Shoreline Translocation during Road Expansion Was Successful for Most Waterbirds but Not for Waders

Abstract

Coastal wetlands are one of the most threatened ecosystems due to, firstly, their relative rarity and, secondly, the strong human interest in these coastal sites for infrastructure development, and recreation. These coastal wetlands also serve as important migration stopover sites for a range of waterbirds. There is great international interest in mitigating the negative effects of human land development and in restoring degraded habitats. I evaluated the response of one waterbird community to the mitigation action implemented during road expansion in an important waterbird coastal habitat in central Norway. Using a novel mitigation method, the shoreline was moved seawards to allow space for a continued shoreline habitat and the extended road. By weekly monitoring the waterbird community during spring and autumn migration periods (n = 7 of each), I found similar biodiversity, evenness, and abundance of geese, dabbling ducks, diving ducks, gulls, and waders when data for the whole wetland were used. However, after construction waders were partially displaced from the zone closest to the road to a zone further away. For some groups of birds, shoreline translocation can be a fruitful mitigation action aiming to reduce the negative effects of infrastructure expansion in coastal areas.

1. Introduction

Coastal wetlands are often exposed to additional threats compared with inland wetlands, as they are popular sites for industrial activity (including aquaculture, buildings, port facilities, and other infrastructure) and recreation. These wetlands face both pollution and overharvesting of marine food resources that are important for waterbirds [1]. Despite the generally small land area covered by wetlands <5% [2], they provide a large contribution to global ecosystem services [3]. Coastal estuaries are especially important wetland ecosystems for bird populations. Many wader species and other waterbirds (e.g., geese, ducks, herons, gulls, and terns) rely on these ecosystems as feeding sites in winter, stop-over sites during migration, and as breeding sites during summer because these habitats provide considerable food resources and shelter [1]. This dependence makes waterbirds vulnerable to the high global degradation of wetlands evident in the last century [4], as seen by the dramatic decline in shorebird populations around the globe [1,5]. Among the main concerns for shorebird conservation are the effects of anthropogenic loss and degradation [1], since ecosystem deterioration may affect survival and breeding success [6,7,8]. Due to an enduring worldwide deterioration of wetland areas [7,9,10] and a decline in some seashore bird populations [5,11], it is of vital importance to halt the destruction of these areas and/or to restore wetland areas.

To alleviate the effects of lost and degraded habitats, a common conservation measure is to restore, rehabilitate, or create habitat, and because restoration investment into wetlands has a high return value, wetlands are among the most commonly restored ecosystems [12]. In Japan, 23,000 rivers and wetlands were restored during the period 1990–2005 [13]. Restoration is now a worldwide phenomenon [13,14,15], mainly restoration of damaged or lost wetlands [15]. A search with the keyword ‘Wetland restoration’ on ISI Web of Science in August 2024 gave more than 10,000 publications.

Restored ecosystems have higher biodiversity compared with degraded ecosystems and possess a level of biodiversity that is similar to that of natural ecosystems (reviewed in [16]). For example, many bird species respond positively by increasing their numbers rapidly after restoration efforts in coastal wetland ecosystems [17]. However, many created ecosystems are not fully successful because birds or other species use these areas to a lesser extent than the original or surrounding natural sites [18,19,20,21]. For some species, nevertheless, artificial wetlands can be beneficial, compared to natural wetlands [22,23], depending on how these artificial wetlands are constructed [24].

Restoration projects can be very costly. For example, in China, the costs of over 200 programs exceeded USD 20.7 billion [15] and, in the USA, one single project has a budget of up to USD 181 million [25]. Because of the huge costs, there is an urgent need for scientists and managers to exchange information on restoration practices and evaluation techniques [13]. However, the focus should be to avoid the destruction of critical wetland areas or to find solutions that minimize the negative effects of interference in such habitats.

In this paper, I evaluate a novel (to my knowledge) mitigation action for securing an important waterbird habitat and migration stopover site during road construction in central Norway. Due to increased traffic, the European route E6 (hereafter E6) passing the city of Stjørdal needed an upgrade. The Norwegian Public Roads Administration conceived a design guide for the project that details how to reduce damage, primarily on waterbird populations, by moving the shoreline and tidal zone before building E6 on top of the old tidal zone [26]. When the new E6 was built in 2009, the tidal zone (hereafter referred to as shoreline) was translocated seawards, aiming to retain its original shape and size. To evaluate the effect of this mitigation action on the local waterbird populations, all waterbirds were systematically counted on a weekly basis during migration before and after shoreline translocation in four defined zones representing increasing distance from the new road. I compare annual aggregated abundances, weekly abundance patterns, and biodiversity measures before and after shoreline translocation to determine the potential effects of the mitigation. Possible negative effects might be found in all zones but most probably only in the zone closest to the road construction. To evaluate the success of the mitigation project, the distant zones were therefore treated as reference areas together with observed patterns in the national trends in the breeding bird survey (BBS) project.

2. Materials and Methods

2.1. Study Site

The Halsøen study site is in central Norway (63.46° N, 10.91° E). It is bordered by the runway of Trondheim Airport Værnes on the south side, a salt meadow and the European Route E6 on the east side, and a narrow peninsula (Langøra) and the fjord toward the west (Figure 1). The site is located at the former mouth of Stjørdal River and was changed to a bay when a new river mouth was constructed south of the airport. This took place in 1957, when the airport (see Figure 1b) was built, and stopped the water flow through Halsøen [27]. The now-closed-off area became a sheltered, shallow marine wetland area connected to the main fjord by a narrow opening in the northwest (Figure 1b). Before expanding the E6 in 2009–2011, the water area comprised 42.6 ha at low tide and was reduced by 4.5 ha by road construction, whereas the intertidal area was 12.0 ha and did not change due to the construction. The Halsøen seashore was 4.7 km long, of which 3.0 km was natural and 1.7 km was enrockments [28].

In total, 175 different bird species have been observed in and around this wetland, which makes Halsøen an area rich in bird diversity considering its small size and northern latitude. Of these species, 21 of 90 waterbirds are breeding at the site, while an additional 25 species are probably breeding in the surrounding area. However, the area is most important for waterbirds during migration and as a wintering site [29].

2.2. Road Construction and Mitigation Design

The former European route E6 was situated approximately where the easternmost lanes are located today (Figure 1) and was a two-lane road without a central barrier separating the lanes. Owing to increasing traffic, an upgrade of E6 to a four-lane road was necessary. Due to limitations posed by existing infrastructure and buildings, the obvious option was to extend the road toward the sea, with construction of the new road lane following the seashore. During the planning of the road construction, and based on impact assessments, the Norwegian Public Roads Administration conceived a design guide for the project with assistance from the Norwegian University of Science and Technology (NTNU) University Museum (Norwegian Public Roads Administration, 2008). The design was to translocate the seashore before road construction. Important waterbird habitat variables include water depth, water level fluctuations, vegetation, salinity, topography, food, and food accessibility (reviewed in [30]). Furthermore, water depth, water level fluctuations, and salinity are important determinants of food availability and accessibility for various bird species. In an effort to retain the shoreline as similar as possible to the original in terms of slope toward the sea, sediments were removed by a mechanical digger from three predetermined intertidal layers (layer 1: between median low tide to median water level, layer 2: between median water level to median high tide, level 3: above median high tide), 50, 50, and 25 cm deep, respectively (see details in Figure 2 and zonation in Figure 3a). These substrates were stored separately while road construction was initiated and were returned shortly afterward to construct the new shoreline, which was carried out, on average, about 40 m westwards and a maximum of about 70 m west of the new road. Consequently, the intertidal area was similar in shape and size as before, but the water area between the shore and the peninsula providing shelter toward the fjord was reduced by approximately 4.5 ha. By 2010, the new seashore appeared as natural as the original (Figure 3b). Thus, conditions for retaining the natural fauna in the sediments and supporting natural revegetation of the shoreline, both important sources of food and cover for waterbirds, were, hopefully, met [30].

The seashore was only partly reconstructed along the first 400 m north of the airport runway, as old filling compounds covered part of the area (Norwegian Public Roads Administration 2008). The next 800 m, to the northern end of zone 1, was fully reconstructed. The removal and reconstruction of the seashore were carried out in 2009, whereas the road construction continued close by, but outside the directly affected wetland area, and was finished in autumn 2011. Along and close to the new road, a 1–1.5 m high land heap of agricultural soil from another place was constructed to shield the wetland area from the road.

2.3. Bird Surveys

Waterbird numbers in Halsøen in each of the zones 1–4 (Figure 1b) were monitored before road construction, in 1995–1999, and after road construction, in 2010 and 2014. All waterbirds were recorded every week from week 15 to week 23 (early April to early June) and from week 28 to week 41 (early July to early October). These periods were selected because Halsøen is a classic stopover site for migrating birds in spring and autumn, and the number of birds and species is normally low in summer and variable in winter. The winter variation is mainly caused when several hundred common eiders (Somateria mollissima) arrive, mainly in zone 3. In addition, parts of the area can sometimes be covered by ice. Weekly counts over several weeks were chosen because the numbers of birds of each species normally vary greatly during the migration periods, and the peak abundance is not necessarily in the same week every year and differs between species. The time spans between the monitoring periods after road construction were planned to control for immediate and longer-term responses of bird species to the road construction and shoreline translocation. The time span from the first monitoring period to shoreline translocation was not intended but is not problematic because construction zone 1 and reference zones 2–4 were investigated simultaneously.

All individual birds were counted and identified at a species level, using binoculars and telescopes and observing from permanent observation points, with clearly defined borders between the zones providing visual access to the whole study site. When observing adults with chicks, only the adults were counted. However, in late summer and in autumn, young birds without a visual connection to their parents were recorded as independent individuals. The timings of the recordings were standardized to allow them to be initiated at about 1/2–3/4 of full tide (about 2 m between low and high tide), in full daylight, and during good weather with little wind and no or nearly no precipitation.

To assess whether waterbirds responded to the new road and shoreline translocation by changing within-site location preferences, the study site was divided into four zones with different distances to the construction area (Figure 1b). Shoreline translocation only affected zone 1, which also became smaller after translocation, while there were no visible physical changes in reference zones 2–4. The zones were kept constant by well-defined stands, and by sighting against clearly defined points on the opposite site. These stands were not affected by the seashore translocation.

A shallow land area in zones 1 and 2 becomes visible at low tide both before and after the road construction. The location of birds in relation to these zones was recorded for all observations. The fieldwork was performed by four individuals who verified their bird identification skills by passing the online exam at www.birdid.no, accessed on 26 August 2024 [31].

2.4. Data Analysis

To assess the potential impact of shoreline translocation and road construction on waterbirds, I calculated biodiversity indices. In this calculation, I included all observed waterbirds and weekly and annual counts for selected groups of ecologically similar waterbirds, as many species are uncommon at the site, making their data unsuitable for statistical analyses. The selected five groups are geese, dabbling ducks, and diving ducks including mergansers, waders, and gulls. The species included in each group are listed in

Table 1. List of species included in each species group used for analysis, including average total annual bird counts ± standard error (weeks 15–23 and 28–41) before and after mitigation (shoreline translocation) for each species group. Group proportion is the contribution of each species to the species groups. Species marked with * were included in national trend analyses.

Species Group	Species	Scientific Name	Group Proportion
			Before	After
Geese (Before: 28.2 ± 22.4, After: 2003.5 ± 171.5)
	Greylag goose	Anser anser	0.329	0.786
	Pink-footed goose	A. brachyrhynchus	0.654	0.214
	Bean goose	A. fabalis	0.003	0.000
	Canada goose	Branta canadensis	0.014	0.000
Dabbling ducks (Before: 2259.4 ± 327.0, After: 2888.5 ± 390.5)
	Mallard *	Anas platyrhynchos	0.886	0.751
	Eurasian wigeon *	A. penelope	0.037	0.070
	Eurasian teal *	A. crecca	0.075	0.179
	Gadwall	A. strepera	0.000	0.000
	Northern pintail	A. acuta	0.001	0.001
	Northern shoveler	A. clypeata	0.001	0.000
Diving ducks (Before: 182.2 ± 54.0, After: 481.0 ± 35.4)
	Tufted duck *	Aythya fuligula	0.035	0.027
	Greater scaup	Aythya marila	0.006	0.001
	Long-tailed duck	Clangula hyemalis	0.013	0.000
	Common goldeneye *	Bucephala clangula	0.228	0.772
	Common eider	Somateria mollissima	0.537	0.000
	Goosander	Mergus merganser	0.028	0.018
	Red-throated merganser *	Mergus serrator	0.153	0.182
Gulls (Before: 1519.4 ± 90.2, After: 2140.0 ± 309.0)
	European herring gull	Larus argentatus	0.085	0.099
	Greater black-backed gull	L. marinus	0.027	0.033
	Common gull *	L. canus	0.454	0.390
	Lesser black-backed gull	L. fuscus	0.000	0.001
	Black-headed gull	Chroicoce. ridibundus	0.407	0.474
	Common tern	Sterna hirundo	0.018	0.004
	Arctic tern	S. paradisaea	0.009	0.000
Waders (Before: 2523.6 ± 167.4, After: 1443.5 ± 72.5)
	Eurasian oystercatcher *	Haematopus ostralegus	0.145	0.484
	Northern lapwing *	Vanellus vanellus	0.570	0.002
	Dunlin	Calidris alpina	0.062	0.037
	Common redshank *	Tringa totanus	0.125	0.284
	Common greenshank *	T. nebularia	0.013	0.023
	Eurasian curlew *	Numenius arquata	0.007	0.048
	Ruff	Philomachus pugnax	0.019	0.012
	Common ringed plover *	Charadrius hiaticula	0.034	0.084
	Grey plover	Pluvialis squatarola	0.000	0.001
	European golden plover *	P. apricaria	0.012	0.003
	Sanderling	Calidris alba	0.000	0.000
	Red knot	C. canutus	0.004	0.000
	Curlew sandpiper	C. ferruginea	0.001	0.000
	Temminck’s stint	C. temminckii	0.000	0.003
	Little stint	C. minuta	0.002	0.002
	Stilt sandpiper	C. himantopus	0.000	0.000
	Common sandpiper *	Actitis hypoleucos	0.005	0.012
	Broad-billed sandpiper	Limicola falcinellus	0.000	0.000
	Spotted redshank	Tringa erythropus	0.000	0.000
	Wood sandpiper *	T. glareola	0.000	0.000
	Black-tailed godwit	Limosa limosa	0.000	0.000
	Bar-tailed godwit	L. lapponica	0.000	0.001
	Whimbrel *	Numenius phaeopus	0.001	0.002
	Common snipe *	Gallinago gallinago	0.000	0.000

. All analyses were performed using R statistical software, version 3.4.1 [32].

To obtain a general overview of changes in bird numbers and biodiversity measures, I calculated the weekly Shannon diversity index (H) and Shannon equitability index (E_H, assessing evenness) and aggregated annual counts of observed birds in each of the five groups of waterbirds each year. Differences in biodiversity indices and annual bird counts were compared between years before and after moving the shoreline (mitigation action) using linear (Shannon diversity index and Shannon equitability index) and generalized linear models (annual bird counts, Poisson-distributed errors corrected for overdispersion) with mitigation action as the explanatory factor. To test whether birds used the different zones at the study site differently before and after moving the shoreline, I ran generalized linear mixed models, using Poisson-distributed error terms, with annual count as the response variable and the interaction between mitigation action and zone ID as explanatory factors.

To assess whether weekly variation in bird numbers, for each year, differed before and after the mitigation action, I ran generalized additive models (GAM, R package mgcv: [33]) separately for each bird group with scaled (mean-centered and corrected for standard deviation), square root-transformed bird counts as the response. The arrival of large flocks of one species was not observed to influence the number of birds in other bird groups.

All models included an autoregressive autocorrelation structure taking into account dependency between weekly counts, which was allowed to differ between years (corAR1 (form = ~week|year)), following suggestions from Zuur et al. [34]. The random effect of year was added as a smoothed term with base smooth = “re”.

To account for the time lags in the investigations, I compared the abundance patterns of bird groups in the directly affected zone, zone 1, with the other zones at Halsøen (Figure 1b). In addition, I compared these patterns with BBS trends between 1996 and 2017 for the same groups, as was carried out in other similar investigations [35,36]. I used the Norwegian Breeding Bird Survey (BBS) data, which consist of bird counts collected as the sum of pair equivalents (defined as observations of a male, a female, a male and female observed together, or a parent with offspring). The data from 1996 to 2008 are from the first national BBS in Norway [37], while this project was gradually replaced with overlap by a new BBS with data from 2005 to 2017 [38]. The new survey contains data from 20 points each at 493 randomly selected sites across the country (details of data collection and design can be found in, e.g., [39,40]). Not all species had adequate representation in the data for trend analysis, but for most groups, the species contributing the highest number of individuals to the community were included (see Table 1). Annual population indices and standard errors from 1996, or the earliest available count year, to 2017 were extracted from linear trend models for each species using BirdStats v 2.1 and the R package rtrim [41]. All the models were corrected for overdispersion and serial autocorrelation. Composite indices and trends with standard errors were calculated for each bird group using the MSI tool [42]. Species were not weighted relative to each other, as relative population abundances changed over time, but I measured the effect of leaving each species out of the group on the trend estimates. Based on these tests, I found no grounds for excluding any of the species (see

Table 2. National trends for each species group when sequentially excluding one species from the group. Abbreviations: anacre = Anas crecca, anapen = Anas penelope, anapla = Anas platyrhynchos, aytful = Aythya fuligula, buccla = Bucephala clangula, merser = Mergus serrator, acthyp = Actitis hypoleucos, galgal = Gallinago gallinago, haeost = Haematopus ostralegus, numarq = Numenius arquata, numpha = Numenius phaeopus, pluapr = Pluvialis apricaria, trigla = Tringa glareola, trineb = Tringa nebularia, tritot = Tringa totanus, vanvan = Vanellus vanellus, chahia = Charadrius hiaticula.

Species Group	Trend Slope ± SE	Trend Class
Dabbling ducks excluding anacre	1.056 ± 0.014	Moderate increase
Dabbling ducks excluding anapen	0.999 ± 0.010	Stable
Dabbling ducks excluding anapla	1.024 ± 0.011	Moderate increase
Diving ducks excluding aytful	1.014 ± 0.038	Uncertain
Diving ducks excluding buccla	0.991 ± 0.030	Uncertain
Diving ducks excluding merser	0.998 ± 0.042	Uncertain
Waders excluding acthyp	0.984 ± 0.007	Moderate decline
Waders excluding galgal	0.985 ± 0.007	Moderate decline
Waders excluding haeost	0.980 ± 0.007	Moderate decline
Waders excluding numarq	0.983 ± 0.007	Moderate decline
Waders excluding numpha	0.986 ± 0.007	Moderate decline
Waders excluding pluapr	0.985 ± 0.007	Moderate decline
Waders excluding trigla	0.984 ± 0.003	Moderate decline
Waders excluding trineb	0.984 ± 0.006	Moderate decline
Waders excluding tritot	0.985 ± 0.007	Moderate decline
Waders excluding vanvan	0.985 ± 0.007	Moderate decline
Waders excluding chahia	0.978 ± 0.006	Moderate decline

).

I did not calculate national trends for geese, and dunlins and ruffs were left out from the “waders” group, as BBS counts were not available for the whole period. The common gull, representing about 40% of the gulls in the investigated area both in spring and in autumn (Table 1), was the only gull species with available data for gulls, and I present trends obtained from BBS as representative for the species group.

3. Results

3.1. Biodiversity and Evenness in Zone 1–4

Waders, gulls, and dabbling ducks dominated the waterbird community at Halsøen before shoreline translocation, while wader dominance decreased, and geese increased dominance after shoreline translocation (

Table 3. Proportional contribution of each species group to the community composition at Halsøen each monitoring year (1995–1999, 2010, and 2014).

Year	Mitigation	Dabbling Ducks	Diving Ducks	Geese	Gulls	Waders
1995	Before	0.285	0.028	0.017	0.228	0.443
1996	Before	0.342	0.023	0.001	0.241	0.393
1997	Before	0.232	0.059	0.000	0.244	0.465
1998	Before	0.403	0.020	0.002	0.200	0.375
1999	Before	0.445	0.017	0.0001	0.255	0.282
2010	After	0.291	0.050	0.214	0.286	0.160
2014	After	0.351	0.057	0.233	0.196	0.162

). Within each species group except dabbling ducks and gulls, the dominant species was different before shoreline translocation compared with after (Table 1). Pink-footed geese were replaced by greylag geese as the dominant species in the goose group, common eiders were replaced by common goldeneyes in the diving duck group, and northern lapwings were replaced by Eurasian oystercatchers in the wader group.

Average diversity measures (Shannon’s H) were slightly higher after shoreline translocation (H = 1.77 ± 0.07) than before (H = 1.69 ± 0.03), while the linear models indicated that the difference in weekly diversity values was not significant (β = −0.08 ± 0.07, t = −1.176, p = 0.241, Figure 4a). The calculated evenness (Shannon’s equitability index, E_H) remained the same before and after shoreline translocation (before: E_H = 0.66 ± 0.01; after: E_H = 0.66 ± 0.03, β = −0.004 ± 0.02, t = 0.182, p = 0.856; Figure 4b). Among the species groups, dabbling ducks and diving ducks excluding eiders showed increased diversity after shoreline translocation and were the only groups with significant differences in diversity before and after road construction and mitigation (

Table 4. Estimated difference in Shannon–Wiener’s diversity index (H) and Shannon’s equitability index (J) and corresponding statistics for each species group before and after shoreline translocation. Positive values indicate an increase in diversity and equitability (evenness) after mitigation compared with before. Significant differences are highlighted with bold text; trends are highlighted with bold italic text.

	Species Group	Estimated Difference Before/After	Statistic	Significance Level
H
	Dabbling ducks	0.183 ± 0.050	F1,159 = 13.410	p < 0.001
	Diving ducks	0.061 ± 0.060	F1,159 = 1.018	p = 0.315
	Diving ducks excl eiders	0.161 ± 0.057	F1,159 = 8.113	p = 0.005
	Geese	0.015 ± 0.020	F1,136 = 0.505	p = 0.479
	Gulls	0.046 ± 0.049	F1,159 = 0.905	p = 0.343
	Waders	0.060 ± 0.079	F1,159 = 0.581	p = 0.447
	Waders excl lapwings	0.072 ± 0.052	F1,159 = 0.218	p = 0.641
J
	Dabbling ducks	0.072 ± 0.052	F1,109 = 1.898	p = 0.171
	Diving ducks	0.026 ± 0.077	F1,99 = 0.117	p = 0.773
	Diving ducks excl eiders	0.140 ± 0.089	F1,88 = 2.480	p = 0.119
	Geese	0.037 ± 0.031	F1,122 = 1.430	p = 0.234
	Gulls	0.042 ± 0.033	F1,159 = 1.672	p = 0.198
	Waders	0.073 ± 0.041	F1,152 = 3.112	p = 0.080
	Waders excl lapwings	0.043 ± 0.093	F1,141 = 0.178	p = 0.674

). Eiders were excluded from some calculations because they were sometimes observed in flocks of several hundred individuals, and these flocks could be observed in the opening toward the sea in zone 3 (Figure 1b). Evenness did not differ before and after shoreline translocation, although waders tended toward higher evenness after shoreline translocation (Table 4).

3.2. Bird Abundances in Zone 1–4

The annual bird counts for waterbirds in Halsøen were significantly higher after shoreline translocation compared with before (Figure 5, Table 1) for diving ducks, geese, and gulls (

Table 5. Model results for differences in annual counts in Halsøen before and after shoreline translocation for dabbling ducks, diving ducks, geese, gulls, and waders (including and excluding northern lapwings). Groups with significant differences are highlighted in bold text.

	Scaled Difference Before/After ± SE	t-Value	p-Value
Dabbling ducks	0.246 ± 0.235	1.047	0.343
Diving ducks	0.971 ± 0.253	3.834	0.012
Diving ducks excluding eiders	1.574 ± 0.196	8.034	0.001
Geese	4.263 ± 0.751	5.677	0.002
Gulls	0.342 ± 0.120	2.860	0.035
Waders	−0.559 ± 0.140	−3.999	0.010
Waders excluding northern lapwings	0.307 ± 0.146	2.110	0.089

). When eiders were excluded from the “diving ducks” group, the difference in annual counts before and after mitigation was even more pronounced (Table 5). Waders, in general, showed an opposite pattern, with higher counts before shoreline translocation than after, but this pattern was mainly due to the drastic reductions in the abundance of northern lapwings. When lapwings were removed from the data, there were no differences in annual counts of waders before and after mitigation (Figure 5, Table 5).

3.3. Bird’s Use of the Different Zones

Across all years and all species, most birds were observed in zones 1 and 3, and their proportions did not depend on shoreline translocation (Figure 6). However, the zone distribution and change in use depended on the species group (Figure 7,

Table 6. Model estimates ± SE and statistics for the distribution of individuals between the four zones 1–4 indicating increasing distance from the road before and after shoreline translocation. Model estimates are provided for the main effects of mitigation, zones, and their interaction for each species group. Significant differences are highlighted in bold text.

Group	Variable	Estimate ± SE	z-Value	p-Value
Dabbling ducks
	Mitigation (difference before/after)	0.401 ± 0.022	527.218	<0.001
	Zone 2 (difference from zone 1)	−1.184 ± 0.028	18.385	<0.001
	Zone 3 (difference from zone 1)	−1.357 ± 0.029	−43.006	<0.001
	Zone 4 (difference from zone 1)	−0.814 ± 0.024	−46.064	<0.001
	Mitigation: 2	−0.034 ± 0.045	−33.850	<0.001
	Mitigation: 3	−0.423 ± 0.054	−0.738	0.460
	Mitigation: 4	−0.856 ± 0.050	−7.824	<0.001
Diving ducks
	Mitigation (difference before/after)	1.463 ± 0.076	19.161	<0.001
	Zone 2 (difference from zone 1)	−1.017 ± 0.118	−8.629	<0.001
	Zone 3 (difference from zone 1)	0.500 ± 0.077	6.500	<0.001
	Zone 4 (difference from zone 1)	−1.048 ± 0.119	−8.792	<0.001
	Mitigation: 2	0.230 ± 0.144	1.597	0.110
	Mitigation: 3	−1.260 ± 0.112	−11.210	<0.001
	Mitigation: 4	−2.009 ± 0.249	−8.082	<0.001
Geese	Mitigation (difference before/after)	6.595 ± 0.578	11.404	<0.001
	Zone 2 (difference from zone 1)	−13.792 ± 346.047	−0.04	0.968
	Zone 3 (difference from zone 1)	3.060 ± 0.591	5.181	<0.001
	Zone 4 (difference from zone 1)	3.206 ± 0.589	5.443	<0.001
	Mitigation: 2	13.306 ± 346.047	0.038	0.969
	Mitigation: 3	−2.117 ± 0.592	−3.576	<0.001
	Mitigation: 4	−4.7682 ± 0.595	−8.021	<0.001
Gulls	Mitigation (difference before/after)	0.100 ± 0.033	3.074	0.002
	Zone 2 (difference from zone 1)	−1.516 ± 0.042	−35.748	<0.001
	Zone 3 (difference from zone 1)	−0.566 ± 0.030	−18.933	<0.001
	Zone 4 (difference from zone 1)	−0.396 ± 0.028	−13.961	<0.001
	Mitigation: 2	0.591 ± 0.066	8.925	<0.001
	Mitigation: 3	0.383 ± 0.050	7.640	<0.001
	Mitigation: 4	0.262 ± 0.049	5.382	<0.001
Waders
	Mitigation (difference before/after)	−1.071 ± 0.040	−26.854	<0.001
	Zone 2 (difference from zone 1)	−0.219 ± 0.021	−10.547	<0.001
	Zone 3 (difference from zone 1)	−1.364 ± 0.031	−44.453	<0.001
	Zone 4 (difference from zone 1)	−1.022 ± 0.027	−37.959	<0.001
	Mitigation: 2	0.139 ± 0.058	2.399	0.016
	Mitigation: 3	1.571 ± 0.059	26.635	<0.001
	Mitigation: 4	0.797 ± 0.062	12.799	<0.001
Waders excluding northern lapwings
	Mitigation (difference before/after)	−0.298 ± 0.043	−6.982	<0.001
	Zone 2 (difference from zone 1)	−0.756 ± 0.036	−20.905	<0.001
	Zone 3 (difference from zone 1)	−1.373 ± 0.045	−30.206	<0.001
	Zone 4 (difference from zone 1)	−0.715 ± 0.036	−20.048	<0.001
	Mitigation: 2	0.680 ± 0.065	10.450	<0.001
	Mitigation: 3	1.585 ± 0.068	23.343	<0.001
	Mitigation: 4	0.495 ± 0.067	7.444	<0.001

). Diving ducks and geese increased their use of zone 1 and decreased their use of zone 3. Waders, including and excluding northern lapwings, were the only species group decreasing its use of zone 1 and increasing its use of zone 3. Other species groups showed no strong changes in the proportional use of zones.

3.4. Seasonal Timing of Bird Abundance Changes in Zone 1–4

The weekly bird counts provide a better resolution on the timing of abundance increase and decrease after shoreline translocation (Figure 4). The models for most groups show that weekly trends differed before and after shoreline translocation compared with a common trend for all years, except for gulls, where a common trend fitted the data best (

Table 7. Analysis of deviance table for the effect of mitigation (before/after shoreline translocation) on weekly trends in bird counts for each species group by comparing a model where weekly trends in counts were fitted with different trend lines before and after mitigation with a model where weekly trends were common before and after mitigation. Difference in deviance and degrees of freedom for the two models are reported along with the corresponding F- and p-values. Significant differences are highlighted in bold text.

	Deviance	Df	F-Value	p-Value
Dabbling ducks	−4.492	−3.761	3.680	0.008
Diving ducks	−9.931	−4.161	2.719	0.030
Geese	−37.613	−1.264	45.245	<0.001
Gulls	−5.022	−3.987	1.729	0.147
Waders	−34.163	−8.433	6.860	<0.001
Waders excluding northern lapwings	−7.961	−3.476	3.752	0.009

, Figure 8). Dabbling ducks showed slightly higher counts after mitigation only during the spring weeks (Figure 8a), whereas diving ducks showed slightly higher counts in early autumn (Figure 8b). Weekly changes in geese and wader numbers responded strongest to mitigation, with a high increase in geese during autumn migration (Figure 8c), whereas wader counts decreased during autumn (Figure 8e). However, the autumn decrease in waders is mainly due to a decrease in northern lapwing abundances, since the autumn difference disappeared when excluding this species (Figure 8f).

3.5. National Trends in Species Groups

Of the three species groups for which I had reliable national count records (dabbling ducks, diving ducks, waders), dabbling ducks showed a moderate increase in population indices (Figure 9a and Figure 10a,

Table 8. Overall national trends (1996–2017) for species groups and trends for the last 6 years (6y, 2012–2017) with corresponding trend classes (following classification from [42]).

Species Group	Overall Trend ± SE	Trend Class	Trend 6y ± SE	Trend Class
Dabbling ducks	1.026 ± 0.010	Moderate increase	1.034 ± 0.082	Uncertain
Diving ducks	1.004 ± 0.032	Uncertain	0.884 ± 0.167	Uncertain
Waders	0.984 ± 0.006	Moderate decline	1.010 ± 0.055	Uncertain
Gulls (L. canus)	0.957 ± 0.006	Moderate decline	1.033 ± 0.057	Uncertain

). Visually, the population trends in diving ducks (excluding eiders) seemed stable (Figure 9b and Figure 10b) but were defined as uncertain (Table 8). The population trends in waders showed a general decline (Figure 9c and Figure 11); the same result was observed when northern lapwings were removed from the data (see Table 2). I only had data on one gull species, the common gull, which also showed a moderate decline from 1996 to 2017 (Table 8). All the species groups and common gulls showed uncertain trends over the past 6 years (since 2012, Table 8).

4. Discussion

Based on before/after comparisons of bird counts and species richness measures, I evaluated whether the mitigation action of translocating the tidal zone at Halsøen seawards to account for an extension of the new European Route E6 negatively affected the waterbird populations at the site.

4.1. Mitigation Impact on Biodiversity and Evenness in Zone 1–4

The biodiversity measures (Shannon’s H) were slightly higher after shoreline translocation compared to before. The evenness in the whole area (Shannon’s equitability index, E_H) remained similar after shoreline translocation. The dominating species within each species group changed for some groups. These changes for the whole area indicate that seashore translocation was mainly successful. For a few species, the changes in the study area followed national trends, for example for the northern lapwing.

4.2. Mitigation Impact on Bird Abundance in Zone 1–4

Most species groups showed an increase in annual counts and minor shifts in weekly counts (except for geese). Geese counts at Halsøen were considerably higher after road construction than they were before and were mainly related to highly increased autumn migration numbers based on the weekly counts (Figure 8c). This is in accordance with the documented increase in population sizes of large herbivorous migratory waterbirds such as the pink-footed goose and the greylag goose in Europe [43]; these were the two most common goose species at Halsøen (Table 1). Pink-footed geese dominated the goose group before mitigation in the late 1990s, corresponding to increased use of central Norwegian stopover sites during migration [44]. Greylag geese dominated the goose group after mitigation. Because most Norwegian-breeding greylag geese migrate in August, whereas greylag geese breeding in northern Norway migrate in September [45], the latter populations probably account for the peak count numbers observed at Halsøen, provided they stop over during migration.

The condition of bird communities is increasingly used to evaluate the success and failure of habitat restoration and mitigation efforts because birds respond rapidly to a collective set of underlying features [46]. Such projects are not necessarily successful, for example, if creation, recreation, or mitigation leads to less frequent use by resident/migrating birds than the original or surrounding natural sites [18,19,20,21]. It is, therefore, encouraging that I found, in this study, clearly higher annual aggregated counts of diving ducks, geese, and gulls and no difference in the counts of dabbling ducks after shoreline translocation and road extension up until five years after habitat mitigation.

However, waders showed more differences before and after mitigation, showing lower counts after mitigation than before and showing a strong decrease in autumn counts. Waders was the only species group with reduced counts after shoreline translocation compared with before. The declining results in annual and weekly counts are mainly due to the decreased abundance of northern lapwings (Figure 5). Northern lapwings have experienced considerably reduced population abundance across Europe [47]. The number of lapwings, in particular, decreased dramatically in Norway during the investigation period, and they became nearly extinct in many areas [48,49,50]. This reduction is mainly due to changes in agricultural practices not related to road construction at Halsøen. The strong decline in the autumn numbers of waders (Figure 8e) disappears when excluding northern lapwings from the group (Figure 8f), whereas the negative national trends in waders (Figure 11) did not depend on the presence of this species in the analyses (Figure 11c). The decline in composite trends in waders supports existing species-specific trends in several publications including waders in Fennoscandia [39,51,52]. These patterns suggest that many wader species still use Halsøen more than expected based on national trends, despite some of the species being especially wary of the presence of roads [53].

The weekly counts indicated only minor changes in time-dependent abundances for both duck groups, indicating increased spring abundances of dabbling ducks and a shift toward higher relative abundance in autumn compared with spring in diving ducks. Consequently, the higher counts of dabbling ducks were expected considering the moderate increase observed in national trends (Figure 9a and Figure 10a). Similarly, the lack of difference in counts of diving ducks was expected based on the stable national trends found for this species group (Figure 9b and Figure 10b). Furthermore, the abundance of ducks did not decline despite the reduced size of the water surface area between the shoreline and Langøra peninsula providing shelter toward the fjord.

Changes in annual and weekly counts of waders corresponded well with the available national trends. National trends for waders were negative even when excluding northern lapwings from the analysis, suggesting that wader abundances (excluding those of northern lapwings) were not negatively affected by the road extension and shoreline translocation in all zones combined.

4.3. Mitigation Impact on Bird’s Use of the Different Zones

Most species groups showed an increased or no change in the use of the area closest to the road. Geese, diving ducks, geese, and gulls used zone 1 closest to the road more frequently or equally often after road construction compared with before, suggesting that they were not negatively affected either by the mitigation action or increased traffic or traffic noise.

However, waders showed more differences before and after mitigation, using the area closest to the road (zone 1) less frequently. The seashore translocation did not especially affect the areas used by lapwings. The reduced use of zone 1 indicates that waders may be vulnerable to the new road and/or increased traffic or that the new seashore offered less food for this group. The sensitivity to road presence seems to be alleviated by reducing the use of zone 1 and increasing the use of zone 3 further away from the road, but less food available in zone 1 after translocation might also be the reason why waders became less abundant in zone 1 (Figure 7). Normally, land reclamation reduces the number of waders in an area [54]. Another reason for the reduction of waders in zone 1 might be reduced overview of possible attacking raptors behind the land heap between the road and zone 1.

Birds displaced from one wetland area might aggregate in other, less-disturbed areas [7]. This might explain why the negative effects on waders in all zones were not very pronounced compared with the national trend even though the use of zone 1 was significantly lower after seashore translocation. On the wintering ground, it was shown that redshanks displaced from one wetland area moved into another area and met competition from the residents in the new area. The displaced birds became lighter and suffered a 44% increase in mortality rate, a change likely to substantially reduce the local population size [6]. Such effects are challenging to measure on migration stopover sites, but, because the wader population has declined, relatively more waders might be found in zone 3 without reducing the carrying capacity in that zone.

4.4. Study Limitations

The main limitation of this study concerns the timing of the before monitoring relative to road expansion and shoreline translocation. The before monitoring was performed about a decade before road construction and shoreline translocation. Therefore, I do not have detailed information on bird counts immediately prior to road construction and shoreline translocation. However, the reference areas in the same area (zones 2–4) were counted at the same time as the construction area, and the changes in bird numbers for the different species groups considered are in line with the national trends calculated for those groups, suggesting that the before–after translocation patterns I observed are valid. Furthermore, the close proximity to the road and the airport, both of which have experienced increased traffic during the study period, makes local population increases prior to shoreline translocation unlikely as both road and airport traffic can have substantial effects on local birds [55,56]. The spatial response of waders by moving from zone 1, with the translocated shoreline, and using zone 3, which is further away from both the road and the airport’s landing field, supports this notion.

4.5. Implications for Habitat Management

Several features of the mitigation design may have contributed to the apparent success, for most waterbird groups, of shoreline translocation caused by the road expansion. Special emphasis was placed on retaining not only the original slope profile of the shoreline, which is important for maintaining access to food for waterbirds with different water levels [30], but also on reusing the original substrate layers from the original shoreline, with the same relative positioning. This probably enhanced the probability that the invertebrate fauna in the sediments recovered readily and would probably be useful in similar mitigation actions elsewhere. Natural revegetation of restored salt marshes seems to improve ecosystem functionality better compared with restoration projects where revegetation is necessary [57]. The availability of food, including both vegetation and invertebrates, and accessibility are also important features determining habitat quality for waterbirds (reviewed in [30]). As birds are excellent indicators of environmental change [58,59,60,61,62,63,64], I used the changes in the bird populations as indicators of the quality of vegetation, invertebrates, and water without investigating these variables. If more detailed information about changes in bird populations after mitigation action is needed, I recommend more detailed investigations on both biotic and abiotic factors.

Although they may be successful otherwise, restored and created habitats and/or ecosystems improve conditions compared with degraded habitats/ecosystems [16] but often fail to provide similar numbers of species or individuals as undisturbed ecosystems [65] or fail to improve conditions for species with high conservation value [66]. Such studies are often based on monitoring efforts covering 10–20 years, and meta-analyses suggest that although animal populations recover rapidly after restoration or mitigation, ecosystem functionality may need decades to recover [2,16]. Monitoring until five years after the mitigation action at Halsøen suggests promising prospects for resident and migrating waterbirds, with increasing or stable bird counts for most groups investigated.

A main contributing factor promoting a successful result of the mitigation action includes the sheltering effect of Langøra peninsula, which protects the shoreline at Halsøen from most of the wave actions in the fjord. This would reduce the negative impact of wave erosion during establishment, and I would not recommend carrying out mitigation actions in exposed areas in the same way as they were performed here.

5. Conclusions

In light of the huge costs connected to wetland restoration and the positive results of this novel method with seashore translocation, I recommend this mitigation action in other similar projects. I found only minimal changes in biodiversity, evenness, and abundance of dabbling ducks, diving ducks, gulls, and waders when data for the whole wetland was used.

However, after construction, waders were partially displaced from the zone closest to the road to a zone further away. It is important to divide the investigated area into separate zones when the effects are investigated. Halsøen is a small area, and the conclusion for the whole area was that there were no negative effects on the waterbirds, not even on waders. However, this conclusion is wrong. By dividing the wetland area into different zones, it was apparent that waders used the zone closest to the road and seashore translocation to a lesser extent than before. Their increased use of a more distant zone masks the real effect of the mitigation action found if only data for the whole area were included.

Despite the current mitigation to retain the important waterbird habitat at Halsøen, the site, like many similar shoreline sites along the Norwegian coast, is still subject to high anthropogenic disturbances, from road traffic, air traffic, industry, and human settlements. Collectively, these factors still pose threats toward the continued suitability of Halsøen for migrating and breeding waterbirds. Waterbirds contribute substantially to ecosystem services (ecosystem processes that directly or indirectly benefit human well-being) [67]; thus, human destruction and disturbance in important waterbird areas [5,9,68,69,70,71] should stop immediately.