Sycamore (Acer pseudoplatanus L.) is an invasive, non-native species in Great Britain and its management in conservation areas is controversial. Climate change adds further uncertainty to decision making. We investigated the role of management history in determining present-day abundance and the effects of climatic variability on growth, photosynthesis and phenology at Wytham Woods, a UK Environmental Change Network (ECN) monitoring site. Relatively few sycamore trees were found in undisturbed ancient, semi-natural woodland and recent plantations, despite being common in other areas of the site. Sycamore grew more slowly than ash (Fraxinus excelsior L.), its principal competitor, but at a similar rate to pedunculate oak (Quercus robur L.) in the period 1993–2005. There were fewer sycamore than ash seedlings, regardless of which species dominated the canopy. Growth of sycamore was slower in dry periods than wet ones and lower photosynthetic rates were measured in canopy leaves under dry compared with wet soil conditions. This study therefore suggests that sycamore does not present a serious threat to undisturbed ancient woodland on the site and that it may eventually decline in areas of the site where it competes with ash, in the absence of disturbance. It may also decline under climate change if summer droughts become more frequent.
Sycamore has expanded its range in north-west Europe in recent centuries. It is non-native in Great Britain and has colonized widely since it was introduced, probably in medieval times (Jones, 1945). It grows quickly compared with most broadleaved species and can produce a good timber crop (Savill, 1997; Binggeli and Rushton, 1999). Its invasive nature has however led to sycamore being regarded as a threat to the conservation of native woodlands, particularly where ash has historically been the dominant tree species (Scurfield, 1959; Binggeli, 1992, 1993; Peterken, 1996). Attempts are often made to remove the species in conservation areas, although this is usually only practical for the most sensitive sites (Morton Boyd, 1992). The view of sycamore as a threat to biodiversity has moderated in recent years as studies have shown that it supports a range of epiphytes, herbivores and ground flora, comparable to those of many native species (Binggeli, 1993; Peterken, 2001). There is also evidence that sycamore and ash each tend to regenerate better under the canopy of the other species and may establish a cyclical pattern with dominance alternating between the two species (Waters and Savill, 1992; Savill et al., 1995). Sycamore may therefore potentially offer good opportunities for combining wood production with the support of biodiversity in some circumstances, such as the creation of new farm woodlands, where preserving existing tree species composition is not a priority. However, sycamore remains a controversial species and the necessity for control to protect conservation sites is still a matter for debate. Much information on the species is essentially anecdotal and there is a need for more detailed scientific study.
Climate change adds to the complexity of the issues surrounding sycamore. Current projections (Hulme et al., 2002) indicate that Great Britain is likely to become warmer, with proportionally less precipitation falling in summer, leading to an increased incidence of summer droughts in southern and eastern areas, including most of England. Modelling by Broadmeadow et al. (2005) indicates that the productivity of sycamore will decline in much of central, southern and eastern England because of its sensitivity to drought. The same projections indicated that most of its competitors, such as ash and pedunculate oak, would be less adversely affected. Drought sensitivity is consistent with sycamore's natural distribution, which is centred on cool, damp, mountainous regions in central Europe (Jones, 1945; Rusanen and Myking, 2003). It is also consistent with the observation that within semi-natural British woodlands, sycamore is most dominant in the relatively cool, wet areas of the north and west (Pigott, 1984; Rodwell, 1991; Forestry Commission, 1997). These conjectures are however based on the extrapolation of correlations with geographical variations in present-day climate. Geographical patterns are not an infallible guide to the climatic sensitivities of species as distributions and productivity can also be influenced by, for example, soil type and management history. It is therefore important to understand the underlying processes which control species responses to climate change and to look for direct evidence of climatic impacts.
There is some evidence of sycamore's drought sensitivity, although examples are not extensive. Lemoine et al. (2001) and Tissier et al. (2004) present evidence from within the natural range of sycamore in France, that its xylem vessels are relatively susceptible to cavitation under dry conditions. During drought in the UK in 1976, Coultherd (1978) reported some death of sycamore, but only in association with sooty bark disease (Cryptostroma corticale).
Rising temperatures may have an effect on the competitive balance between sycamore and other species, distinct from any effects of an increased frequency of droughts. In particular, the sensitivity of phenology to temperature is well known and earlier leafing of trees in recent decades has been documented at the European scale (Menzel et al., 2006) and within the UK (Sparks and Carey, 1995; Fitter and Fitter, 2002). There is some evidence (Fenn, 2005; T. Sparks, unpublished data) that the timing of sycamore leafing is more sensitive to temperature than that of ash and that its growing season is therefore lengthening to a greater extent. In these circumstances, warming may have the opposite effect to that of drought by giving sycamore a competitive advantage.
Further understanding of sycamore's ecology and ecophysiology is therefore necessary to develop a better understanding of its likely responses to climate change. Recent decades have seen a wide range of different weather conditions in Britain (Figure 1). The summer of 1995 was very dry (Marsh, 1996) and was followed by 2 years of unusually low rainfall; the period 1998–2002 was, however, marked by very high precipitation. The summer of 2003 was extremely hot and dry. This range of weather conditions provided an opportunity to investigate the effects on sycamore. We have done this by bringing together a range of physical and biological monitoring data collected at Wytham Woods, in southern England, under the UK Environmental Change Network (ECN) programme. This is a well-studied and instrumented site where it is also possible to take account of management history and to gain access to the tree canopy, by means of a walkway. This range of research and monitoring allows us to address the following questions:
1 Is there any difference in the extent to which sycamore has colonized different areas of Wytham Woods? Do different management histories make the stand more or less susceptible to invasion?
2 Is there any evidence that sycamore is out-competing ash or growing faster?
3 Has tree growth decreased or mortality increased in sycamore trees during dry compared with wet periods?
4 Does tree growth reflect changes in photosynthesis in the canopy during wet and dry periods?
5 Is there any interspecific difference in phenological responses to temperature which might affect the outcome of competition in the long term?
Throughout the paper we compare sycamore with ash because of the interest in their relative competitive advantages and pedunculate oak, which is also found widely in British woods. Oak and sycamore are both accessible from a canopy walkway at Wytham. Earlier work in the canopy showed that photosynthesis of sycamore was lower than that of oak (Morecroft and Roberts, 1999). This work was carried out in the summer of 1996, which was a relatively dry year, following an extremely dry summer in 1995, leading to reduced soil moisture levels. One possible explanation for the difference in photosynthetic rates was a greater sensitivity of sycamore than oak to these dry conditions. A Ph.D. study (Stokes, 2002) made similar measurements on the same trees in wetter conditions in 1999 and 2000 and gave us the opportunity to test this hypothesis.
Materials and methods
Wytham Woods (51° 46′ N, 1° 20′ W; UK National Grid: SP 46 08) cover ∼400 ha and include a wide range of different soil and vegetation types. It has been a research site, owned and managed by Oxford University, since the 1940s. Present and historical management of the site are well documented (Grayson and Jones, 1955; Gibson, 1986) and tree, shrub and ground layers have been monitored since the mid 1970s (Dawkins and Field, 1978; Kirby et al., 1996; Kirby and Thomas, 2000). Since 1992, the site has been part of the ECN, under which climate, air pollution, soils and selected animal populations have been monitored in addition to further recording of vegetation and tree growth (www.ecn.ac.uk; Sykes and Lane, 1996).
1 Undisturbed ancient semi-natural woodland: Ancient woodland is woodland which has had a continuity of forest cover since ∼1600; the period for which historical records are usually available in England (Peterken, 1981). At Wytham, this woodland was managed as a ‘coppice with standards’ system (mixture of coppice stools interspersed with full-height trees). However, coppicing was discontinued over the course of the twentieth century and these areas have been largely unmanaged for between 40 and 100 years (differing locations were abandoned at different times). Hazel (Corylus avellana L.) is the most frequent coppice species and pedunculate oak the most frequent standard.
2 Disturbed ancient woodland: Ancient woodland areas which were formally managed as coppice with standards, but converted to high forest during the twentieth century. Timber has been extracted at various times but they have not been clearfelled. Extensive natural regeneration has occurred, along with some localized planting.
3 Secondary woodland: Areas which have naturally reverted to closed canopy woodland over the last 200 years, having previously been grassland or wood pasture (with scattered trees but no continuous canopy). A small amount of localized planting has taken place and there has been some timber extraction.
4 Nineteenth-century plantations: Formerly open areas which were planted in the nineteenth century. This planting was largely ornamental with widely spaced trees, particularly of beech (Fagus sylvatica L.). Management has been minimal in recent decades.
5 Twentieth-century plantations: Plantations, mostly of beech and pedunculate oak, mixed with exotic conifers in some places, planted between 1950 and 1970. Some were planted on grassland and others on cleared ancient woodland areas. Most have been managed by thinning, following standard forestry practice.
For the last 30 years, only the twentieth-century plantations have been subject to silvicultural management, entailing occasional thinning.
A survey of tree species in 294 sample plots of 10 × 10 m (0.01 ha), systematically located on a 100-m grid, was carried out in the summer of 1993 and in a few cases of 1994 (subsequently we refer to 1993 to include plots surveyed in either year), following the ECN ‘baseline’ survey methodology (Sykes and Lane, 1996). Forty one of these plots were randomly selected for on-going monitoring, with tree diameter at breast height (d.b.h. – diameter at 1.3 m) measured every 3 years and tree height every 9 years. Up to 10 trees, over 5 cm d.b.h. were selected on the basis of proximity to random coordinates in each plot and the d.b.h. measured with a diameter tape. Tree height was measured with a hypsometer (Blume-Leiss altimeter, Berlin-Steglitz, Germany) at a known distance (measured with a tape measure) from each tree. Trees were classified according to their crown classes (Sykes and Lane, 1996):
1 Dominant – trees with crowns extending above the general level of the crown cover and receiving full light from above and partly from the side.
2 Subdominant – trees with crowns forming the general level of the crown cover and receiving full light from above but comparatively little from the sides.
3 Intermediate – trees shorter than those in the two preceding classes but with crowns extending into the crown cover formed by dominant and co-dominant trees; receiving a little direct light from above but virtually none from the sides.
4 Suppressed – trees with crowns entirely below the general level of the crown cover, receiving little light either from above or from the side.
We report data for the period 1993–2005, comprising four measurement periods of 3 years each. Tree seedlings (young trees or shrubs, grown from seed, with a d.b.h. less than 0.5 cm) were counted by species in each plot in 10 400 × 400 mm quadrats (selected with a random number table), within the plots.
The date of first leafing in all three species across the site as a whole was recorded from 1994 onwards. This was defined as the first observation of a leaf having fully emerged from the bud (but not expanded). Observations were made by professional scientists working on the ECN programme and visiting the same areas of the site regularly (several times per week).
Meteorological variables, including temperature and precipitation, were monitored with an Automatic Weather Station (Didcot Instruments, Didcot, UK) at a grassland area in the middle of the site as part of the ECN programme (Morecroft et al., 1998).
A scaffolding walkway, ∼12 m above the ground, gave access to the canopies of five sycamore and five oak trees. The sycamore trees were ∼50 to 100 years old (27 cm mean d.b.h.), the oak 150–200 years old (66 cm mean d.b.h.). The location is described in more detail by Morecroft and Roberts (1999) and Roberts et al. (1999).
In this paper, we compare rates of photosynthesis in 1996, when soil conditions were dry with those made in 1999 and 2000 when they were wet. Measurements were taken throughout the growing season when conditions were suitable (particularly that the leaves were dry). In 1996, an ADC LCA 2 infrared gas analyser with PLC(B) leaf chamber (ADC Ltd, Hoddesdon, Herts, UK) was used. In 1999 and 2000, this was replaced by a PP Systems CIRAS 1, with PLC(B) leaf chamber (PP Systems, Hitchin, Herts, UK). For each species, five leaves exposed to full sunlight in the upper canopy were measured on each of the five trees (intermediate and shade leaves were also sampled; data not presented). In 1996, a single measurement was taken for each leaf, in the later years a mean of five measurements over 1 min was used for each leaf. Measurements were taken in the middle part of the day (09.00–14.00 GMT) when gas exchange rates were maximal (Stokes, 2002). In order to estimate maximum net photosynthetic rates (Amax), only those measurements made when the light was saturating for photosynthesis are included here. This was defined as a photosynthetic photon flux density of greater than 1000 μmol m−2 s−1, as measured by the leaf chamber sensors. Further details can be found in Morecroft and Roberts (1999) and Stokes (2002).
Data were analysed using Systat 11 (Systat Software Inc., 2004). Differences in distribution of species between contrasting areas were tested using chi-squared tests. d.b.h. increments were analysed using analysis of variance (ANOVA) and repeat measures analysis of variance (RMANOVA). Initial investigations indicated that diameter increment in all species was related to starting d.b.h., so d.b.h. in 1993 was included as a covariate in ANOVA and RMANOVA models. Gas exchange measurements were compared visually with means and standard errors. Trees which died during the period and those for which measurements started after 1993 were not included in the analysis of growth. Phenological data were tested for correlations with temperature in preceding periods.
Distribution of sycamore
Sycamore was the second most frequently occurring tree species in the baseline survey plots (Table 1), with ash the most frequent species. Sycamore was not randomly distributed (chi-squared test: χ2 = 67.09, degrees of freedom (df) = 4, P < 0.0001) but was found more frequently in plots located in secondary woodland, disturbed ancient woodland and nineteenth-century plantations (Figure 3a). Proportionally, fewer plots in undisturbed ancient woodlands and twentieth-century plantations had sycamore trees. There were significant differences in the frequencies with which sycamore occurred with other species (χ2 = 20.09, df = 9, P = 0.017). The species with which it was most frequently associated was ash (Figure 3b); they were found together in 56 of the 143 plots in which sycamore occurred. Other species with which sycamore frequently occurred included elder (Sambucus nigra L.) and hawthorn (Crataegus monogyna Jacq.). Ash is also non-randomly distributed (χ2 = 14.24, df = 4, P = 0.007) being most common in secondary woodland (60 per cent of plots) and least common in ancient woodland (28 per cent of plots). It therefore shows similar patterns to sycamore, although the differences are less pronounced.
|Species||Number of plots|
|Species||Number of plots|
Growth and mortality
Between 1993 and 2005, mean diameter growth was lower in sycamore (1.5 ± 0.3 cm) than either ash (4.5 ± 0.5 cm) or oak (2.2 ± 0.5 cm). ANOVA showed a significant effect of species (F = 17.049, df = 2, P < 0.001) and starting d.b.h. (F = 15.239, df = 1, P < 0.001). Because of the effect of starting d.b.h., size and growth were investigated separately in canopy dominant, sub-dominant and intermediate trees (there were too few suppressed ash and oak trees to compare). In each of these three categories, the increment in d.b.h. was larger in ash than sycamore or oak (Figure 4a). The starting d.b.h. of sycamore and ash were very similar in canopy dominant and intermediate trees (Figure 4b); among sub-dominant trees, it was higher in sycamore than ash (Kolmogorov–Smirnov test, P = 0.041) (Figure 4b). Sycamore and ash increments were also compared in the 12 plots where the two species occurred together; in all plots, mean diameter growth of ash was higher than that of sycamore. The difference in growth rates between sycamore and ash is therefore not likely to be an artefact of size or canopy position. The mean starting diameter of canopy dominant oak trees was, however, substantially larger than those of ash and sycamore and the greater growth increment in oak than sycamore presumably reflects this; sub-dominant and intermediate oaks were similar in size to sycamore and similar increments were found (Figure 4b).
Measurements in 1993 and 2002 indicated that height had increased most in ash trees (1.7 ± 0.4 m), with little or no height growth in sycamore (0.2 ± 0.3 m) and oak (0.3 ± 0.7 m).
d.b.h. growth in each of the 3-year intervals was examined separately (Figure 5a) for each of the three species by RMANOVA. To facilitate comparison between species in different periods on a like-for-like basis, trees over 50 cm d.b.h. (six oaks, one sycamore, one ash) were excluded. Sycamore was the only species in which there was a significant difference with time (Table 2), with growth highest during the period 1999–2002 (Figure 5a), which climate records show was 24 per cent wetter than the other periods (Figure 5b). There was no significant difference between periods in oak (Table 2), although it showed a similar pattern to sycamore (Figure 5a). Ash grew more than the other two species in all periods (Figure 5a) with no significant difference between time periods (Table 2).
|Time||d.b.h. in 1993||Time × d.b.h.|
|Time||d.b.h. in 1993||Time × d.b.h.|
Time with starting d.b.h. (1993) as a covariate.
Mortality of the three species was low and there was no significant difference between species. Of the 290 trees and shrubs which were originally monitored in the plots, 31 died between 1993 and 2005 including three of the original 54 sycamores, two out of 21 oaks and one out of 54 ash.
There was no evidence that sycamore seedlings were more abundant under ash canopies, or vice versa, in the monitoring plots in either 1993 or 2002 (Figure 6). The larger baseline survey also failed to show any effect (data not presented). However, numbers of seedlings of all species were very low, particularly in 1993 and very few were older than 1 or 2 years old; saplings were almost absent from the wood.
Of the 3 years in which photosynthesis was measured in the canopy, Amax values in sycamore were lowest in the dry year, 1996 (Figure 7); The highest mean Amax on any particular day in 1996 was only 6.1 μmol m−2 s−1, compared with 12.1 μmol m−2 s−1 in 1999 and 10.4 μmol m−2 s−1 in 2000. Values of Amax in oak were also lower in 1996 (Morecroft et al., 2003) but the difference between years was smaller. Comparing the two species in different years, it can be seen that oak consistently had higher Amax than sycamore, but the difference between the two species was greatest in the 1996. This difference in Amax was associated with differences in stomatal conductance (the capacity for CO2 to diffuse into the leaves, which is largely regulated by the number and degree of opening of the stomata) of the two species.
Sycamore came into leaf earlier than the other species in 12 of the 13 years (Figure 8a). The date of leafing was significantly correlated with March temperature in all three species (Table 3, Figure 8b). This correlation was less strong in ash than the other two species and its leafng was more strongly correlated with mean March–April temperature than that for March alone (Table 3).
Bold text indicates significant differences: *P < 0.05; **P < 0.01.
Is there any difference in the extent to which sycamore has colonized different areas?
It is striking how little sycamore has colonized the undisturbed ancient woodland areas of Wytham Woods, despite its abundance in other parts of the site. This may reflect poor establishment and growth in shade as these areas have been largely unmanaged for much of the twentieth century. This is consistent with the observations of Morton Boyd (1992) and Savill (1997) that sycamore tends not to become dominant in woodlands with a dense canopy. There was also relatively little colonization of recent plantations which were managed for timber production; in this case, establishment of sycamore may have been prevented by weed control and thinning as well as low light levels following canopy closure. It is also possible that more colonization of recent plantations would take place with time.
Is there any evidence that sycamore is out-competing ash or growing faster?
We found no evidence of sycamore out-performing ash, its main competitor, either among mature trees or seedlings.
d.b.h. increment was consistently higher in ash than sycamore throughout the period and ash also showed greater height growth. Drought sensitivity (discussed further below) appears to have contributed to sycamore's overall lower growth rate, but it still grew less than ash during the relatively wet period, 1999–2002. Data from the Radcliffe Meteorological Station, 5 km away from the site in Oxford, showed that this was the eighth wettest such period since 1767. It is therefore unlikely that the growth of sycamore would have exceeded that of ash under any climatic conditions over the last two centuries at this site. One possible factor contributing to the lower overall growth rates of sycamore is damage by grey squirrel (Sciurus carolinensis), which sycamore is vulnerable to (Mayle et al., 2007). In writing about Wytham, Elton (1966, p. 227) commented that squirrels ‘cause tremendous damage to sycamores by stripping the bark, frequently killing parts or all of the younger trees’. Squirrel damage has continued at the site, despite regular control by shooting and poisoning; we did not, however, identify it as a cause of mortality in our study.
Ash seedlings were more abundant than those of sycamore in both 1993 and 2002. During the period of our study, the site has been subject to grazing pressure from large deer populations (fallow, Dama dama, muntjac, Muntiacus reevsii, and roe, Capreolus capreolus) (Kirby and Thomas, 2000; Morecroft et al., 2001; Perrins and Overall, 2001). Linhart and Whelan (1980) reported that sycamore seedlings were more adversely affected by sheep grazing than those of ash and they may also be more sensitive to deer herbivory. There was no evidence of sycamore seedlings performing better under ash canopies and ash growing better under sycamore canopies, contrasting with the results of Waters and Savill (1992) and Savill et al. (1995), despite the study being carried out on the same site. This may, however, be obscured by the high levels of deer herbivory.
Has tree growth decreased or mortality increased in sycamore trees during dry compared with wet periods?
The proposition that sycamore is drought sensitive is supported by the fact that it grew more during the 1999–2002 wet period than the dry ones. On this basis, it is also clear that ash is less drought sensitive than sycamore; the evidence for oak is more ambiguous. There was, however, no evidence of increased mortality in dry periods. The relatively dry 3-year intervals (1993–1996, 1996–1999, 2002–2005) were in fact slightly wetter than average in the Radcliffe meteorological record. The dry weather that they included, particularly the summers of 1995 and 2003 (Figure 1), may, to some extent, have been offset by the wetter weather in the 3-year periods (by, for example, ensuring high soil–water contents at the start of the summer). In the context of climate change, further work is required on the interactive effects of precipitation at different times of year.
Does tree growth reflect changes in photosynthesis in the canopy during wet and dry periods?
The reduction in sycamore growth during dry conditions is most easily interpreted as a result of reduced photosynthesis resulting from stomatal closure. The gas exchange measurements support this interpretation, demonstrating that sycamore photosynthesis was lower in the drier conditions of 1996 than in 1999 and 2000. Photosynthesis of oak was slightly lower in 1996 than 1999 and 2000 but to a lesser extent than sycamore. The difference between the two species observed by Morecroft and Roberts (1999) in 1996 is therefore likely to be an effect of differing responses to the dry conditions. Sooty bark disease is promoted by hot, dry conditions (Coultherd, 1978; Desprez-Loustau et al., 2006) and was associated with sycamore mortality in the 1976 drought, but no outbreaks were noted in this case.
Is there any interspecific difference in phenological responses to temperature which might affect the outcome of competition in the long term?
The sensitivity of phenology to temperature has been demonstrated for all three species, with similar relationships to temperature. The time series is relatively short and the different species’ sensitivities to different periods in the spring (which may differ in their relative warmth in different years) complicates the interpretation of results. It is, however, unlikely that a lengthening of the growing season would give sycamore a competitive advantage over ash. Not only is there little difference in responsiveness to temperature but also solar radiation increases substantially over the course of the spring. An extension of ash's growing season in late April or early May would have a proportionally bigger impact on total carbon uptake than a similar extension of sycamore's growing season, earlier in the year. The predicted increase in drought frequency would also tend to outweigh the effects of an earlier start to the growing season.
Conclusions and application
These results suggest that sycamore does not currently pose a serious threat to the undisturbed ancient woodland at Wytham and that it is not out-competing ash in the rest of the woods. The evidence is that ash is growing faster and producing more seedlings than sycamore. Sycamore is likely to have been planted at Wytham from the early nineteenth century onwards (Elton, 1966), but many of the present trees have naturally regenerated and this appears to have been favoured by the conditions in the secondary and disturbed woodlands. In contrast, the minimum intervention regime in the undisturbed ancient woodland has presented few opportunities for sycamore to gain a foothold. It is possible that sycamore would eventually decline in those areas where it currently coexists with ash, in the absence of active management.
The resistance of ancient woodlands to sycamore invasion under a minimum intervention regime may not be so clear cut in other situations; in particularly, sycamore is likely to be more of a threat in wetter areas. However, climate change will tend to decrease sycamore growth over much of England, if, as projections suggest, summer droughts increase in frequency. Our results provide empirical support for the projections of Broadmeadow et al. (2005), indicating a decline in the productivity of sycamore with climate change over much of England. Sycamore is therefore likely to be a reduced threat to conservation in future; however, as a timber crop, foresters may find it a less productive species.
Natural Environment Research Council (ECN programme and a studentship to V.J.S.).
Conflict of Interest Statement
We are grateful to the University of Oxford for allowing access to Wytham Woods and to Nigel Fisher, Nick Crawford and Kevin Ewart for their help and cooperation. We thank the many people who have contributed to establishing and maintaining the long-term monitoring, especially Carolyn Walls and Andrew Whitehouse. The late John Roberts gave much useful advice and encouragement, as well as collaborating in making the earlier measurements of gas exchange, together with Rebecca Hopkins. Peter Savill, two referees and the editor made helpful comments on earlier drafts of the paper.