Biol Invasions DOI 10.1007/s10530-017-1422-4
INVASION NOTE
Continuing influences of introduced hedgehogs Erinaceus europaeus as a predator of wader (Charadrii) eggs four decades after their release on the Outer Hebrides, Scotland John Calladine . Elizabeth M. Humphreys . Lucy Gilbert . Robert W. Furness . Robert A. Robinson . Robert J. Fuller . Nick A. Littlewood . Robin J. Pakeman . Johanne Ferguson . Charles Thompson
Received: 6 May 2016 / Accepted: 29 March 2017 Ó Springer International Publishing Switzerland 2017
Abstract Non-native predators can cause major declines or even localised extinctions in prey populations across the globe, especially on islands. The removal of non-native predators can, therefore, be a crucial conservation management tool but there can be challenges when they are viewed as charismatic in their own right. Four decades after their introduction to islands in the Outer Hebrides, Scotland, European hedgehogs Erinaceus europaeus continue to be an important nest predator for a declining population of breeding waders. Where hedgehogs were rare, clutch survival rates (assessed using nest temperature loggers) of five species of waders (dunlin Calidris alpina,
J. Calladine (&) E. M. Humphreys British Trust for Ornithology (Scotland), University of Stirling, Stirling FK9 4LA, UK e-mail:
[email protected] L. Gilbert N. A. Littlewood R. J. Pakeman The James Hutton Institute, Craigiebuckler, Aberdeen AB15 8QH, UK R. W. Furness MacArthur Green, 95 South Woodside Road, Glasgow G20 6NT, UK R. A. Robinson R. J. Fuller British Trust for Ornithology, The Nunnery, Thetford, Norfolk IP24 2PU, UK J. Ferguson C. Thompson Scottish Natural Heritage, Stilligarry, South Uist HS8 5RS, UK
lapwing Vanellus vanellus, redshank Tringa totanus, snipe Gallinago gallinago and ringed plover Charadrius hiaticula) were higher than where hedgehogs were relatively more abundant. Hedgehogs were the most frequent nest predator identified using cameras. However, factors influencing population sizes of breeding waders are complex and unlikely to be attributable to a single species of predator. The interactions between predation, land use, habitat and the changes in each deserve further attention. Keywords Machair
Predation Shorebird Non-native
Introduction The now connected islands of South Uist, Benbecula, North Uist and Berneray (hereafter referred to, collectively, as Uist), in the Outer Hebrides archipelago off the west coast of Scotland, hold important concentrations of ground-nesting waders (Charadrii) (Fuller et al. 1986). Particularly important areas are on the west coasts of the islands where extensive machair (a vegetated plain of wind-blown shell sand and a rare habitat type in its own right) is found and grades into the acidic peat-based moorlands of the islands’ interiors (Ritchie 1979; Pakeman et al. 2011). Interactions of land use, hydrology, soil salinity and pH gradients create a fined-grained complex of habitats
123
J. Calladine et al.
and habitat gradients, which, in combination with a restricted suite of predators (compared to most mainland situations), supported breeding dunlin Calidris alpine, ringed plover Charadrius hiaticula, redshank Tringa totanus, lapwing Vanellus vanellus, snipe Gallinago gallinago and oystercatcher Haematopus ostralegus, at a combined average density[90 breeding pairs per km2 (Fuller et al. 1986; Fuller and Jackson 1999). The only mammalian predator native to Uist is the Eurasian otter Lutra lutra, however, European hedgehogs Erinaceus europaeus were introduced to the southern end of South Uist in 1974 (Angus 1993; Jackson 2007). By the mid-1990s, hedgehogs had spread throughout the western coastal areas of South Uist and northwards onto Benbecula and the southern part of North Uist (Jackson et al. 2004). Between 1983 and 2014, there was an 18% decline in the total number of breeding waders (Fuller et al. 2010; Calladine et al. 2015); dunlin and ringed plover populations declined the most, while oystercatcher and redshank populations both increased. Changes tended to be more negative in South Uist and Benbecula, where introduced hedgehogs had become established (Jackson and Green 2000; Jackson 2001; Jackson et al. 2004), than for North Uist where hedgehogs were still rare. Strong evidence, based on a combination of hedgehog exclusion experiments and differential spatial patterns of change in wader populations on Uist between the 1980s and 2000, implicated hedgehog predation of wader eggs as an important factor in population declines (Jackson and Green 2000; Jackson 2001; Jackson et al. 2004). A programme to control hedgehogs had largely removed them from North Uist by the mid-2000s, but was subject to public criticism (e.g. Warwick 2012) with a resulting requirement to further assess the role of hedgehog predation in wader population declines. Although there had been an overall decline in the combined populations of breeding waders in the 40 years since the introduction of hedgehogs to Uist, the assemblage remained at an average of 73 pairs of waders per km2, but its composition had changed considerably (Fuller et al. 2010; Calladine et al. 2015). In this study we: (a) assess how predation rates of wader nests differed between areas with contrasting densities of hedgehogs and whether they have changed since the 1990s; and (b) quantify the levels of nest predation by introduced hedgehogs relative to that by other, mostly native, predators.
123
Methods Two broad study areas for 2012–2014 were selected, one where hedgehogs were relatively abundant (South Uist) and the other where hedgehogs were rare (North Uist; Fig. 1). In 2013, hedgehog presence was recorded, as part of a parallel survey, using a combination of footprint tunnels, camera traps and searches using sniffer dogs. Although hedgehogs were detected in (and removed from) one area of North Uist, encounters by all methods were much lower there than at South Uist (Scottish Natural Heritage, unpublished data). Within each study area, representative ranges of machair and associated habitats were present. To estimate the point at which incubation ceased, temperature loggers (Thermochron ibuttons) were placed in 1222 wader nests. Each logger was programmed to record temperature every 30 min, sealed in plastic, and placed below the eggs within the nests and anchored by wire threads into the soil. Most loggers were retrieved as soon as possible after the expected hatching date. Incubation was considered to have finished when a clear diurnal cycle of temperature variation in the downloaded trace commenced. The outcomes of 281 nests (a subset of the 1222) were also monitored using motion-triggered video cameras (Bushnell ‘trophy cams’), principally to identify nest predators but also to validate outcomes inferred from the temperature loggers. The habitat immediately surrounding each nest was assigned to one of three classes: (1) cultivated machair, which included stubble and recent fallow as well as land under active cultivation; (2) pasture machair, including both semi-natural and sown grassland, dunes and their slacks and saltmarsh; and (3) ‘blackland’, the interface between machair and peatlands, which was entirely pasture including rushes and mires. Habitats and their extent were not plotted or otherwise mapped because of time constraints, rather information on habitat mosaics (for purposes reported here limited to the distance from a nest to a habitat type other than the one in which the nest was located) was derived from a combination of two existing data sets (Morton et al. 2011; Dargie 1998). Clutch survival was estimated separately for North and South Uist for five wader species (dunlin, lapwing, redshank, ringed plover and snipe). The daily mortality rate was calculated by dividing the total number of clutches that were known to have failed by the cumulative total number of days that all nests were monitored (excluding those for which outcomes were
North Uist (low hedgehog density)
Continuing influences of introduced hedgehogs Erinaceus europaeus
South Uist (high hedgehog density)
10 km
Fig. 1 Map of two broad study areas used in 2012–2014. Note that the low hedgehog density area comprises three discrete areas in and off North Uist
uncertain) (Mayfield 1975). From this we derived the overall hatching rate which is the proportion of clutches that survived to hatch at least one chick, assuming a mean incubation period of 19 days for snipe, 21.5 days
for dunlin, 24 days for redshank and ringed plover and 30 days for lapwing (Robinson 2005). We tested for differences in the duration of incubation (from deployment of the loggers) between
123
J. Calladine et al.
the high and low hedgehog density areas, in combination with other covariates, using a binomial generalised linear model (GLM, Aebischer 1999) (all statistical analyses used the program SAS v9.2). In the model, success or failure to hatch (a binomial variable, 0 or 1), with the measured incubation duration as the binomial denominator, was the dependent variable. Explanatory terms used in the GLM were: study area (n = 2, as a proxy of hedgehog density); wader species (n = 5, to account for interspecific differences in the incubation periods); habitat classification (n = 3: cultivated machair, pasture machair and blackland); distance to the nearest neighbouring habitat type (a continuous variable included as a simple measure of habitat heterogeneity, though it should be acknowledged that this may not be representative of how waders perceive habitat heterogeneity); the date of logger deployment (Julian date as a continuous variable); year (n = 3); and whether the nests were monitored using cameras or not (n = 2: Yes, No). The original models also included four interaction terms: (1) habitat 9 site; (2) habitat 9 species; (3) start date 9 site; and (4) camera 9 start date. The measured incubation period would be influenced by the date when the loggers were first placed in nests relative to when eggs were laid. Inclusion of the start date within the models would account for such variation and the start date 9 site interaction terms would suggest if there was any systematic bias in logger deployment. Inclusion of whether or not nest cameras were deployed aimed to assess whether the additional disturbance or marking of nests (with a visible camera) could have had a measurable influence on clutch survival, with the associated interaction recognising that cameras would have been deployed on the earlier nests of target species that were found. Post-hoc pairwise comparisons of model least square means were used to identify differences in recorded incubation durations between areas of contrasting relative abundance of hedgehogs.
Results Interpretable data were retrieved from loggers in 1007 nests. The Mayfield estimates of mean clutch survival were greater in low hedgehog density areas than in high hedgehog density areas (v2 = 16.95, P \ 0.001;
123
Table 1). Mayfield estimates of clutch survival rates provide a useful comparison between areas, however they are subject to bias, and notably nests found later during incubation will tend to have apparently higher survival rates. This will lead to an overestimation of clutch survival rates and potentially confound any recorded differences between areas if there were any systematic biases in laying dates and the timing of nest monitoring. However, measured incubation duration also differed between study areas as well as between species and start date of nest monitoring (Table 1). No apparent differences were identified between years or habitat classifications and none of the interaction terms was statistically significant. Similarly, there was no apparent difference between nests monitored by a camera or not. Post-hoc pair-wise comparisons showed that wader clutches survived for longer on North Uist than on South Uist (Table 1). As expected, shorter incubation periods were recorded at nests where monitoring started later (a significant negative association with start sate); however, an absence of significant interaction terms suggested there was no systematic difference in the dates when loggers were deployed in nests between the two study areas. On North Uist (low hedgehog density), 29% of recorded incubations ended during darkness compared to 45% for South Uist (high hedgehog density) (v2 = 25.8, P \ 0.001), which is consistent with the hypothesis of a greater incidence of nocturnal predation on South Uist. Based on captured video images and direct observation, 50% of nests were confirmed as lost to predation on South Uist compared to 18% on North Uist (v22 = 30.6, P \ 0.001; 8 and 5% respectively were lost through agricultural practices). Hedgehogs were by far the most common predator identified from the nest cameras, with all confirmed observations being on South Uist (Table 2).
Discussion This is the first attempt to determine the relative impacts of nest predation by introduced hedgehogs and other predators on an important breeding wader assemblage of coastal habitats in the Outer Hebrides. Clutch survival was generally lower in the area where hedgehogs were more abundant than where they were rare. Hedgehogs were the most frequently photographed nest predators, even when data were
Continuing influences of introduced hedgehogs Erinaceus europaeus Table 1 The effects of a range of factors on the recorded durations of incubation of wader clutches using nest temperature loggers in two study areas in Uist, Scotland 2012–2014: (a) Estimated clutch survival rates to full term in the two study areas of contrasting hedgehog relative abundances; and (b) a Wader species
summary of the general linear model outputs for the main effects [species; study site (as a proxy for hedgehog relative abundance); start date of nest monitoring; habitat classification] and their interaction terms on the recorded duration of incubation
Clutch survival (mean and 95% CLs), sample size North Uist Hedgehogs rare
South Uist Hedgehogs abundant
Dunlin
0.73 (0.67–0.76), n = 47
0.32 (0.23–0.39), n = 63
Lapwing
0.55 (0.52–0.57), n = 266
0.32 (0.29–0.34), n = 345
Redshank
0.44 (0.35–0.50), n = 77
0.11 (0.05–0.16), n = 80
(a)
Ringed plover
0.37 (0.24–0.47), n = 26
0.26 (0.18–0.33), n = 69
Snipe
0.49 (0.27–0.61), n = 15
0.19 (0.03–0.33), n = 19
Effect
Model v2
df
P
(b) Species
4
5.20
0.27
Study site
1
4.99
0.02
Start date
1
6.76
0.01
Habitat
2
0.18
0.91
Distance (to neighbouring habitat)
1
0.61
0.44
Year
2
0.17
0.68
Camera
1
1.03
0.32
Study site 9 habitat Species 9 habitat
2 8
0.76 4.10
0.68 0.77
Start date 9 study site
1
1.17
0.28
Distance 9 study site
1
1.84
0.17
Camera 9 start date
1
2.05
0.15
Species 9 study site
5
0.76
0.94
Year 9 study site
2
0.05
0.83
Direction of effects on duration of incubation where they were statistically significant Study site: North Uist [ South Uist (by 57 h) Start date: Estimate of slope -0.77 (SE 0.98) Species: dunlin [ redshank (by 93 h); dunlin [ ringed plover (by 64 h); dunlin [ snipe (by 106 h); lapwing [ redshank (by 54 h)
combined across both areas. Avian predators of eggs (mostly gulls) could have been under-sampled by the cameras because video recordings of predation by gulls were usually of birds flying away, while hedgehogs remained at the nests typically for several minutes, allowing ample time for the cameras to trigger. Nonetheless, the greater incidence of nocturnal predation (as suggested by nest temperature loggers) in the high hedgehog density study area is strong circumstantial evidence for the greater
importance of hedgehogs as nest predators. There was no evidence for influences on clutch survival from habitat or from position of a nest within a habitat mosaic, however, that may to some extent reflect the coarse scale at which habitat data were collected or were otherwise available. In the late 1990s, hedgehogs accounted for 55% of nest failures (n = 131 over 3 years) on South Uist outside of exclosures (Jackson 2001). In the present study, we estimated that 21% of all confirmed nest
123
J. Calladine et al. Table 2 The percentage of failed nests attributed to different predators on in 2012–2014 within the two study areas of contrasting hedgehog relative abundance, North Uist (hedgehogs rare) and South Uist (hedgehogs abundant) Predator
All wader nests
Percent of each monitored species
North Uist
South Uist
Lapwing
Hedgehog
0
12
5
21
9
7
3
Otter
0
0.7
\1
\1
0
1
0
Cat
0
0.4
0
0.6
0
0
0
Rat
2.6
0
1
3
2
0
3
Common gull
1.0
3.0
\1
1
2
5
0
Redshank
Dunlin
Ringed Plover
Snipe
Black-headed gull
0
0
\1
\1
0
0
0
Great black-backed gull
0.3
0
\1
\1
0
0
0
Herring gull
0
0.4
\1
\1
0
1
0
Raven
0
0.4
\1
\1
0
0
0
Other/unknown avian
0
1.5
1
1
0
3
3
Sample of camera-monitored nests
n = 98
n = 183
n = 93
n = 56
n = 62
n = 64
n=6
failures on South Uist were due to hedgehogs (58% of all nests failed to hatch, where outcomes could be confirmed, of which 12% were lost to hedgehogs; Table 2). Reasons for lower level of nest predation by hedgehogs in 2012–2014 compared to those recorded in the 1990s could include: (a) changes in hedgehog density or distribution; and (b) a response by the waders to nest preferentially in areas where they are less vulnerable to predation by hedgehogs. The percentage of nests preyed on by hedgehogs in this study was greatest for redshank and dunlin and least for lapwing and ringed plover (Table 2). The differences between species are consistent with those reported for the late 1990s (Jackson and Green 2000), but with the notable difference that our study revealed that levels of predation of ringed plover nests by hedgehogs equalled those of lapwings. Based on low predation rates and observed behaviour of hedgehogs, which tended to avoid areas favoured by nesting ringed plovers, the latter species has previously been assumed to be relatively unaffected by hedgehogs (Jackson and Green 2000). Our study suggests that the difference in nest susceptibility between dunlin and ringed plover may not be as great as previously suggested, or else other (as yet unidentified) conditions have changed to make ringed plovers more susceptible. Calladine et al. (2014) found changes in vegetation appeared to have little effect on breeding densities of most waders, however, changes in ringed plover densities were associated with more of those vegetation changes (positive associations with shade
123
intolerance, soil acidity, fertility and reduced salinity) than other species. Ringed plovers could have become more vulnerable to predation by hedgehogs in response to changing conditions, perhaps by nesting more frequently in areas where hedgehogs now have greater access. The factors influencing population sizes of breeding waders on the Uist machair will be complex and are unlikely to be attributable solely to predation of eggs, especially by a single species (Fuller and Jackson 1999; Calladine et al. 2014). For example, several locations within Uist have never supported hedgehogs but experienced declines in lapwing, ringed plover and dunlin (Fuller and Jackson 1999; Calladine et al. 2014, 2015). Other possible drivers of population change include changing survival rates (of juveniles and adults) resulting from impacts locally or away from the breeding grounds and the impact of predation on chick survival remains poorly understood on Uist. Acknowledgements This study was commissioned by Scottish Natural Heritage and we are grateful to Des Thompson, Ian Bainbridge, Iain MacLeod, David MacLennan and Gwen Evans for their support. We are especially grateful to all the crofters and grazing clerks who facilitated the fieldwork. Fieldwork was managed by Charles Thompson and undertaken by Andrew Airnes, Stephen Bentall, James Bray, Christian Christodoulou-Davies, Terry Fountain, Liza Glesener, Susan Holoran, Colin Kerr, Mairi MacCormick, Andy Roberts, and Yvonne Townsend, with additional nest finding support from other SNH staff in the Argyll & Outer Hebrides Unit, in particular Flora Donald and Patrick Hughes. Further advice was provided by Ruth Mitchell (JHI) and members of Scottish
Continuing influences of introduced hedgehogs Erinaceus europaeus Natural Heritage’s Scientific Advisory Committee and Expert Panel, especially Nicholas Aebischer, Chris Spray and Colin Shedden. The manuscript was improved by the comments of two anonymous reviewers.
References Aebischer NJ (1999) Multi-way comparisons and generalized linear models of nest success: extensions of the Mayfield method. Bird Study 46:22–31 Angus S (1993) Hedgehogs in the Western Isles in 1991. Hebrid Nat 11:73–76 Calladine J, Pakeman RJ, Humphreys E, Huband S, Fuller RJ (2014) Changes in breeding wader assemblages, vegetation and land use within machair environments over three decades. Bird Study 61:287–300 Calladine J, Humphreys EM, Boyle J (2015) Changes in breeding wader populations of the Uist machair between 1983 and 2014. Scott Birds 35:207–215 Dargie T (1998) Sand dune vegetation survey of Scotland: Western Isles. SNH, Redgorton Fuller RJ, Jackson DB (1999) Changes in populations of breeding waders on the machair of North Uist, Scotland, 1983–1998. Wader Study Group Bull 90:47–55 Fuller RJ, Reed TM, Buxton NE, Webb A, Williams TD, Pienkowski MW (1986) Populations of breeding waders Charadrii and their habitats on the crofting lands of the Outer Hebrides, Scotland. Biol Conserv 37:337–361 Fuller RJ, Humphreys EM, Wilson JD, Hoccom DG, Calladine J (2010) Changes in the breeding wader population of the machair of the Western Isles, Scotland, between 2000 and 2007. Bird Study 57:121–124
Jackson DB (2001) Experimental removal of introduced hedgehogs improves wader nest success in the Western Isles, Scotland. J Appl Ecol 38:802–812 Jackson DB (2007) Factors affecting the abundance of introduced hedgehogs (Erinaceus europaeus) on the Hebridean island of South Uist in the absence of natural predators, and the implications for ground-nesting birds. J Zool 271:210–217 Jackson DB, Green RE (2000) The importance of the introduced hedgehog (Erinaceus europaeus) as a predator of the eggs of waders (Charadrii) on the machair in South Uist, Scotland. Biol Conserv 93:333–348 Jackson DB, Fuller RJ, Campbell ST (2004) Long-term population changes among breeding shorebirds in the Outer Hebrides, Scotland, in relation to introduced hedgehogs. Biol Conserv 117:151–166 Mayfield HF (1975) Suggestions for calculating nest success. Wilson Bull 7:456–466 Morton D, Rowland C, Wood C, Meek L, Marston C, Smith G, Wadsworth IC, Simpson IC (2011) Final Report for LCM2007. CEH, Lancaster Pakeman RJ, Huband S, Kriel A, Lewis R (2011) Changes in the management of Scottish machair communities and associated habitats from the 1970s to the present. Scott Geogr J 127:267–387 Ritchie W (1979) Machair development and chronology in the Uists and adjacent islands. Proc R Soc Edinb 77B:107–122 Robinson RA (2005) BirdFacts: profiles of birds occurring in Britain & Ireland. BTO Research Report 407, BTO, Thetford. http://www.bto.org/birdfacts. Accessed on 03 Sept 2014 Warwick H (2012) Uist hedgehogs—lessons learnt in wildlife management. Br Wildl 24:111–116
123
