Modelling large-scale relationships between changes in woodland deer and bird populations



1. There is increasing evidence from local studies carried out in several parts of the world to suggest that increases in the abundance of deer may be depressing population levels of breeding woodland bird populations that are associated with dense understorey habitats. We examine whether habitat modification by deer is likely to be a factor contributing to recent large-scale population declines of woodland birds in lowland England.

2. Novel analytical methods are applied to extensive national bird and deer monitoring data, to examine whether populations of 11 woodland bird species that are associated with dense understorey habitats in lowland England may have been depressed following increases in the abundance of three widespread and abundant deer species: Reeves’ muntjac Muntiacus reevesi, roe deer Capreolus capreolus and fallow deer Dama dama. An additional four woodland bird species that are not specifically associated with understorey habitats are included as controls.

3. For five of the 11 understorey species considered, there is evidence that increases in deer are associated with large-scale depression of abundance or population declines in lowland England. Of these species, we suggest that the impacts of deer are likely to have been greatest for two species of conservation concern, common nightingale Luscinia megarhynchos and willow tit Poecile montanus.

4.Synthesis and applications. Large-scale analyses of extensive national monitoring data provide evidence for a negative association between three widespread and abundant deer species and several woodland bird populations that are associated with dense understorey habitats. These findings are consistent with those from small-scale experimental studies and indicate that deer-related habitat modification may be affecting some bird species on far larger scales than previously appreciated. Mainly through their effects on understorey vegetation, high deer populations are now likely to be affecting woodland biodiversity over large parts of lowland England and deer management plans, involving integrated exclusion and culling of deer, need to be co-ordinated on large scales. It is suggested that such management plans could most usefully target areas that still support relatively high populations of species that are sensitive to deer. Knowledge about the form of relationships between deer abundance and habitat quality for birds and other biodiversity is an important knowledge gap that needs to be addressed if sound, collaborative deer management plans are to be developed.


The ranges and population sizes of deer are increasing in several parts of the world resulting in intensified browsing pressure within many forest, woodland and agricultural landscapes (DeCalesta 1994; Côtéet al. 2004; Dolman et al. 2010). These changes appear to be caused by multiple factors including changes in land use, climate change and the range expansion of non-native deer (Fuller & Gill 2001, 2010; Dolman & Wäber 2008). By altering ecosystem structure and function, large herbivores can act as keystone species. Within woodland, high levels of deer browsing can modify vegetation structure and lead to shifts in plant species composition, potentially releasing ecological cascades affecting a range of trophic levels (Fuller & Gill 2001; Côtéet al. 2004). Evidence recently compiled by Dolman et al. (2010) demonstrates that rapid increases in deer populations in lowland Britain are causing changes in woodland biodiversity especially in ground flora, tree composition, small mammals and birds. The situation in Britain is especially interesting in that the deer populations consist of a mixture of two native and four non-native species, all of which are expanding their range (Dolman et al. 2010; Fuller & Gill 2010). Most evidence of deer-related impacts on biodiversity comes from studies conducted at a local scale. However, Dolman et al. (2010) suggest that impacts are becoming increasingly widespread and should not be viewed as local phenomena.

In England, several woodland bird species have declined markedly in abundance in recent years (Hewson et al. 2007; Baillie et al. 2009) and are now red- or amber-listed ‘Birds of Conservation Concern’ (Eaton et al. 2009). These include lesser spotted woodpecker Dendrocopus minor L., common nightingale Luscinia megarhynchos L., song thrush Turdus philomelos L., willow warbler Phylloscopus trochilus L., willow tit Poecile montanus L. and marsh tit P. palustris L. A number of hypotheses have been proposed to explain these declines (Fuller et al. 2005). Of these, there is growing evidence that reduced regeneration of woody vegetation and understorey density as a consequence of increasing deer populations (Gill & Fuller 2007) may have contributed to the decline of some woodland species that are dependent on dense understorey habitats for nesting or foraging (Gill & Fuller 2007). Supporting this idea, research from North America has demonstrated that changes in understorey vegetation through deer browsing accounts for considerable spatial variability in the abundance and diversity of bird populations (DeGraaf, Healy & Brooks 1991; DeCalesta 1994; Allombert, Gaston & Martin 2005). Experimental work on nightingales from England that examined the effects of deer on fine-scale distribution and habitat use of this species further support the conclusion that an increased deer population is likely to have contributed to declines of this particular species (Holt, Fuller & Dolman 2010). There is also experimental evidence that the abundance of a wider range of migrant birds as well as dunnocks Prunella modularis L. can be reduced by deer browsing (Gill & Fuller 2007; Holt, Fuller & Dolman 2011). Whilst structural modification of vegetation is the main mechanism by which deer are likely to impact on birds, sustained browsing may induce longer-term shifts in tree species composition that could affect habitat quality through altered food availability or vegetation structures (Fuller 2001).

Despite the emerging body of evidence of impacts at local scales, to the best of our knowledge, no large-scale analyses of bird and deer monitoring data has examined relationships between deer browsing on bird populations at a national scale. For conservation purposes, it is essential to assess the evidence of whether processes operate at wider population levels and are not purely local in effect. This paper examines whether changes in abundance of woodland birds potentially sensitive to deer browsing correlate with spatial variations in relative abundance of deer. It specifically addresses the question of whether habitat modification by deer has contributed to recent large-scale population declines of woodland birds.

Materials and methods

Survey methods

This study uses data from the UK Breeding Bird Survey (BBS). The BBS uses a formal sampling framework where 1-km squares are randomly selected according to a stratified sampling design, and each square is surveyed annually. Fieldwork involves two early morning visits to each square where birds together with large and easily detectable mammal species are counted along two 1-km transects. This includes the three most widespread and abundant deer in England, Reeves’ muntjac Muntiacus reevesi Ogilby 1839, roe deer Capreolus capreolus L. and fallow deer Dama dama L. Red deer Cervus elaphus L., sika deer Cervus nippon Temminck 1838 and Chinese water deer Hydropotes inermis Swinhoe 1870 also occur in England but are relatively localised (Dolman et al. 2010) and poorly monitored by the BBS (Wright et al. 2009). Whilst observer counts will underestimate the true number of deer and birds present, there is good evidence for birds (Buckland, Goudie & Borchers 2000) and deer (reviewed in Morellet et al. 2007) that such standardised monitoring produces measures of abundance that correlate well with other methods of intensive density estimates. In this study, we use BBS data from England for the period 1995–2006, during which all three most widespread deer species increased significantly in abundance (Wright et al. 2009).

We consider the impact of the three most widespread and abundant species of deer on 15 bird species in lowland England, exclusively passerines that are associated with woodland habitat (Table S1, Supporting Information). Eleven of these species (dunnock Prunella modularis L., nightingale, song thrush Turdus philomelos L., willow warbler, willow tit, marsh tit, bullfinch, blackcap Sylvia atricapilla L., chiffchaff Phylloscopus collybita L., and blackbird Turdus merula L.) depend on dense understorey for nesting and/or foraging. Four of the species (blue tit Cyanistes caeruleus, nuthatch Sitta europaea, robin Erithacus rubecula and chaffinch Fringilla coelebs) do not depend so strongly, if at all, on the field or shrub layers but are included as control species here. The 11 understorey species were chosen on the basis that they are widely believed to be dependent on dense understorey habitats, are potentially vulnerable to habitat modification by deer browsing (Fuller 2001) and had been recorded on sufficient BBS sites for change in relative abundance to be monitored (Baillie et al. 2009).

Analytical methods

Deer have relatively large home ranges, though this varies between species, so the surveyed sites may not be at the best scale at which to measure the local abundance of these species. In addition, because deer are not very detectable in woodland, the observed rates of change of deer are likely to have many errors associated with them. For these reasons, we used generalised additive models (GAMs) with logarithmic link and quasi-Poisson error structure to remove some of this variability by producing a smoothed relative abundance surface for each deer species across England (after Wood 2006). GAMs have the advantage over GLMs for this purpose in that they incorporate a nonparametric smooth, enabling us to explore spatial patterns of relative abundance which are not constrained to have a specific form. The amount of smoothing was optimised automatically within the model-fitting procedure using generalised cross validation (GCV) scores. A gamma value of 1·4 was included in each model to reduce the possibility of overfitting (Wood 2006). We fitted a model to the data for each year separately, in which easting and northing were expressed as a two-dimensional smooth. Ten landcover variables (Centre for Ecology and Hydrology, CEH aggregate classes; Fuller et al. 2002) were also included as predictor variables. We conducted the analyses in R (R development core team, 2009) and used the ‘mgcv’ package (Wood 2006) to fit the GAMs.

The residuals and Moran’s I values provided no evidence that spatial autocorrelation was important in these models. Examination of the residuals and calculation of the percentage of deviance explained by each model suggested that model fits were good. Additional model evaluation was carried out for each deer species and year by removing 10% of site counts, using the remaining 90% to predict counts for the removed 10%, and repeating for each of 10 nonoverlapping samples of 10%. Correlations between observed and predicted counts were examined to assess predictive ability of the models. Whilst the correlation coefficients for these species were not particularly high, they were considered satisfactory considering the small 1-km scale of the predictions and that we were predicting counts rather than presence/absence (Table S2a, Supporting Information). In addition, these values compare favourably to a comparison of observed values produced using the whole data set (Table S2b, Supporting Information). Comparing the observed values in three pairs of years (1995/1996, 1999/2000 and 2004/2005) reveals a high degree of inherent variability in observed counts (Table S2c, Supporting Information). This allowed us to determine the proportion of the natural variability which the model validation described, which was high (Table S2c, Supporting Information). The fitted models were used to produce year-specific modelled deer abundances across England for each deer species for each surveyed BBS site. These predicted deer abundances were used as predictors in the following analyses.

Changes in the abundance of birds and deer may be driven by different, but concurrent processes, and for this reason, it is easy to misinterpret correlations at a large spatial scale. To examine whether there is a relationship between deer and bird populations, we used an analytical approach developed by Freeman & Newson (2008), which increases the efficiency of data use and the precision of parameter estimates over previous methods for examining the importance of ecological interactions in monitored populations. Freeman & Newson (2008) modelled a change in the abundance of one species as a response variable, in relation to the abundance of another species at the site level. In our case, a change in abundance of a focal bird species in relation to the abundance of roe deer, fallow deer and Reeves’ muntjac. Several biotic and abiotic covariates were also included to control for environmental variability and to provide greater context to our question regarding the influence of deer abundance upon change in bird numbers. To increase the focus of the study to sites where deer may reasonably have an impact on bird populations associated with dense understorey habitats, we restrict the analyses to 1936 woodland and farmland BBS squares in lowland England. This was done by (i) excluding 1-km squares classified as marginal upland or true upland by the CEH Land Classification System (Bunce et al. 1996) and (ii) using CEH Landcover Map 2000 data (Fuller et al. 2002) to exclude any squares with 25% or more urban/industrial cover (landcover type 21), i.e. major urban areas. Considering the mobility and landscape use by deer (Dolman et al. 2010), the impact of deer on bird population change may vary depending on the extent of woodland and farmland in the wider area around a 1-km BBS square. For each BBS square, we calculated the percentage coniferous woodland, broadleaved woodland and farmland (landcover types 14–23) in the wider area within and around 1-km squares of interest (1-km square of interest and neighbouring squares, i.e. nine 1-km squares). The interactions between these three habitat types and each deer species were included as covariates in the following models.

To determine the impact of a single deer species on a bird species, it was necessary to control for the impact that other deer may have on the bird species of interest, by including all three abundant deer species in the model. The impact of deer would be mainly through an indirect effect on the understorey vegetation, which may not have an impact on the bird species immediately. We therefore considered three sets of models that considered either no lag, a 1-year or a 2-year lag for the effect of deer. Because there is a loss of 1 year of data with each year lag, further year lags were not considered. If we were to consider all combinations of lagged and not lagged for each deer species, we would have to fit 33 = 27 models for each bird species, which is clearly impractical and would considerably increase the problem of multiple testing and interpretation. For this reason, we fitted the three sets of models, with the same number of years used in each model and Akaike Information Criterion (AIC) used to identify the best fitting model for subsequent analysis. Model selection was then carried out on the following variables using only the best fitting time-lag.

Weather may influence bird abundance independently of deer, so we included changes in temperature (mean daily minimum temperature) and rainfall (days with rainfall ≥1 mm) during the preceding breeding season (using species-specific breeding seasons, see Table S1, Supporting Information). For resident species, weather variables from the preceding winter (December–February) were also included as covariates in the model. We used spatial monthly weather data provided at a 5-km resolution by the Meteorological Office through the UK Climate Impact Programme matched to the appropriate 1-km square (UKCIP, Multiple deer species were incorporated into the Freeman & Newson’s (2008) model as follows: Suppose that Ni,t is the count of the focal bird species, Pi,t the modelled count of the deer species at site i in year t, Rt is the instantaneous rate of change of the focal bird species population during the period t − 1 to t in the absence of any covariate effect and α is the effect on that rate of change of the deer species. This allows us to separate out the contribution that individual deer species make to change in the bird population of interest α, over and above other potential unidentified drivers of bird population change Rt that are not accounted for by deer and other covariates included in the model. We assume that the observed counts Ni,t have a Poisson distribution and account for large-scale annual changes and the local effect of the deer species. The deer–bird relationship adopted here models the rate of change in the focal bird species abundance as a linear function of the deer abundance, with α as the slope and year-specific intercepts, Rt. Parameter α therefore measures the response in rate of change of the focal bird species per unit of deer abundance; if it is negative, it indicates that the instantaneous rate of change of the focal bird species declines with deer abundance and vice versa. The general model where the effect of deer abundance is not lagged is shown below (a model with 1- or 2-year time-lags is produced by modifying the t subscripts to t − 1 or t − 2 subscripts for the deer terms). This gives us the following model. For illustration, we present the interactions between each habitat variable and each deer variable as a single term in equation 1, rather than specifying each of the 9 habitat*deer interaction terms.

image(eqn 1)

Models were implemented by fitting generalised linear models using the genmod procedure in sas (SAS Institute 2001) with Poisson errors. Nonsignificant weather and habitat*deer interaction terms were removed by stepwise backwards deletion. We accounted for variation between sites by adopting fixed site effects as well as year effects, and treated consecutive observations at a site as independent Poisson-distributed observations. Examination of the residuals suggested that serial correlations were minimal, and model fits were good.

To help interpret the importance of any significant negative relationships identified here, we estimated the impact of deer on national populations of these species by considering the deer coefficients in relation to the change in modelled deer abundance on BBS squares. This was done by rearranging equation 1 to estimate the annual rate of change in the focal bird species at each site as a result of changes in deer abundance as shown in equation 2. Specifically, this equation estimates the average rate of change for one site i in the presence of one deer species from year t to year T.

image(eqn 2)

Excluding BBS squares where deer were never recorded for each of the remaining sites, the modelled deer abundance was inserted into the equation along with the model coefficients describing the effect of deer abundance α, and the Rt’s, describing the effect on annual growth rate in the absence of deer. The equation could then be used to estimate the effect of each deer species P on the population growth rate of each focal bird species at each site, by setting α = 0 to estimate the growth rate with no deer present. An average effect of each deer species on each bird species was then calculated across sites.

We also considered whether any broad pattern emerged across the full suite of deer coefficients by calculating a weighted mean of deer coefficients across bird species for understorey and nonunderstorey specialists separately. This provided a way of evaluating the relative impact of the different deer species.


Maps of predicted relative abundance for the three deer species for the first and last years 1995 and 2006 are shown for illustrative purposes in Fig. 1. These maps clearly indicate range expansion and population growth in all three species of deer over this short period of time. Reeves’ muntjac has spread strongly eastwards from central England to East Anglia where a high-density population has become established. Roe deer have also increased substantially in East Anglia but also show marked increases through much of southern England. Fallow deer are expanding and increasing through much of central, southern and eastern England.

Figure 1.

 Modelled relative abundance of Reeves’ muntjac, roe deer and fallow deer in 1995 and 2006. The legends are the same within species.

Examining relationships between deer and the 11 ‘understorey species’, using the appropriate year lag measure of deer abundance (Table S1, Supporting Information), we found six significant associations at the 5% level (Table 1) from a possible 33. Of these, all six suggested negative associations (Table 1). A negative relationship indicates that the interannual change in the size of the bird population became either less positive or more negative as the size of the deer population increased. Although based on a smaller sample of sites (21 and 49 BBS squares) than the other bird species considered here, the strongest and most highly significant negative associations were between roe deer and nightingale, and between roe deer and willow tit. A smaller sample size will reduce the power to detect a relationship should it exist, and there is always the possibility with a smaller sample size that the sites reporting these species are not representative of the populations of these species as a whole. However, BBS sites are chosen according to a random sampling design, and we have no reason to expect that sites reporting these species are not representative. Whilst the value of α is not completely straightforward to interpret because its units are modelled numbers of deer, for roe deer and nightingale an α of −0·065 corresponds to an approximate 6·3% fall in growth rate for every modelled roe deer, suggesting that this effect is not trivial. Further negative associations were found between roe deer and chiffchaff, between Reeves’ muntjac and both song thrush and willow warbler, and between fallow deer and song thrush. Based on the number of hypothesis tests carried out, even in the absence of a genuine association, one might expect spurious significant results for one or two species at the 5% level, although unlikely to the extent and significance levels here. Following the arguments of Moran (2003), we did not adjust significance level here on account of this multiple testing.

Table 1.  Change in bird abundance in relation to abundance of three deer species according to BBS data 1995–2006. Species scientific names are given in Table S1 (Supporting Information). P-values are: *<0·05, **<0·01, ***<0·001. Significant results are highlighted in bold. Columns are estimated coefficients for the predictors, their standard errors in parenthesis and the magnitude of the effect, Exp(inline image)
SpeciesReeves’ muntjacRoe deerFallow deer
Coefficient inline imageExp(inline image)Coefficient inline imageExp(inline image)Coefficient inline imageExp(inline image)
  1. aWeighted by 1/variance.

Species associated with dense understorey
 Dunnock0·000 (0·012)1·0000·003 (0·002)1·003−0·004 (0·003)0·996
 Nightingale−0·106 (0·160)0·899 −0·065 (0·031)*0·9370·023 (0·085)1·023
 Blackbird−0·006 (0·008)0·994−0·003 (0·002)0·9970·001 (0·001)1·001
 Song Thrush0·033 (0·013)*0·968−0·002 (0·003)0·9980·011 (0·004)**0·989
 Garden Warbler−0·029 (0·021)0·9710·007 (0·009)1·0070·030 (0·010)1·030
 Blackcap−0·004 (0·015)0·996−0·001 (0·005)0·999−0·004 (0·005)0·996
 Chiffchaff0·007 (0·015)1·0070·015 (0·005)**0·985−0·003 (0·005)0·997
 Willow Warbler0·037 (0·012)**0·9640·000 (0·004)1·0000·004 (0·002)1·004
 Bullfinch−0·012 (0·024)0·9880·007 (0·007)1·007−0·010 (0·007)0·990
 Marsh tit−0·003 (0·024)0·9970·004 (0·009)1·004−0·002 (0·004)0·998
 Willow tit0·048 (0·120)1·0490·060 (0·029)*0·942−0·009 (0·028)0·991
 Weighted mean for guild (SE)a−0·011 (0·005)0·989−0·002 (0·001)0·998−0·000 (0·001)1·000
Not specifically associated with understorey
 Blue Tit−0·003 (0·010)0·997 0·005 (0·002)*1·005−0·001 (0·003)0·999
 Nuthatch0·034 (0·026)1·0350·008 (0·009)1·0080·012 (0·011)1·012
 Robin−0·010 (0·009)0·990 0·006 (0·002)*1·0060·000 (0·002)1·000
 Chaffinch−0·022 (0·011)0·9780·003 (0·003)1·003−0·005 (0·004)0·995
 Weighted mean for guild (SE)a−0·006 (0·006)0·9940·005 (0·005)1·005−0·000 (0·002)1·000

Of the four control species that are not believed to be associated with dense understorey habitats, two of the twelve possible correlations between abundance of deer and change in bird populations were significant at the nominal 5% level, and were both positive associations with roe deer. The magnitude of each effect was small, with an α of about 0·006 in both cases, which is equivalent to a <1% increase for every modelled roe deer. It is conceivable that these species benefit from the habitat structures created by deer, either directly or indirectly through reduced competition with some understorey species.

Calculating a weighted mean coefficient across bird species split into understorey specialists and control species for each deer species, produces for understorey species a mean that was negative for Reeves’ muntjac, roe and fallow deer. For fallow deer, the mean negative association is negligible, whilst for roe deer, the association was small (−0·002) and equivalent to a <0·2% reduction in bird population growth rate in relation to each additional modelled deer. For Reeves’ muntjac, the mean association (−0·011) was larger and was equivalent to a 1·1% reduction in bird population growth rate for each additional modelled deer. For the control species, the weighted mean coefficient was negative for Reeves’ muntjac (−0·006), although the standard error associated with this estimate is large, is positive for roe deer (0·005) and close to zero (−0·000) for fallow deer.

To help interpret the importance of the significant negative relationships identified here for nightingale, song thrush, chiffchaff, willow warbler and willow tit, we estimated the impact of deer on national populations on these species by considering the deer coefficients in relation to the change in modelled deer abundance on BBS squares. Because this considers change in deer on all BBS sites where deer are present in one or more years, regardless of the number recorded, these impacts can perhaps be considered as conservative at a national scale (and see discussion below). Local impacts where deer are present or at higher abundance are likely to be more extreme than the large-scale averages. In addition, there is the possibility that multiple deer species may impact on bird populations at a single site. Of 881 BBS squares reporting one or more deer species, 654 (74%) of sites reported one deer species, 190 (22%) reported two deer species and 37 (4%) reported all three deer species. We consider here the percentage impact of Reeves’ muntjac, roe deer and fallow deer on each of these five woodland bird species where a significant relationship was found (Fig. 2). Whilst the exact values should be interpreted with caution, they highlight that the impacts of deer are likely to have been greatest for nightingale and willow tit.

Figure 2.

 Modelled average impact of each deer species on nightingale, song thrush, chiffchaff, willow warbler and willow tit. Impact is expressed as an average percentage decline in each species at a site that can be explained by deer alone and x-axis shows years projected over the 11-year BBS period.


The maps of deer distribution in England generated in this study are unique, as far as we are aware, in that they indicate fine-scale spatial change in abundance, not merely distribution (Fig. 1). The patterns are broadly in agreement with those shown by Aebischer, Davey & Kingdon (2011) at a county scale based on game bag records. For each of the three deer species, they indicate both a clear expansion of range and increasing relative abundance over just a 10-year period. The rapidity with which these changes are occurring is striking, and it is likely that habitat modification as a consequence of browsing will now be occurring over a large part of lowland England. However, there is regional variation in the dynamics of deer populations, with the southern and eastern parts of England being more affected than the northern half of the country. It is worth noting that Hewson et al. (2007) reported considerable regional variation in population changes in woodland birds and that significant population declines were more frequent in the east of England and south-east England than in northern England.

A number of factors are likely to have contributed to the increase of muntjac, roe and fallow deer in England (reviewed by Fuller & Gill 2001; but see also the study by Côtéet al. 2004). The expansion was aided by deliberate and accidental release of captive deer during and since the 19th centuries. However, two particular changes in land management are highly relevant factors. First, a widespread switch from spring-sown to autumn-sown cereals in the mid-20th century (Chamberlain et al. 2000) is likely to have provided an increased winter food source for deer. Secondly, the area of woodland has increased steadily in England during the last 70 years from 755 000 ha in 1947 to 1 130 000 ha by 2010 (Forestry Commission 2010). Furthermore the extensive, mainly pre-Second World War lowland conifer plantations such as Thetford Forest in east England have probably become large sources of deer for colonising the surrounding regions. The hunting of deer is also subject to more controls and management, which has reduced the scale of hunting (Gill 1990), and there are no large predators to provide natural regulation of deer numbers. With a trend towards warmer winter and spring weather, which are correlated with increased recruitment and overwinter survival, it is predicted that climate change will further boost deer populations in the future (Irvine et al. 2007).

We provide statistical evidence that recent increases in common deer species may have contributed, at least in part, to the large-scale population declines of five of eleven bird species considered here that are associated with dense understorey habitat for nesting and/or foraging. These results are largely in agreement with intensive studies carried out at a local level (Gill & Fuller 2007; Holt, Fuller & Dolman 2011). The strongest evidence for a link between deer abundance and population change was for nightingale and willow tit. We acknowledge that because no correlative study can prove causation, it is possible that the relationships obtained may be caused through a correlation with some unmeasured factor. The sample size for nightingale was small, but the effect size, as measured by the coefficient, was relatively large. There is independent experimental evidence that deer browsing can reduce breeding habitat suitability for nightingales (Holt, Fuller & Dolman 2010). This combined with the results of the current analysis strongly suggests that deer are contributing to population decline in nightingales in England. It is unlikely that deer are the sole driver of decline in this species because larger scale factors are probably implicated (Fuller et al. 2005); nonetheless, increasing browsing pressure is probably having a widespread effect on breeding habitat quality for the species. In the case of willow tit, it has been suggested that habitat change as a result of deer browsing could be implicated in the population decline (Fuller et al. 2005), but there is no previous published evidence to support that hypothesis. In a study of paired occupied and abandoned sites, the former were found to have higher soil dampness, but there was no evidence that habitat structure was linked to site occupancy (Lewis et al. 2007). However, this latter study was based on just nine pairs of sites and did not control for habitat and landscape type. The timing of local population responses to habitat change may vary according to landscape context, broad habitat type and availability of microhabitat features. It is clear that there has been a huge decline in the willow tit population in southern England over the last 20 years (BTO/SOC/BWI Atlas 2007–2011, unpublished). Whilst it is unlikely that this can be attributed solely to deer, it remains possible that changes in woodland structure linked with deer browsing is one of several factors reducing habitat quality for willow tits. Other factors may include drying out of woodland (Lewis et al. 2007) and successional change in scrub and woodland habitats (Fuller et al. 2005).

More generally it is interesting that the mean deer coefficient calculated across bird species was most negative for the two most widespread and abundant deer species, Reeves’ muntjac and roe deer. The fact that the relationship was negligible for fallow deer may suggest that the generalised impacts of this species are less than for roe deer and Reeves’ muntjac. It needs to be stressed that where fallow deer do occur, they can have an extreme impact on woodland structure because of their gregarious behaviour (Dolman et al. 2010). There are sufficient differences in the ecology and behaviour of Reeves’ muntjac, roe and fallow deer to expect them to induce different impacts on vegetation structure (Dolman & Wäber 2008; Fuller & Gill 2010). This may be a reason why no individual bird species showed a consistent relationship with all three deer species. There is a need to develop a better understanding of the differential effects of these deer species on vegetation composition and structure and biodiversity.

There are several reasons why the present results should be treated as conservative. Firstly, it is important to highlight that our predictor variables are subject to sampling error, and some are themselves predictions from other models. Such error could lead to attenuation of the estimated coefficients (Carroll et al. 2006), which would underestimate the number and level of significant effects. For this reason, the impact of deer could be greater and could affect more species than suggested here. Secondly, the BBS data are not derived purely from woodland habitats; they are drawn from a large random set of sample 1-km squares across English lowland landscapes. Whilst the bird species examined are mainly associated with woodland, many also occur in a wider range of habitat types such as hedgerows (Fuller et al. 2005). Impacts of deer on woody vegetation structure appear to be most evident within woodland, so one might expect relationships to be strongest between deer abundance and bird declines within woodland habitats. Thirdly, there is large environmental variation associated with analyses such as those presented here that examine relationships of population change with potential causal factors over large spatial areas. This may tend to reduce the chance of detecting clear-cut relationships compared with fine-scale studies that can control for some of this variation. For conservation purposes, it is essential that processes that are trivial or purely local in effect are distinguished from more important processes affecting populations over much larger scales (Gill, Norris & Sutherland 2001). Therefore, assessing the evidence of whether processes operate at a wider population level is essential and we advocate a combination of intensive (local) and extensive (regional) approaches wherever possible. In this context, there may be an argument in favour of relaxing the level of evidence required from extensive studies, in effect acknowledging the difficulty of detecting a signal relating to one specific factor.

In conclusion, we provide evidence that in lowland England, deer may be having a large-scale impact on the populations of a number of woodland bird species that are associated with dense understorey habitats. We are not suggesting that deer are the only, or even the main, factor driving population declines; many other factors are potentially implicated (Fuller et al. 2005). It is also the case that several of the largest declines in woodland bird species (e.g. lesser spotted woodpecker Dendrocopos minor, spotted flycatcher Muscicapa striata) are in species that are not specifically associated with understorey habitats (Fuller et al. 2005). There is still much to learn about the mechanisms underlying the declines of many species of woodland birds and a need for a better understanding of the implications of intensified deer browsing, not just for habitat quality of woodland birds but for wider biodiversity. This research needs to take account of both the effects of different deer densities and different assemblages of deer. It is also likely that deer browsing may interact with woodland management such that effects may be more pronounced in some systems than others (e.g. DeGraaf, Healy & Brooks (1991). Knowledge about the form of relationships – these are unlikely to be linear – between deer abundance and habitat quality for birds and other biodiversity is an important knowledge gap that needs to be addressed if sound deer management plans are to be developed (Fuller & Gill 2001). Deer management aimed at reducing the impacts of deer typically takes the form of excluding deer through the use of various types of fencing and/or culling of deer (the latter is subject to The Deer Act 1991; Dolman et al. 2010). Control strategies are generally advocated at landscape scales, rather than at the scale of individual sites and need to involve multiple partners (Fuller & Gill 2001; Dolman et al. 2010). The findings of this paper serve to emphasise the importance of developing co-ordinated strategies on large spatial scales for minimising deer impacts. With numbers and ranges of deer predicted to expand even further (Irvine et al. 2007), we suggest that such strategies should be targeted on areas that continue to support concentrations of species that are especially vulnerable to modification of habitat structures by deer.


We are extremely grateful to the BBS volunteers. We thank Mark Eaton, Robin Gill, Ian Mitchell, David Noble and James Pearce-Higgins for their comments. The BBS is a Partnership between the BTO, JNCC (on behalf of CCW, NE, CNCC and SNH) and RSPB. Comments from the editors and two anonymous reviewers improved previous versions of this manuscript.