The responses of saplings to being browsed

6.1 The ability of saplings to survive and resprout after browsing varies with
• the age and species of tree or shrub;
• the proportion of the plant removed;
• the type of damage;
• the plant's growing environment; and
• the season in which the browsing occurs.

6.2 Mature trees are relatively resistant even if their lower branches are completely removed up to the grazing height of the herbivores. Young trees are more at risk as a greater proportion of total foliage area may be removed. Trees up to 2 m tall are vulnerable and saplings of less than 50 cm are often killed.

6.3 Even so, birch, rowan and pine can all tolerate removal of up to half of the current year's growth. The effect of removal of more than this differs with species. Pine seems least tolerant of damage. Two year old pine usually dies if eaten back into the previous season's wood whereas most birch and all rowan are likely to survive. Other broadleaves are probably similarly resistant.

6.4 Trees which are tolerant of browsing may be kept small above ground but develop an extensive root system, enabling them to respond rapidly if browsing pressure is reduced or removed. In general, recovery is likely to be better on sites with favourable growing conditions including good soil nutrient levels and climatic conditions although, under these conditions, saplings may have to compete with a vigorous growth of ground layer vegetation.

6.5 Browsing on seedlings and saplings in early summer, when nutrient concentrations in the shoots and leaves are high, is likely to be more detrimental than browsing in late summer or winter, when most of the nutrients have been re-allocated to stems and roots. This applies to birch and may also apply to other broadleaved species. Coniferous species store nutrients in their leaves as well.

7. What herbivore densities will allow tree regeneration?

7.1 Given the range of factors involved, and that large herbivores do not usually distribute themselves evenly across the range available to them, it is not surprising that there are no clear thresholds for the numbers of deer or sheep at which adequate regeneration may occur. The few studies which have been carried out in this area can, however, give us an idea of the range of herbivore densities that might be appropriate. It should be borne in mind that, even at low overall densities, grazing pressures in preferred areas can be very high.

7.2 Hester, Mitchell and Kirby (1996) studied the effects of sheep grazing and browsing on regeneration in an upland wood in Cumbria at three stocking densities and in either summer or winter. The study looked at germination and survival rates of seedlings and saplings of downy birch, rowan and ash.

7.3 Although germination rates were higher under heavier grazing pressures {2.1-3.8 sheep ha¯¹), probably because of the higher number of regeneration niches, there were still large numbers of seedlings appearing even in the most lightly grazed plots {0.6-1.2 sheep ha¯¹).

7.4 Seedling survival was higher at lower grazing intensities, and in summer-grazed plots. The latter was probably because summer grazing prevented the build-up of ground vegetation that would compete with seedlings.

7.5 At high levels of summer grazing (2.1-3.4 sheep ha¯¹), almost all the saplings were browsed whereas at low levels (0.6-0.9 sheep ha¯¹), only 25% of the saplings were browsed. In winter, browsing rates on saplings were generally lower, with 40% of saplings remaining undamaged even at 3.8 sheep ha¯¹. The higher browsing rates in summer are likely to be because the saplings had leaves at this time.

7.6 A parallel study of browsing of planted saplings also found a decreasing impact as sheep densities decreased, apparently because the sheep browsed saplings less frequently when more ground-layer vegetation was available. However, planted saplings were more heavily browsed on average than naturally regenerated saplings, which is commonly the case.

7.7 The winter browsing of naturally regenerated, leafless birch saplings at Glen Feshie has been found to be linearly related to local densities of red deer (Miller & Cummins 1998) with the browsing rate ranging from about 5% of birch saplings being browsed per week at about 200 red deer km¯² (2 ha¯¹) to < 1 % of saplings browsed per week at less than about 25 red deer km¯² (0.25 ha¯¹). Deer numbers on low ground are often over 150 km¯² in winter.

7.8 Several attempts have been made to estimate the density of red deer that might allow successful woodland regeneration. Fraser Darling (1947) guessed at 6 km¯², Holloway (1967) found partial regeneration at 4 km¯² and unimpeded regeneration at 2 km² and Mitchell, Staines & Welch (1977) found regeneration occurring at about 10 km¯². The latter was at a site with a relatively high density of saplings of around 4,000 saplings ha¯¹. Ratcliffe (1988) found that in forests there was a heavy impact on most broadleaved trees at red deer densities of about 7 deer km¯².

7.9 At Creag Meagaidh, following culling of red deer, regeneration of birch from previously suppressed saplings appears to be taking place at densities of 8 red deer km¯². At Abernethy Forest densities of up to 12 km¯² in 1988/9 were reduced by culling to about 5 km¯² by 1993. Roe were also culled but the proportion of the population this represented is not known. A 20% increase in the number of established seedlings and saplings was recorded over this time, with rowan accounting for 68% of new saplings. Most of the rest were pine (the first species to respond), with some birch and willow.

7.10 Densities of herbivores low enough to permit regeneration may still be high enough to affect the composition of ground layer vegetation. At about 5 red deer km¯², Ratcliffe (1988) suggested that plant diversity in forests would be reduced.

8. Planning regeneration

8.1 Thus, some general conclusions on the impact of grazing on tree regeneration in the uplands are as follows.

• There will normally be insufficient seeds of wind-dispersed species to initiate much regeneration at distances of more than about 1 km from existing woods. Seed densities fall off rapidly with distance from the parent tree.
• Ground conditions may be unsuitable for regeneration, especially if there is a deep raw litter layer or dense vegetation.
• Grazing and trampling may assist initial establishment of seedlings by suppressing competing vegetation, creating regeneration niches and limiting field vole abundance.
• Grazing and browsing impact on regenerating trees is likely to be concentrated in the most heavily used areas such as those providing high quality grassland or winter shelter.
• Seedlings are less likely to be browsed where the surrounding vegetation partly conceals them but if the vegetation is dense there will be more competition with the seedlings for resources.
• The greatest grazing and browsing pressure will fall on the most preferred species present.
• There may be less browsing impact on saplings where sufficient, preferred foods are available elsewhere.
• Stopping grazing (e.g. by fencing) may lead to increases in grass growth, field vole numbers and seedling destruction by voles.
• Some trees are relatively tolerant and will recover if pressure reduces or ceases.
• Tolerance and recovery will be better on sites where the growing conditions are more favourable but herbivores may selectively browse those on more nutrient-rich sites.
• Deer are more likely to browse saplings than are sheep or cattle.

8.2 This suggests that total exclusion of grazing may be a mistake where there is a lack of germination and /or low densities of seedlings. Where adequate numbers of seedlings or saplings already exist, very low grazing pressures may be the best means of allowing these young trees to establish. Herbivores are an important component of natural ecosystems, so the preferred policy is to achieve regeneration without the need for their total removal.

8.3 Options include:

• Where there are sufficient numbers of suppressed seedlings or saplings already present, the elimination, or drastic reduction, of grazing and browsing pressure will allow the young trees to 'get away'. This regime should be continued until the saplings are tall enough to withstand browsing.
• Where there are insufficient numbers of seedlings or saplings present, the instigation of a low and/or periodic grazing regime may be appropriate. This could be brought about by a reduction in overall stock, or wild herbivore, densities or by periodic, or seasonal, exclusion of stock by fencing. Animal numbers might have to be progressively reduced until the optimum balance between the positive and negative effects of herbivores is reached. The regime should be continued until the required density of saplings becomes tall enough to withstand browsing. This is likely to be about the lifetime of a conventional fence. It should, however, be borne in mind that, if fences are used, they can have a number of important side effects (see the SNH I & A Note 59 Fences and upland conservation management).

• Replacement of red deer with sheep may lead to lower browsing rates on saplings.
• If ground layer vegetation is dense, mechanical scarification may be needed to increase the number of regeneration niches. Pigs and cattle can also be used for this purpose.
• Felling, or coppicing, of selected trees will increase the amount of light available and may increase the density of seedlings and/or suckers of light-demanding species.
• If the supply of seed is limiting, it may be necessary to increase the density of trees nearby either by encouraging the spread of woodland from existing sites or by planting trees to provide a seed source. This is a long term strategy and, if regeneration is wanted in the short term, direct planting will be necessary.

9. Further reading

1. Beaumont, D., Dugan, D., Evans, G. & Taylor, S. (1995) Deer management and tree regeneration in the RSPB reserve at Abernethy Forest In: Aldhous, J.R. (Ed). Our Pinewood Heritage. Forestry Commission, RSPB, SNH.

2. Fraser Darling, F.F. (1947) Natural history in the Highlands and Islands. Collins, London.

3. Harmer, R. (1995) Natural regeneration of broadleaved trees in Britain: III. Germination and establishment. Forestry, 68, 1-9.

4. Hester, A.J. & Miller, G.R. (1995) Scrub and Woodland Regeneration: Prospects for the Future. In: Thompson, D.B.A., Hester, A.J. & Usher M.B. (Eds). Heaths and Moorland: Cultural Landscapes. HMSO, Edinburgh.

5. Hester, A.J., Mitchell, F.J.G. & Kirby, K.J. (1996) Effects of season and intensity of sheep grazing on tree regeneration in a British upland woodland. Forest Ecology and Management, 88,99-106.

6. Hodge, S. & Pepper, H. (1998) The prevention of mammal damage to trees in woodland. Forestry Commission Practice Note No. 3.

7. Holloway, C.W. (1967) The effects of red deer and other animals on the natural regeneration of Scots pine. Ph.D. thesis. University of Aberdeen.

8. Kinnaird, J.W. (1974) Effect of site conditions on the regeneration of birch (Betula pendula Roth and B. pubescens Ehrh.). Journal of Ecology, 62, 467-472.

9. Miller, G.R. & Cummins, R.P. (1998) Browsing by red deer on naturally regenerated birch and juniper saplings on wintering ground at Glen Feshie. Scottish Forestry, 52,138-145.

10. Miller, G.R., Cummins, R.P. & Hester, A.J. (1998) Red deer and woodland regeneration in the Cairngorms. Scottish Forestry, 52, 14-20.

11. Miller, G.R., Kinnaird, J.W. & Cummins, R.P (1982) Liability of saplings to browsing on a red deer range in the Scottish Highlands. Journal of Applied Ecology, 19, 941-951.

12. Mitchell, B., Staines, B.W. & Welch, D. (1977) Ecology of red deer. Institute of Terrestrial Ecology, Cambridge.

13. Pepper, H. (1998) The prevention of rabbit damage to trees in woodland. Forestry Commission Practice Note No. 2.

14. Ratciiffe, P.R. (1988) The management of red deer populations resident in upland forests. In: D.C. Jardine (Ed). Wildlife Management in Forests. Proceedings of a symposium held by the Institute of Foresters.

15. Staines, B.W., Balharry, R. & Welch, D. (1995) Moorland management and impacts of red deer. In: Thompson, D.B.A., Hester, A.J. and Usher M.B. (Eds). Heaths and Moorland: Cultural Landscapes. HMSO, Edinburgh

10. Authors

John Andrews
Andrews Ward Associates
17 West Perry
PE18 0BX

Gordon Miller
 School Hill
 AB31 5TQ

Helen Armstrong
Woodland Ecology Branch
Forest Research
Northern Research Station
EH25 9SY

11. Acknowledgements

Comments were kindly provided by Alison Hester (Macaulay Land Use Research Institute), Richard Thompson, Brenda Mayle, Jonathan Humphrey and Robin Gill (Forest Research), Kate Holl and Roland Stiven (SNH), Steve Palmer (Institute of Terrestrial Ecology), Lars Edenius (Swedish University of Agricultural Sciences), Pete Reynolds (Capreolus consultants), Fiona Stewart and Dick Balharry.

12. Contacts for advice and information

Kate Holl
Land Use Group
Scottish Natural Heritage
2, Anderson Place
Tel: 0131447 4784

Duncan Stone
Land Use Group
Scottish Natural Heritage
27 Ardconnel Terrace
Tel: 01463 712221

  Back to menu