Do bridge construction activities influence birds using the River Mersey, in northwest England?

ABSTRACT

Capsule: Bridge construction works over a major river did not adversely impact the overall wintering and breeding bird assemblage, although contrasting species-specific trends in breeding number were noted.

Aims: To determine whether the wintering water-bird and breeding bird assemblage using the River Mersey, in northwest England, was influenced by bridge construction works.

Methods: Common bird census and wintering bird surveys were carried out before and during construction of the Mersey Gateway bridge. The species abundance and richness of wintering water-birds, assigned to a group (‘Waterfowl’, ‘Wader’, ‘Wetland species’ and ‘Gull’) and breeding birds, assigned to a group (‘Salt marsh species’, ‘Waterfowl species’, ‘Wader species’, ‘Warbler species’ and ‘Other passerine species’), were separately compared before and during the works.

Results: We found little evidence that the bridge construction displaced wintering water-birds and instead found that for most bird groups, their assemblages were higher during the construction works, compared to before. Similarly, we found no evidence that the bridge construction reduced breeding density of any bird group. Contrasting species-specific trends were, however, noted, with some species only recorded breeding during the construction works, and other species recorded in higher breeding densities before the works.

Conclusion: The inadvertent increase in the mud-flats around the bridge structure since the works resulted in more mud-flat being exposed for a greater amount of time, and thus increased the foraging opportunities for many birds. This may at least partly explain why overall wintering bird abundance and richness increased since the works began. Although most breeding birds either increased or remained stable, in number since the works, a small number reduced. This may have been a consequence of the removal of suitable nesting habitat. This study provides a preliminary assessment of the influence of bridge construction on birds.

Human disturbance can influence birds in a number of ways, including displacing birds from optimal habitat and reducing breeding success and/or population size (Carney & Sydeman 1999, Sandvik & Barrett 2001, Burton et al. 2002). Birds being displaced from feeding grounds can potentially have negative impacts on bird survival through increased energy expenditure and reduced food intake, particularly when surviving cold weather (Pfister et al. 1992, Goss-Custard & Verboven 1993, Burton et al. 2002), although the overall impact will depend on factors such as the availability of alternative feeding grounds and the duration of the disturbance (Gill et al. 2001). Impacts on birds will principally be dependent on the species of bird (Blumstein et al. 2005) and type of disturbance, with the influence of recreational human activities on birds typically being the main topic of investigation (Carney & Sydeman 1999, Martín et al. 2015).

Other forms of human disturbance, such as construction works associated with infrastructure developments, have received considerably less attention and most of these have principally considered the impact of works on the population size and distribution of wintering water-birds (Burton et al. 1996, Marsden 2000, Burton et al. 2002). To our knowledge we are not aware of any studies which have specifically assessed the impact of bridge construction activities on birds. Instead a small number, based in North America, have compared wildlife populations (not including water-birds) before and after bridge construction but not during (Vance et al. 2013), and, including water-birds, only after bridge construction (Vance et al. 2012).

It is the widely believed convention that those bird species which require the most protection are those that are most vulnerable to anthropogenic disturbance (Burger 1981, Klein et al. 1995) and are the ones which are most likely to be scared off by human activity. The Mersey Estuary Special Protection Area (SPA) in northwest England (53°20′44.11″N, 2°45′19.48″W) supports internationally important wintering wetland bird populations, including Common Redshank Tringa totanus, Eurasian Teal Anas crecca and Common Shelduck Tadorna tadorna (Joint Nature Conservation Committee 2017). This SPA lies approximately 1.8 km west of the location of a new cable stayed road bridge (Mersey Gateway) which, since May 2014, has been under construction across the River Mersey, linking the towns of Runcorn and Widnes. Therefore any displacement of these internationally important bird populations from suitable foraging habitat has the potential to have adverse implications, particularly given that high levels of foraging are likely to be critical in the winter and that foraging opportunities for these birds to feed on mud-flats is further limited by the tidal cycle.

Our study aims to determine whether there is any evidence that the construction of the Mersey Gateway bridge has displaced over-wintering water-birds adjacent to the Mersey Estuary SPA. This is particularly prudent given that there is a void in the knowledge of the implications of active bridge construction on water-birds. Furthermore, given that there is a variety of habitats (e.g. salt marsh, woodland, habitat boundary vegetation and watercourses) in close proximity to the site of the bridge and these habitats support a diverse breeding bird assemblage, we also aim to assess the potential impact of the bridge construction works on this assemblage.

Methods

The survey area (approximately 370 ha) extended approximately 1.5 km to the east, and approximately 1.4 km to the west of the Mersey Gateway bridge location (53°21′15.86″N, 2°42′76.06″W; Figure 1).

Figure 1. Survey area (marked as a polygon), with Mersey Gateway bridge location shown, between Runcorn and Widnes in northwest England.

Breeding bird survey

Within the survey area, salt marshes, watercourses (Latchford Canal and Astmoor Lagoons), boundary habitats (hedgerows, scrub and treelines) and Wigg Island Nature Reserve were included for the breeding bird surveys. Over a 4-year period (2012, 2013, 2015 and 2016), during each breeding bird season (March–July inclusive), ten surveys were carried out, following the Common Bird Census (CBC) methods (Marchant 1983). Eight of the CBC surveys were undertaken in the morning (all commencing before 09:00 hrs, typically before 07:30 hrs) and two in the early evening until dusk. The surveys were undertaken by an experienced ornithologist, with binoculars, walking along a pre-determined route and recording all birds seen or heard. All bird records were marked onto base maps and bird territories were determined using standard territory mapping techniques, which involved using consistent identification at particular points and cues of territoriality and breeding (Marchant 1983). All surveys were carried out in appropriate weather conditions, avoiding heavy rain and wind. The survey durations were approximately 2.5 and 3 h respectively for the Widnes and Runcorn sides of the River Mersey.

The 2012 and 2013 surveys signify the period when no bridge construction was being carried out. Activities associated with the bridge construction were initiated in May 2014 with works on a temporary haul road across the salt marshes, and a trestle bridge, including piling activity taking place within the estuary. As 2014 signified a period when only preliminary works were being undertaken for the bridge construction, data from this year was omitted from the analyses. During the 2015 and 2016 surveys a full suite of bridge construction activities were proceeding, including works on haul roads and working platforms across the salt marsh, a temporary trestle bridge and coffer dams (including impact and vibration piling), building of the bridge abutments and main road deck, and movements of safety and monitoring boats around the construction works. These works included the use of a wide and varied range of mechanical plant and excavators, haulage trucks, ‘crawler’ cranes, the continuous import of construction material and concrete to the construction site, use of large-scale mobile scaffolding systems, installation and use of three tower cranes (over 100 m high), concrete pours lasting a continuous 36 hours in some cases, and associated noise and displacement activities that may be expected on a large-scale infrastructure civil engineering project. The survey years in this study therefore fall into one of two distinct categories: before construction and during construction.

All data analyses were conducted in R 3.4.0 (R Core Team 2017). Birds were assigned into one of five groups: ‘Salt marsh species’, ‘Waterfowl species’, ‘Wader species’, ‘Warbler species’ and ‘Other passerine species’. For the analyses only those species which have experienced recent declines and thus require the greatest conservation attention (Amber and Red List species; Eaton et al. 2015) are included. The only exception to this, are Green list warblers which are included, given that many of these are migratory and only summer in the UK, and many migratory songbirds are sensitive to population declines, given the nature of their life cycle (Sillett & Holmes 2002).

To test the first hypothesis of whether the bridge construction had an impact on breeding birds, we modelled territory number of each bird group against the period of bridge construction (before or during) using the non-parametric Mann Whitney–Wilcoxon test. For the second hypothesis, we tested whether there were annual differences in the territory number, by modelling territory number, for each bird group, against year, using the non-parametric Kruskal–Wallis test (followed by a Mann Whitney–Wilcoxon post hoc test if a significant result was returned). This second analysis was to determine whether annual fluctuations were more likely to be influencing the bird breeding assemblage than bridge construction works.

Wintering bird survey

Over a 5-year period (‘winters’ of 2011–13 and 2015–16; starting in August 2011) water-bird surveys were conducted in line with the Wetland Bird Survey (WeBS) methodology (British Trust for Ornithology 2017), with surveys carried out once every month, August–March. These involved an experienced ornithologist, with the aid of a telescope, visiting five pre-determined points along the edge of the mud-flats and recording all water-birds using the estuarine habitat during low tide (typically within 3 h of low tide). Water-birds using salt marshes and watercourses (Latchford Canal and Astmoor Lagoons, and the western extreme of St Helens Canal) were also recorded, and some of these habitats (particularly the watercourses) were walked, in order for all water-birds to be detected. Water-bird records were marked onto a base map. Surveys were carried out during periods of suitable weather conditions, with heavy rain and winds avoided. Construction activities during these surveys were the same as those highlighted above. Water-birds were assigned into one of four groups (Table 1): ‘Waterfowl’, ‘Wader’, ‘Wetland species’ (which consisted of other wetland species, such as herons and rails) and ‘Gull’.

Table 1. Recorded wintering water-bird species and their groups used in the analyses. Values in brackets are the mean counts and the range across all survey years. The variation in the counts of different species within each group was considerable (e.g. a peak of one Avocet Recurvirostra avosetta and a peak of 1703 Lapwing Vanellus vanellus, with a mean of 232, in the Wader group) and therefore it is appreciated that some of the more commonly recorded species will have a greater influence on the results of the analyses. However, by grouping together it meant that all species could be considered, whilst ensuring that the quality of the dataset was not undermined by an excess number of zero counts.

Waterfowl	Wader	Wetland species	Gull
Gadwall Anas strepera (7; 0–70)	Avocet Recurvirostra avosetta (1; 0–1)	Coot Fulica atra (28; 0–106)	Black-headed Gull Chroicocephalus ridibundus (661; 150–2235)
Greylag Goose Anser anser (6; 0–52)	Black-tailed Godwit Limosa limosa (6; 0–194)	Cormorant Phalacrocorax carbo (9; 0–26)	Common Gull Larus canus (5; 0–30)
Mallard Anas platyrhynchos (55; 3–164)	Common Snipe Gallinago gallinago (4; 0–55)	Grey Heron Ardea cinerea (2; 0–6)	Glaucous Gull Larus hyperboreus (1; 0–2)
Mute Swan Cygnus olor (17; 0–46)	Common Sandpiper Actitis hypoleucos (1; 0–12)	Great Crested Grebe Podiceps cristatus (1; 0–2)	Great Black-backed Gull Larus marinus (10; 0–53)
Pochard Aythya ferina (3; 0–40)	Curlew Numenius arquata (9; 0–44)	Kingfisher Alcedo atthis (1; 0–1)	Herring Gull Larus argentatus (54; 0–590)
Common Shelduck Tadorna tadorna (35; 0–156)	Dunlin Calidris alpina (27; 0–556)	Little Egret Egretta garzetta (1; 0–10)	Lesser Black-backed Gull Larus fuscus (187; 10–1430)
Shoveler Anas clypeata (14; 0–87)	Golden Plover Pluvialis apricaria (10; 0–360)	Little Grebe Tachybaptus ruficollis (1; 0–16)	Little Gull Hydrocoloeus minutus (1; 0–2)
Eurasian teal Anas crecca (116; 0–842)	Green Sandpiper Tringa ochropus (1; 0–1)	Moorhen Gallinula chloropus (4; 0–14)	Yellow-legged Gull Larus michahellis (1; 0–2)
Tufted Duck Aythya fuligula (15; 0–118)	Jack Snipe Lymnocryptes minimus (1; 0–2)	Water Rail Rallus aquaticus (1; 0–2)
Wigeon Anas penelope (4; 0–90)	Lapwing Vanellus vanellus (232; 0–1703)
	Little Ringed Plover Charadrius dubius (1; 0–14)
	Oystercatcher Haematopus ostralegus (1; 0–5)
	Common Redshank Tringa totanus (17; 0–158)
	Ringed Plover Charadrius hiaticula (3; 0–53)
	Spoonbill Platalea leucorodia (1; 0–2)
	Woodcock Scolopax rusticola (1; 0–1)

Note: The lowest possible mean count given is 1, and the value was rounded up even if the mean was lower than 0.5 for this extremely low bird count.

All data analyses were conducted in R 3.4.0 (R Core Team 2017). To test the first hypothesis of whether the bridge construction had an impact on water-birds, for each group, we modelled the peak and mean species abundance and species richness separately (all fitting a Poisson distribution), using a generalized linear model, as a function of time of bridge construction (‘before’ or ‘during’) and stage of winter period (‘early’, August-November, or ‘late’, December–March). In the analyses ‘during construction’ and ‘early in the winter period’ were both set at 0 (the intercept).

For the second hypothesis, we tested whether there were annual differences in the species abundance and richness of each water-bird group, by modelling species abundance and species richness separately against year, using the non-parametric Kruskal–Wallis test. This was followed, if a significant result was returned, by a post hoc test (the non-parametric Mann Whitney–Wilcoxon test) to determine specifically the years in which species abundance and richness differed. This analysis was to determine whether annual fluctuations were more likely to be influencing the water-bird assemblage than the bridge construction works.

Results

Breeding bird assemblage

We found that the period of the bridge construction works had no significant effect on the territory number of any of the five bird groups (Table 2). Despite this, the territory number was consistently higher during the bridge construction works compared to before the works for all groups, although differing species-specific responses were identified (Table 2). The territory number of the combined breeding bird assemblage was not significantly different between the periods during and before the construction works.

Table 2. Territory number of all breeding bird species before and during bridge construction works, along with the change (%) in number during versus before the works, and the results of the Mann Whitney–Wilcoxon test which tested whether territory number of each of the five bird groups (and all species combined) differed before and during the works. The values in bold signify that the species territory number was lower during the bridge construction works, compared to before the works.

Bird species	Mean territory number before construction (Range is in brackets)	Mean territory number during construction (Range is in brackets)	Change in mean number during versus before construction (%)	Change in peak number during versus before construction (%)	Mann Whitney–Wilcoxon test result of territory number during versus before construction
Salt marsh species
Meadow Pipit Anthus pratensis^A	35 (33–37)	43 (41–45)	18.6	17.8	W = 10.5, P = 0.261
Reed Bunting Emberiza schoeniclus^A	8 (4–12)	16 (15–16)	50	25
Skylark Alauda arvensis^R	38 (35–41)	39 (39–39)	2.6	−5.1
Waterfowl species
Gadwall Anas strepera^A	1 (1–1)	2 (2–2)	50	50	W = 17.5, P = 0.116
Mallard Anas platyrhynchos^A	1 (0–2)	2 (1–2)	50	0
Mute Swan Cygnus olor^A	1 (0–1)	1 (0–1)	0	0
Shoveler Anas clypeata^A	0	1 (1–1)	100	100
Wader species
Lapwing Vanellus vanellus^R	0	1 (0–2)	100	100	W = 5.5, P = 0.536
Common Redshank Tringa totanus ^A	2 (0–3)	2 (1–2)	0	−50
Warbler species
Blackcap Sylvia atricapilla ^G	18 (14–22)	17 (16–17)	−5.9	−29.4	W = 147.5, P = 0.656
Cetti’s Warbler Cettia cetti^G*	1 (0–1)	1 (0–1)	0	0
Chiffchaff Phylloscopus collybita^A	9 (8–9)	9 (9–9)	0	0
Common Whitethroat Sylvia communis^G	9 (6–12)	14 (14–14)	35.7	14.3
Garden Warbler Sylvia borin^G	0	1 (1–1)	100	100
Grasshopper Warbler Locustella naevia^R	0	1 (0–1)	100	100
Reed Warbler Acrocephalus scirpaceus^G	7 (6–7)	9 (8–10)	22.2	30
Sedge Warbler Acrocephalus schoenobaenus^G	12 (11–12)	11 (9–13)	−9.1	7.7
Willow Warbler Phylloscopus trochilus^A	13 (11–14)	6 (3–9)	−116.7	−55.6
Other passerine species
Bullfinch Pyrrhula pyrrhula^A	2 (1–2)	4 (3–5)	50	60	W = 36, P = 0.305
Dunnock Prunella modularis^A	15 (13–16)	19 (17–21)	21.1	23.8
Linnet Linaria cannabina^R	6 (4–7)	4 (2–5)	−50	−40
Mistle Thrush Turdus viscivorus ^R	0	1 (1–1)	100	100
Song Thrush Turdus philomelos^R	3 (0–5)	7 (6–7)	57.1	28.6
All species combined	8 (0–41)	9 (0–45)	11.1	8.8	W = 886, P = 0.177

Note: The superscript after the bird’s name refers to the conservation status (Eaton et al. 2015), with ^G = Green List, ^A= Amber List, ^R = Red List. * signifies Wildlife & Countryside Act (1981, as amended) Schedule 1 species.

With regards to the influence of year on the territory number of the five groups; year did not significantly influence the territory number of any group (Figure 2). For salt marsh and waterfowl species the territory numbers were typically higher during, compared to before, the construction works but the difference was not significant (Kruskal–Wallis test; χ²= 1.846, df = 3, P = 0.605, χ²= 4.59, df = 3, P = 0.204, respectively). For wader species, the territory number was lowest in 2013 (before the construction works), with no evidence of any breeding species in this group, although given that low numbers of breeding wading species were recorded in all survey years (peak of 3 territories in 2012), unsurprisingly the difference between years was not significant (Kruskal–Wallis test; χ²= 2.349, df = 3, P = 0.503). The territory number of warbler species was typically comparable across all survey years and the difference was not significant (Kruskal–Wallis test; χ²= 0.291, df = 3, P = 0.962). For other passerine species the territory number was lowest in 2012, but the difference between the years was not significantly different (Kruskal–Wallis test; χ²= 1.987, df = 3 P = 0.575).

Figure 2. Box and whisker plots of breeding bird territory number for each survey year, for the bird groups. ‘Other passerine species’ are termed ‘Passerine Species’ in the plots. The three horizontal lines in the boxes represent the 25% quartile, the median and the 75% quartile; in some cases one or more of these values are identical and thus fewer than three lines are shown. The whiskers represent minimum and maximum values and are only presented when they differ from the 25% and 75% quartiles respectively.

Wintering bird assemblage

The time of the bridge construction had a significant influence on the abundance and species richness of waterfowl, wader and wetland species, with the number greater during the bridge construction works compared to before (typically P < 0.001 for abundance and P < 0.05 for richness). Gull abundance was significantly greater before the works (P < 0.001), but gull species richness was not influenced by the works (P = 0.247; Table 3).

Table 3. Generalized linear models of wintering water-bird abundance and species richness as a function of time of bridge construction (‘before construction’ and ‘during construction’) and stage of winter period (‘early winter’ and ‘late winter’). For the explanatory variable, time of bridge construction, during construction is set at 0 (the intercept) and for stage of winter period, early winter is set at 0 (the intercept).

Response variable	Time of bridge construction		Stage of winter period
	Parameter estimate ± se	P value	Parameter estimate ± se	P value
Waterfowl peak abundance	−0.036 ± 0.023	0.106	0.143 ± 0.023	<0.001
Waterfowl mean abundance	−0.340 ± 0.030	<0.001	0.250 ± 0.030	<0.001
Waterfowl species richness	−0.353 ± 0.132	0.007	0.174 ± 0.132	0.188
Wader peak abundance	−0.385 ± 0.021	<0.001	0.703 ± 0.022	<0.001
Wader mean abundance	−0.580 ± 0.029	<0.001	0.672 ± 0.029	<0.001
Wader species richness	−0.365 ± 0.164	0.026	−0.067 ± 0.164	0.682
Wetland species peak abundance	−0.151 ± 0.055	0.006	−0.270 ± 0.055	<0.001
Wetland species mean abundance	−0.353 ± 0.074	<0.001	−0.239 ± 0.073	0.001
Wetland species richness	−0.339 ± 0.163	0.037	−0.173 ± 0.163	0.291
Gull peak abundance	0.988 ± 0.015	<0.001	0.112 ± 0.014	<0.001
Gull mean abundance	0.772 ± 0.018	<0.001	−0.038 ± 0.017	0.025
Gull species richness	0.181 ± 0.156	0.247	0.078 ± 0.150	0.601

The stage of winter influenced the abundance of waterfowl, wader and wetland species, with abundances significantly greater in the late winter for all bird groups (typically P < 0.001). The species richness of waterfowl, wader and wetland species was however unaffected by the stage of winter. Gull abundance was influenced by the stage of winter with peak gull abundance higher later in the winter (P < 0.001), and mean gull abundance higher earlier in the winter (P = 0.025). Gull species richness was unaffected by the stage of winter (P = 0.601; Table 3).

With regards to the influence of year on the abundance and species richness on the four groups; year influenced both waterfowl abundance and species richness (Kruskal–Wallis test; χ² = 31.3, df = 4, P < 0.0001 and χ² = 25.7, df = 4, P < 0.0001, respectively; Tables 4 and 5), with abundance and species richness lowest in 2011 and 2012 (before construction works) and typically, significantly lower than in 2013, 2015 and 2016 (2015–16 during construction works; Figure 3, Tables 4 and 5). Wader species richness, but not abundance, was influenced by year (Kruskal–Wallis test; χ²= 10.4, P = 0.034), with the richness in 2011 significantly lower than in 2015 (Mann Whitney–Wilcoxon test; W = 5, P = 0.003; Tables 4 and 5). For wetland species abundance and richness, both were influenced by year (Kruskal–Wallis test; χ² = 16.2, df = 4, P = 0.003 and χ² = 14.8, df = 4, P = 0.005, respectively; Tables 4 and 5), with the values in 2011 consistently lower than in 2013, 2015 and 2016 (Figure 3). Gull abundance and species richness were not influenced by year (Figure 3).

Figure 3. Box and whisker plots showing wintering water-bird abundance and species richness for each survey year, for the four bird groups. The three horizontal lines in the boxes represent the 25% quartile, the median and the 75% quartile; in some cases one or more of these values are identical and thus fewer than three lines are shown. The whiskers represent minimum and maximum values and are only presented when they differ from the 25% and 75% quartiles respectively.

Table 4. Differences in wintering water-bird species abundance and richness in response to year. 2011–13 were before bridge construction began, 2015–16 were during bridge construction.

Response variable	Overall comparison	Post hoc testing
		2011 (A) vs 2012 (B)	2011 (A) vs 2013 (C)	2011 (A) vs 2015 (D)	2011 (A) vs 2016 (E)	2012 (B) vs 2013 (C)	2012 (B) vs 2015 (D)	2012 (B) vs 2016 (E)
Waterfowl abundance	χ^2 = 31.298; P < 0.0001	W = 2901.5; P = 0.253	A < C; W = 2192; P = 0.0002	A < D; W = 2210; P = 0.0003	A < E; W = 1924; P < 0.0001	B < C; W = 2419.5; P = 0.005	B < D; W = 2443; P = 0.007	B < E; W = 2117.5; P = 0.0001
Waterfowl species richness	χ^2 = 25.674; P < 0.0001	A < B; W = 13; P = 0.035	A < C; W = 1; P = 0.0007	A < D; W = 4; P = 0.002	A < E; W = 0; P = 0.0004	B < C; W = 5; P = 0.004	B < D; W = 12.5; P = 0.037	B < E; W = 1; P = 0.0009
Wader abundance	χ^2 = 8.722; P = 0.068	-	-	-	-	-	-	-
Wader species richness	χ^2 = 10.438; P = 0.034	W = 22.5; P = 0.306	W = 18; P = 0.129	A < D; W = 5; P = 0.003	A < E; W = 9; P = 0.013	W = 28; P = 0.667	W = 15.5; P = 0.073	W = 20; P = 0.196
Wetland species abundance	χ^2 = 16.248; P = 0.003	A < B; W = 2166.5; P = 0.039	A < C; W = 1924; P = 0.002	A < D; W = 1964.5; P = 0.004	A < E; W = 1776; P = 0.0002	W = 2352; P = 0.287	W = 2424.5; P = 0.462	W = 2199.5; P = 0.088
Wetland species richness	χ^2 = 14.782; P = 0.005	A < B; W = 9.5; P = 0.014	A < C; W = 6; P = 0.005	A < D; W = 8; P = 0.009	A < E; W = 4.5; P = 0.003	W = 21.5; P = 0.245	W = 17; P = 0.106	W = 16; P = 0.081
Gull abundance	χ^2 = 2.974; P = 0.562	-	-	-	-	-	-	-
Gull species richness	χ^2 = 5.946; P = 0.203	-	-	-	-	-	-	-

Notes: Overall comparisons of abundance and species richness of each bird group were conducted using the non-parametric Kruskal–Wallis test, and if significant (result marked in bold), followed by Mann Whitney–Wilcoxon post hoc test. ‘-’ in the table represents that post hoc tests were not conducted as the Kruskal–Wallis test returned a non-significant result. Mann Whitney–Wilcoxon post hoc test do not automatically correct for multiple comparisons so Bonferroni adjustments were conducted when interpreting significance levels and thus only post hoc tests with P < 0.01 are considered significant (in bold); results that were significant prior to Bonferroni adjustments but not subsequently are in italics. ‘<’ signifies the year in which bird abundance and/or species richness was higher in the output of the Mann Whitney–Wilcoxon tests. For all Kruskal–Wallis tests the degrees of freedom = 4.

Table 5. Differences in wintering water-bird species abundance and richness in response to year. 2013 was before bridge construction began, 2015–16 were during bridge construction.

Response variable	Overall comparison	Post hoc testing
		2013 (C) vs 2015 (D)	2013 (C) vs 2016 (E)	2015 (D) vs 2016 (E)
Waterfowl abundance	χ^2 = 31.298; P < 0.0001	W = 3207; P = 0.982	W = 2948; P = 0.385	W = 2965; P = 0.417
Waterfowl species richness	χ^2 = 25.674; P < 0.0001	W = 35.5; P = 0.704	W = 14.5; P = 0.052	W = 18; P = 0.128
Wader abundance	χ^2 = 8.722; P = 0.068	-	-	-
Wader species richness	χ^2 = 10.438; P = 0.034	W = 22; P = 0.243	W = 26; P = 0.507	W = 37.5; P = 0.533
Wetland species abundance	χ^2 = 16.248; P = 0.003	W = 2726.5; P = 0.562	W = 2435; P = 0.502	W = 2324; P = 0.254
Wetland species richness	χ^2 = 14.782; P = 0.005	W = 26.5; P = 0.550	W = 24; P = 0.387	W = 28.5; P = 0.706
Gull abundance	χ^2 = 2.974; P = 0.562	-	-	-
Gull species richness	χ^2 = 5.946; P = 0.203	-	-	-

Overall comparisons of abundance and species richness of each bird group were conducted using the non-parametric Kruskal–Wallis test, and if significant (result marked in bold), followed by Mann Whitney–Wilcoxon post hoc test. ‘-‘ in the table represents that post hoc tests were not conducted as the Kruskal–Wallis test returned a non-significant result. Mann Whitney–Wilcoxon post hoc test do not automatically correct for multiple comparisons so Bonferroni adjustments were conducted when interpreting significance levels and thus only post hoc tests with P < 0.01 are considered significant. For all Kruskal–Wallis tests the degrees of freedom = 4.

Discussion

Breeding bird assemblage

We found no evidence that the bridge construction activities negatively impacted the breeding assemblage of any of the bird groups, with the territory number of most species actually highest during, compared to before, bridge works. The territory number of the five bird groups was also not significantly affected by year, indicating that breeding bird assemblages surveyed in this study were typically not affected by either bridge construction activities or year.

Although Vance et al. (2013) did not compare breeding bird assemblages during bridge construction in North America, they did find that breeding bird species richness and diversity was largely comparable between post- and pre-bridge construction. Similarly our results support this finding that most species surveyed did not adversely respond to bridge construction and it had no measurable negative impact on breeding birds. In fact, although in small numbers, there were five species that were only recorded as breeding within the survey area during the period of construction works and not prior to the works (three of which are Red List species; Grasshopper Warbler Locustella naevia, Lapwing Vanellus vanellus and Mistle Thrush Turdus viscivorus).

On the contrary, when regarding the mean and peak breeding densities of individual species surveyed, three bird species were consistently recorded in higher breeding densities prior to the construction works compared to during the works (Blackcap Sylvia atricapilla, Linnet Linaria cannabina and Willow Warbler Phylloscopus trochilus). Although the breeding density range for the Red Listed Linnet pre- and during construction had considerable overlap (4–7 and 2–5 pairs respectively), the higher breeding density of the Green Listed Blackcap and Amber Listed Willow Warbler prior to construction works, compared to during the works, was more marked (Blackcap 14–22 and 16–17, and Willow Warbler 11–14 and 3–9, respectively). The Blackcap and Willow Warbler are both summer visitors to the UK (although some Blackcaps now over-winter). Both species nest close to or on the ground, and there is some overlap with their preferred nesting habitat, with Blackcap favouring well-wooded areas with scrub and dense hedgerows, and Willow Warbler favouring woodland edge, more open areas, with scrub and tall hedgerows (Ferguson-Lees et al. 2011). The bridge construction, and temporary works, resulted in the considerable loss of vegetation, including scrub, short sections of hedgerow and salt marsh, and therefore there is evidence that removal of this habitat may have displaced some breeding pairs, particularly Willow Warblers, where the mean breeding density during the construction works, compared to prior to construction, fell by around half. We cannot dismiss the possibility that reduced breeding numbers of these two warbler species was also, at least partially, related to increased noise from the bridge construction works.

Bridge construction can have a series of effects (positive or negative) on wildlife and this will be, at least, partly dependent on the affected species (Vance et al. 2012). In the USA, some bird species, such as Rock Pigeon Columba livia and Cliff Swallow Petrochelidon pyrrhonota, were found to breed at higher densities closer to bridges, whilst some passerine species are negatively impacted, as suitable habitat was lost during construction (Vance et al. 2012, 2013). Our study found that bridge construction activities had no significant effect on the breeding densities of the surveyed bird species overall, yet some species-specific differences in response to the construction activities were notable.

Wintering bird assemblage

There was no evidence that bridge construction negatively impacted the wintering assemblage of waterfowl, wader or wetland species using the survey area. On the contrary, the species abundance and richness of all of these groups was notably higher during the construction works, compared to before the works started. Our results suggest that particularly low counts of species in these groups in 2011 (and also, albeit to a lesser extent, 2012) is the main cause for this trend. A possible explanation could be annual fluctuations, and specifically a reduced wintering water-bird assemblage using the site in 2011 due to untested environmental factors (e.g. winter temperature). A second plausible explanation is that birds were actively encouraged to the survey area by the construction activity. There was evidence that since construction works started, mud-flats around the bridge structures have increased where sediment has gathered as part of dynamic movements of sediment across the area (Adrian Wright, AECOM Principal Coastal Engineer, unpublished data), and some species (such as Common Shelduck) were seen in relatively high numbers feeding in close proximity to the construction works (C. Bonnington & D. Smith, pers. obs.). This increase in the elevation of the mud-flats around the construction area results in less inundation by the tides and exposes the mud-flats for longer periods, increasing foraging opportunities for water-birds.

The only evidence for an unfavourable influence, albeit limited, of the bridge construction works on wintering water-birds is the higher abundance of gulls before the works (particularly in 2013) compared to during the works, suggesting possible displacement. However, this finding could be explained by seasonal variations or a reflection of national gull declines (Eaton et al. 2015). In addition, a landfill site, adjacent to the Mersey estuary and approximately 6 km to the east of the survey area was closed down in January 2017 and this resulted in a dramatic decline in gull numbers on the estuarine mud-flats in this location (P. Oldfield & T. Parker, Halton Environment Round Table, pers. obs.).

Although some studies have documented that construction works, including the construction of a bridge within a large-scale development, displace water-birds (Marsden 2000, Burton et al. 2002), our results typically do not support this. On the contrary, we provide evidence that the assemblage of most water-birds has actually increased during the construction works compared to before they were initiated. The construction works may have inadvertently created suitable foraging sandbanks for many water-birds. Furthermore, our survey area may be sufficiently sized to allow any over-wintering water-birds that were displaced by the construction works to remain within the area, and hence there was limited, if any, decreases in bird abundance or species richness. Indeed we encourage future investigation, building on our preliminary study, to compare the habitat use of birds in relation to construction activity and proximity to bridge works, to allow any localized displacement to be detected.

Conclusion

Our results provide preliminary, novel information into the possible impact of bridge construction activity on breeding and wintering birds. The results should not be used as categorical evidence that overall the breeding and wintering bird communities are unaffected by works, but instead should be used as baseline data, for continued monitoring of the ecological impact of other bridge developments. This is prudent particularly when we consider that other bridge developments will involve potentially different activities and may be impacting different bird species.

The Mersey Gateway bridge development plan included an ecological enhancement strategy within the salt marsh areas to include the creation of scrapes, the use of sluice gates in creeks and conservation grazing by cattle; actions designed to provide suitable habitat for a variety of breeding and wintering bird species. It is predicted that habitat enhancement will ensure that the breeding bird assemblage within the survey area increases compared to the levels from before, and during, construction works. This is especially true for wader species, such as Lapwing and Common Redshank, which have shown increases in breeding density in areas of cattle grazing (Norris et al. 1997) and indeed for Lapwing in a grazing trial within the upper estuary (Smith 2013), although the density of cattle grazing should be managed to prevent trampling of nests by cattle or a reduction in vegetation structure to prevent egg predation (Sharps 2015).

Continued monitoring of the breeding and wintering bird communities using the survey area during the Mersey Gateway’s operational phase, by the Mersey Gateway Environmental Trust, will provide a vital comparison to allow the success of the ecological enhancement measures to be assessed. Furthermore these surveys will enable us to determine whether the trends reported in this study endure beyond the bridge construction. If they do, it can be concluded, like we have here, that the bridge construction has had little, if any, impact on the breeding and wintering bird assemblages present, and that the notable influence on most birds is in fact, perhaps rather surprisingly, a positive one.

Acknowledgements

We thank the Mersey Gateway Crossings Board for use of the data and Merseylink Construction Joint Venture (comprising Kier Construction, FCC and Samsung) for engaging positively with survey results and ecological enhancement programmes. We thank Paul Oldfield, Mark Hampton and an anonymous reviewer for comments on a previous draft, and Marc van Biljon and Emma Clark for producing Figure 1.