Grazing management can counteract the impacts of climate change-induced sea level rise on salt marsh-dependent waterbirds

Authors


Correspondence author. E-mail: kc@dmu.dk

Summary

  1. Climate change–induced rises in sea level threaten to drastically reduce the areal extent of important salt marsh habitats for large numbers of waterfowl and waders. Furthermore, recent changes in management practice have rendered existent salt marshes unfavourable to many birds, as lack of grazing has induced an increase in high-sward communities on former good-quality marshes.
  2. Based on a high-resolution digital elevation model and two scenarios for projected rise in near-future sea levels, we employ an ArcMap allocation model to foresee the areal loss in salt marsh associated with these changes. In addition, we quantify the areal extent of inadequate salt marsh management in four EU Special Protection Areas for Birds, and demonstrate concurrent population dynamics in four species relying on managed habitats. We conclude by investigating potential compensation for climate change–induced salt marsh losses by means of more efficient management.
  3. Our models indicate that by the end of this century 15·3–43·6% of existent salt marshes will be flooded due to rising sea levels, and that inadequate managed salt marsh presently makes up around 51·1% of total marshes. Thus, re-establishing extensive areas of well-managed marshes might counterbalance the loss expected from rising sea levels during the next century.
  4. In addition to positive effects on plant diversity, this will benefit energetically challenged herbivorous waterfowl such as light-bellied brent geese Branta bernicla hrota L. and increase potential recovery of wader populations with unfavourable conservation status such as black-tailed godwit Limosa limosa L., dunlin Calidris alpina L. and ruff Philomachus pugnax L.
  5. Synthesis and applications. Implementing environmentally friendly management schemes based on extensive grazing (around 1 cow per hectare) is an important initiative to counteract the accelerating climate change–induced habitat loss in near-coastal areas across the globe, and to secure priority salt marsh habitats that support internationally important populations of breeding, wintering and staging waterfowl. However, this may only be a temporary solution that will have to be supplemented by increased reintegration with the sea and managed retreat of seawalls or near-coastal agricultural areas to effectively safeguard the future salt marsh biome.

Introduction

Coastal habitats all over the world are important for breeding, staging and wintering waterbirds (Bellrose 1980; Morrison & Ross 1989; Scott & Rose 1996), many of which are presently considered as having unfavourable conservation statuses (Delany et al. 2009). One such habitat is the Atlantic coastal salt marsh found in north-western Europe (Doody 2008a), which stands out as an avian hot spot in an increasingly culturally modified landscape. The collective presence of shallow waters and lowland salt marshes in coastal areas forms a stopover environment suitable to satisfy the enormous demand for food and rest required by many waterbirds, and is an important feeding ground for many herbivorous waterbirds throughout the year. In this respect, the shallow coastal waters, marshes and wetlands of Denmark are of outstanding international importance for breeding, staging and wintering waterbirds in the Western Palearctic (Scott & Rose 1996; Delany et al. 2009). Exploited food resources in salt marsh communities include common salt marsh grass Puccinellia maritima Huds., red fescue Festuca rubra L., sea plantain Plantago maritima L., seaside arrowgrass Triglochin maritima L., common glasswort Salicornia europaea L., sea aster Aster tripolium L. and sand-spurrey Spergularia sp. Presl. Collectively, these plants create a terrestrial food basis important to leaf- and seed-eating waterfowl (e.g. Summers et al. 1993; Therkildsen & Bregnballe 2006) and are often the only alternative to aquatic feeding when water levels are high (Clausen 2000). Herbivorous species with special affinity to these areas include Eurasian wigeon Anas penelope L. and two subspecies of brent geese Branta bernicla L. During their spring and autumn stopovers, as well as winter staging in European coastal areas, these species traditionally foraged on aquatic macrophytes such as Zostera and Ruppia (Clausen & Percival 1998). However, recent declines in the available area of these habitats have forced a substantial increase in terrestrial habitat use among these species (Clausen et al. 2012). As a result, salt marsh vegetation is now an important source of nourishment during most of these birds' annual cycle, and essential to pre-migration spring fattening in both dark-bellied brent geese B. b. bernicla L. and light-bellied brent geese Branta bernicla hrota Müller (Ganter, Prokosch & Ebbinge 1997; Clausen & Percival 1998).

European salt marshes are also key habitats for many populations of non-herbivorous waders (Delany et al. 2009), many of which rely on salt marshes both as preferred breeding sites and as important foraging habitats at high tide. Their presence in, and exploitation of, this habitat is frequently a major justification for salt marsh areas being included in estuarine Special Protection Areas for Birds (SPAs) designated and protected in Denmark under the EU Birds Directive legislation (Pihl et al. 2006). Denmark holds a higher percentage of Atlantic salt marsh than any other North European country (>25% of total area) (Doody 2008a). The total area of coastal salt marsh in Denmark amounts to 350 km2 (Buttenschøn 2007), and monitoring and preservation of this specific and rare habitat is highly prioritized. Most Danish salt marshes are protected under national and international legislation, that is, either included in the SPA network or designated as Special Area of Conservation (SAC) under the EU Habitats Directive. The Natura 2000 network (combined SPA and SAC) collectively embraces c. 330 km2 (equivalent to 94·3% of total area) as protected land (Buttenschøn 2007; Doody 2008a).

The current climatic predictions with warmer temperatures, higher precipitation and sea level rise threaten to substantially reduce the total area of salt marsh habitats by means of coastal squeeze (Hughes 2004; FitzGerald et al. 2008). Theoretically speaking, net habitat loss will be kept to a minimum, as climatic changes should merely result in an inland migration of coastal habitats. Consequently, these important staging areas should display an inward geographical shift. Nonetheless, this development can only be expected from the dynamics of a natural landscape, where recurrent inundation of lowland plant communities allows for an inward niche displacement of salt marsh plants. Due to the ever-growing anthropogenic pressures on modern coastal areas, the vast majority of European coastlines are bordered by dikes, dunes, roads and elevated field boundaries, completed to effectively protect human interests in terrestrial areas. These countermeasures have efficiently prevented the natural migration of coastal habitats and might cause substantial future losses in coastal areas due to coastal squeeze (Boorman 1992; Moeslund et al. 2011). The effects of this prediction will be an increasing inaccessibility of waterfowl food resources. Shaughnessy et al. (2012) have taken an important first step to decipher the complex interplay between sea level rise and aquatic food resources, but so far only few attempts have been made to quantify the future loss of salt marsh due to coastal squeeze, and none of them have addressed potential impacts on coastal birds. Uncertainties associated with models of climate change and sea level rise are undoubtedly partly responsible for this shortage, and limitations associated with inadequacies of traditional elevation models are an obvious further reason. Nonetheless, recent developments in both knowledge and models have made these limitations manageable, and allow for cautious predictions of future salt marsh loss. In this study, we employ a highly accurate novel Danish digital elevation model (DEM) to precisely examine the effects of sea level rise on salt marsh habitat loss. This DEM is based on remote sensing by light detection and ranging (LiDAR) technology (Vierling et al. 2008), enabling the fine-resolution spatial detail necessary to accurately predict water flows in natural landscapes.

As a further threat to salt marsh habitats, many areas are deteriorating as a consequence of modern farming practices and resultant insufficient management. Salt marshes are vulnerable plant communities dependent on continual grazing or hay harvest, and the suitability as habitat for both herbivorous and non-herbivorous waterbirds relies on these extensive farming practices (Vulink, van Erden & Drent 2010). Mowing or grazing of salt marshes is essential to maintain the low-sward, high-nutrient plant communities necessary to sustain breeding and foraging birds, and once these practices are stopped, breeding waders and herbivorous birds abandon these areas (Thorup 1998; Durant et al. 2008). The absence of proper salt marsh management results in successional processes where brackish non-tidal marshes become increasingly dominated by high-sward communities dominated by common reed Phragmites australis Trin. (Vestergaard 1998), and saline tidal marshes turn into communities of cordgrass Spartina sp., wild rye Elymus athericus (Link) Kerguélen and sea purslane Atriplex portulacoides L. (Jensen 1985; Bos et al. 2002). These changes also involve a development towards less digestible or palatable plants with lower contents of easily digestible plant constituents (protein, lipids, soluble carbohydrates) and higher contents of poorly digestible fibres and secondary metabolites (e.g. tannins) (Summers et al. 1993). The general decline in appropriate salt marsh management has indeed also been noted as one of the major management issues that have led to unfavourable conservation status of three rare breeding waders in Denmark, that is, black-tailed godwits Limosa limosa L., dunlin Calidris alpina L. and ruff Philomachus pugnax L. (Thorup 2004).

Quantifying the loss of short-turf salt marshes associated with declining management practices has received very little attention, and so far, no attempts have been made to assess potential consequences for birds. Here, we aim to assess the potential increase in short-turf salt marsh from suitable management of currently unmanaged areas, and discuss the results in the light of the expected loss of salt marsh associated with higher sea levels and climate change. Historically, most of these marshes were extensively grazed, but arrival of modern husbandry carried with it abandonment of this practice (Buttenschøn 2007). Specifically, we aim to (i) briefly describe the changes in distribution of four focal species of waterbirds associated with our study sites; (ii) assess the areal losses of salt marsh habitats under two scenarios of rising sea levels; (iii) quantify the adequacy of salt marsh management practices in formerly grazed areas important to birds; and (iv) investigate whether some of the losses associated with climate change can be compensated by changes in salt marsh management.

Materials and methods

Focal Species, Populations and Counts

We deal with four species of conservation concern dependent on grazed salt marshes, for which Denmark is host to internationally important proportions of their respective flyway-populations.

The light-bellied brent goose is our principal focal species and the one we used to guide the selection of study sites. This species was chosen because the whole East Atlantic flyway-population is concentrated within our study area during spring from March until the end of May, and lower proportions (40–80%) are found here from September to February (Denny et al. 2004). With only 7600 birds, it is the smallest goose population in the Western Palearctic (Fox et al. 2010). We also include three species of breeding waders that are red-listed in Denmark: dunlin, ruff and black-tailed godwit. These are all focal species for a national action plan for threatened meadow birds (Miljøministeriet 2005) that was launched following an alarming conservation status report (Thorup 2004). All four species have been assigned an unfavourable conservation status in Denmark (Pihl et al. 2006), and the same holds true for the three waders in the European Union (BirdLife International 2004).

Herbivorous brent geese are highly dependent on salt marshes as alternative feeding areas when water levels or depletion renders submerged or intertidal foods energetically unfavourable as feeding areas (Clausen 2000; Clausen et al. 2012). The population has gradually spread from only five regularly used and internationally important sites in the 1960s (Denny et al. 2004) to 19 sites today (this study) – a development that has been driven by eutrophication impacts resulting in large reductions in the extent of seagrass beds on four of the previously most important staging areas (Clausen & Percival 1998; P. Clausen unpublished data). Dunlin, ruff and black-tailed godwit have abandoned former breeding sites throughout Denmark, and all the main breeding sites are now found in the northern and western parts of the country (Thorup 2004). It is believed that this development is due to cessation of cattle grazing and hay harvest in combination with increased drainage in the eastern parts, paralleled by a more wader-friendly management with grazing or mowing and/or wetting of fresh- and saltwater marshes in a dozen of sites in the northern and western parts of the country (Thorup 2004). All four study species are counted on an annual basis under Danish national monitoring programmes, and regular counts exist back to the 1960s (Denny et al. 2004; Thorup 2004). For a more comprehensive overview of the species, their flyway-affinity, listings, conservation status, counts and numerical developments, see Appendix S1 (Supporting Information).

Sea Level Rise

Since 2007, when the Intergovernmental Panel on Climate Change published their fourth and latest assessment report (AR4) (IPCC 2007), several studies have indicated that the predicted sea level rise in this report was greatly underestimated (Horton et al. 2008; Pfeffer, Harper & O'Neel 2008; Grinsted, Moore & Jevrejeva 2009). Moreover, current emissions of greenhouse gases seem to follow a trajectory equal to, or even well above, the highest emission scenarios outlined by the IPCC (OECD/IEA 2008; Meinshausen et al. 2009). In an attempt to cover both the presently-estimated most likely climatic path and account for the recent updates in knowledge on sea level changes, we have chosen to focus on (i) the original A2 scenario described by IPCC (IPCC 2007) and (ii) an updated projection of sea level rise expressed by Horton et al. (2008) originating from the same A2 scenario (Table 1). Recent publications have forecasted sea level rise during the same period of at least 80 cm (Pfeffer, Harper & O'Neel 2008) or even well above 1 m (Grinsted, Moore & Jevrejeva 2009), but in this study we have chosen to rely on numbers in the middle range. To acknowledge the restricted potential for inland expansion of current salt marsh habitats in modern agricultural landscapes, we assumed a model denying inward displacement of agricultural land. Our flooding calculations therefore depict a future situation where no precautionary measures have been taken to secure continued salt marsh distribution at the expenses of other land use interests.

Table 1. Projected mean and range of year 2090–2099 sea level rise based on the original and updated versions of the A2 scenario
ScenarioSea level rise (m)
MeanRange
IPCC0·370·23–0·51
Horton et al.0·770·68–0·89

To quantify potential salt marsh loss in the wake of predicted sea level changes, we employed an ArcMap 10.0 cost allocation procedure based on a novel Danish remote sensing–based digital elevation terrain model (DK-DEM/terrain), with a horizontal spatial resolution of 1·6 m and a vertical accuracy of 5·9 cm (SE: 3·44 cm) (Rosenkranz & Frederiksen 2011). To include large geographical scale, and ensure ecological relevance to herbivorous species of waterbirds, we focused our analysis on a subset of salt marsh areas within sixteen Danish EU Special Protection Areas (SPA's no. 1, 2, 8, 12, 15, 23, 25, 27, 36, 38, 39, 40, 53, 57, 76 and 112) and three sites outside the SPA network (Fig. 1), chosen because these currently have international importance to the strictly herbivorous East Atlantic flyway-population of light-bellied brent geese. Collectively, the included areas hold more than 126 km2 of near-coastal Atlantic salt marsh. Spatial delimitation of salt marsh areas was based on the extraction of extensive survey data from The Danish Nature and Environment Portal (Danmarks Miljøportal: http://www.miljoeportal.dk), where complete geo-referenced data on distribution of salt marsh are available due to its legal protection under the national Nature Conservation Act § 3 and the EU Habitats Directive Annex 1 (habitat types 1310, 1320 and 1330). This enabled exceptional fine-scale digitized data on geographical distribution of salt marsh within the study sites, and the digitized data layer was then used as an overlay on the DEM, ensuring that only actual salt marsh areas were included in the model. The model procedure works by first identifying all cells in the DEM consisting of salt marsh with an altitude lower than projected sea level rise, and proceeds with a flooding algorithm which ensures that only cells in connection with the coastline (or cells connected to the sea only by areas lower than the elevated water level threshold) will be inundated. This procedure guaranteed that higher situated areas and salt marsh protected by dikes, dunes, etc., are unaffected by the rising water. Based on the number of flooded cells under the two scenarios, we calculated the areal loss of salt marsh for three coastal areas in Denmark (the Wadden Sea, the Limfjord and the Baltic Sea) and determined an overall estimate of how rising sea levels might impact future marshes.

Figure 1.

The sixteen Danish EU special protection areas for birds (SPAs) (numbers) and three non-SPAs (letters) included in the model of sea level rise. Salt marshes within these areas are split into three geographical regions (1: The Wadden Sea, 2: The Limfjord and brackish lagoons, 3: The Baltic Sea).

Management

Recent salt marsh management practices were determined in four EU SPAs (No. 1: Nibe Bredning, 15: Randers & Mariager Fjords, 25: Mågerodde and 27: Agerø) also included in the flooding analysis. Collectively, these areas hold 39·9 km2 of salt marsh. Based on the criteria outlined in Table 2, management was classified on a plot-to-plot basis into three groups (well managed, poorly managed or unmanaged), using a composite measure relying on vegetation height and grazing intensity. Doody (2008b) gives a comprehensive overview of the grazing management regimes occurring on salt marshes, and the three groups defined here correspond very well to vegetative states 2, 3 and 4 in that work (Table 2). Grazing schemes on individual patches were determined from existence or the absence of both new and old signs of grazing (corrals, drinking troughs, footprints, grazed vegetation), and were easily recognized in most cases. Data gathering in SPA 1 took place in May–June 2008 by the consultancy ornit.dk (Kjeldsen & Nielsen 2009), and we were granted access to their GIS-data by Aalborg municipality who had paid for the assessment. SPA 15 was mapped in November 2010, and SPAs 25 and 27 in 1993 (Clausen & Percival 1998). The total salt marsh area assigned to each of the three management practices was calculated from digitized field maps in ArcMap 10.0 (see Fig. S2, Supporting Information).

Table 2. Criteria for classifying salt marsh areas as well, poor or unmanaged
ManagementVegetative state (cf. Doody 2008b)Vegetation height (cm)Grazing/mowing schemeValue to herbivorous birdsStructural characteristics
Well managed2<15RegularlyImportantHeterogeneous
Poorly managed315–30IrregularlyOf less interestIntermediate
Unmanaged4>30NoneUnusedHomogenous

Results

Bird Numbers and Distribution

During the past 40 years, light-bellied brent geese have spread to a suite of new sites, in all of which salt marsh forms an important supplementary alternative to the preferred submerged seagrass and algae food (Fig. 2). This importance has become increasingly higher because of the previously mentioned losses of submerged food resources. At the same time, several sites have been abandoned by the three rare breeding waders (Fig. 2). Despite the implementation of a national action plan aimed at improving conservation status of dunlin, ruff and black-tailed godwit (Miljøministeriet 2005), our results show that this has been insufficient to stop the decline in numbers of breeding pairs and occupied sites for all three species (Fig. 3). Since 1970, the per cent of the national wader populations found in our study sites have remained fairly stable for both dunlin (22% in 1970 and 26% in 2011) and black-tailed godwit (6% in 1970 and 8% in 2011), whereas ruff has disappeared from all but one of our study sites that held 10% of the national population in 1970, and only one pair and 2% of the population in 2011 (see Appendix S2, Supporting information).

Figure 2.

Changes in the distribution of (a) staging light-bellied brent geese and breeding, (b) dunlin, (c) ruff and (d) black-tailed godwit from the 1960s to 2011. For the brent geese, the circles give the proportion of the entire flyway-population found on average during 1968–1974 (closed circles, estimated population 1950 birds) and 2004–2011 (open circles, population 7600 birds). Note that the relative importance of four sites has declined since the 1960s, but Special Protection Area 27. (marked by an *) has an increasing importance, and it can thus not be seen on the map that this site was used also during 1968–1974. For the three wader species, dots indicate the presence of the species, with closed dots showing breeding sites used during 1964–1972 (redrawn from Thorup (2004)) but abandoned today, and open dots breeding sites in 2010–2011.

Figure 3.

Development in the number of breeding pairs (solid lines) and sites (dashed lines) occupied by dunlin, ruff and black-tailed godwit in Denmark from the mid-1960s until today.

Sea Level Rise

Based on the expected rise in global sea levels during the next century presented in Table 1, the modelled average loss in salt marsh within the protected areas included in this analysis amounted to 15·3% (19·3 km2) (IPCC scenario) and 43·6% (54·8 km2) (Horton et al. scenario, Table 3, Fig. S1, Supporting Information). Beneath this overall mean, substantial regional variation was found in the susceptibility of local salt marshes to rising water levels. Areas in the Wadden Sea seemed fairly robust to future inundation (2·6% and 5·2% loss under the two scenarios), whereas the Danish Limfjord was much more vulnerable to the effects of climate change (20·0% and 53·6% loss, respectively). This variation is likely to be explained by local topography resulting from the differences in historic coastal protection. Coastlines facing the North Sea (e.g. SPA 53 and 57) are frequently subjected to periods of storm-induced extreme water levels and therefore often protected by man-made dikes to prevent flooding in these areas (CPSL 2005).

Table 3. Forecasted loss of salt marsh habitats in three Danish coastal areas in the years 2090–2099 based on the original and updated versions of the A2 sea level rise scenarios
Scenario RegionSPAsAreal salt marsh loss
IPCCHorton et al.
%km2%km2
  1. SPAs, special protection areas.

1: Wadden Sea53, 572·60·65·21·3
2: Limfjord1, 8, 12, 25, 27, 38, 39, 40, A20·013·853·637·1
3: Baltic Sea2, 15, 36, 76, 112, B, C15·24·852·316·4
Total 15·319·343·654·8

Management

The results from the mapping of management practices in SPA 1, 15, 25 and 27 showed that 48·9% of total salt marsh in these areas (corresponding to 19·5 km2) was well managed, whereas poorly managed and unmanaged areas accounted for 18·9% (7·5 km2) and 32·2% (12·9 km2), respectively. However, the results showed substantial differences between the three areas (Table 4), indicating that salt marsh management practices are much more efficient in SPA 1, 25 and 27 compared to SPA 15. Well-managed areas were dominated by salt marsh grass, red fescue, sea plantain, common glasswort and seaside arrowgrass, while red fescue, sea aster, sea purslane and cordgrass were frequent on poorly managed marshes. Unmanaged marshes consisted almost exclusively of high homogenous reed beds. In addition to the positive effect on species composition, grazing marshes were also characterized by being structurally more diverse with a mosaic of short vegetation, tussocks and occasional wet depressions. This heterogeneity is especially important to attract breeding waders (Table 2).

Table 4. Distribution of salt marsh practices (well, poorly and unmanaged) in special bird protection areas (SPAs) 1, 15, 25 and 27. Management practice on the latter two was mapped in 1992–1993 (Clausen & Percival 1998)
ManagementSPA 1 (%)SPA 15 (%)SPA 25 + 27 (%)Total (%)
Well managed52·912·381·448·9
Poorly managed31·717·57·518·9
Unmanaged15·470·211·132·2

Discussion

Future declines in available salt marsh habitat seem to be an inevitable consequence of the rising waters associated with climate change. However, the extent of this decline is highly dependent on accurate predictions of both sea level rise and sedimentation/accretion processes. Based on two recent scenarios employed in our model, the percentage loss of salt marsh varies between 15·3% and 43·6%. These numbers are generally in good agreement with other studies on salt marsh loss. Boorman (1992) found a local areal loss above 40% in Great Britain with a global sea level rise of 0·5 m, and in Denmark, Moeslund et al. (2011) also found losses very similar to the results presented in this study. The large difference between the estimates from our two scenarios highlights the importance of clarifying the exact increase in oceanic water levels when trying to forecast the environmental impacts associated with climatic changes. That said, both of these numbers indicate significant reductions in important bird habitats. Such losses can have negative impacts on a number of birds relying on these areas, and may exert further pressure on herbivorous birds that in recent years have seen significant declines in aquatic food sources (Clausen & Percival 1998; Burkholder, Tomasko & Touchette 2007). Clausen et al. (2013) found that compared to birds foraging in traditional coastal habitats, brent geese foraging on agricultural land were energetically constrained due to an elevated energy expenditure in these areas. Hence, it is well documented that for at least some herbivorous birds, loss of salt marsh habitats may potentially have negative effects on fitness. If herbivorous birds are increasingly forced into nearby agricultural fields, this might result in elevated crop damage and negative socioeconomic impacts (van Roomen & Madsen 1992). Noting that the entire population of East Atlantic light-bellied brent geese rely on salt marsh during spring-staging in Denmark, it could be detrimental for the future conservation status of this small and vulnerable population if no action is taken to counteract the impacts of climate change on Danish coastal marshes. The loss of salt marsh due to inundation might of course to some degree result in a corresponding increase in shallow water macrophyte communities as marshes become increasingly flooded (Shaughnessy et al. 2012). While this may potentially improve conditions for some of the herbivorous inhabitants, this development is of little use to waders dependent on suitable breeding conditions. Furthermore, most of the Atlantic salt marshes included in this study are found in non- or only moderate tidal areas, where the ‘rescue-effect’ of increasing accretion rates is unlikely (see discussion below).

The change in management practice on salt marsh habitats has had further negative impacts on avian wildlife relying on these areas. Approximately half of the salt marsh areas included in this study were poorly managed or not managed at all. As both herbivorous waterbirds and breeding waders are dependent on short grass communities to meet nutritional demands and secure satisfactory breeding conditions, these areas are effectively excluded as good habitat and abandoned when management ceases (Thorup 1998). Grazing by the geese themselves might, in certain highly exploited areas, aid the maintenance of a short-turf community in the winter staging period, but as noted by Olff et al. (1997) the combined effects of high summer primary productivity and annual movements of the birds imply that in the absence of further management initiatives, continued vegetation succession is inevitable. However, the fact that currently less than half of salt marshes are well managed creates an opportunity for buffering the effects of climate change. Assuming that the areal quantification of management practices in SPAs No. 1, 15, 27 and 25 is representative of all sites included in this study, the possible gain in good-quality salt marsh through suitable management on poorly and unmanaged areas could amount to 64·4 km2. Implementing this strategy would more than counterbalance the loss of salt marsh from rising sea levels, and result in a net surplus of salt marsh habitats available to birds. To some extent of course, this management solution to the coastal squeeze problem is only temporary if sea levels continue to rise even further in the future. Accepting the disappearance of large salt marsh areas may not fulfil demands of the EU Habitats Directive, which states that ‘The conservation status of a natural habitat will be taken as ‘favourable’ when: – its natural range and areas it covers within that range are stable or increasing’. Hence, loosing half of the marshes may indeed be interpreted as ‘unfavourable’. Nonetheless, considering the relative slow impacts of climate change processes, implementation of efficient management schemes could be an efficient ‘quick fix’ to temporarily manage many species of birds relying heavily on these habitats. To secure long-term sustainability of coastal marshes, these initiatives will have to be supplemented by more comprehensive managed realignment and reintegration with the sea, effectuated by controlled retreat of seawalls or near-coastal agricultural areas (Wolters, Garbutt & Bakker 2005). Thorough evaluation of the effects of managed realignment is scarce, but available data suggest an initial lack of species richness and composition behind natural marshes (Wolters, Garbutt & Bakker 2005; Mossman, Davy & Grant 2012). In the long run however, these measures are vital to safeguard the future salt marsh biome. From a bird's perspective, it must be emphasized that the sole establishment of new areas is insufficient to ensure habitat availability. In order to prevent vegetation succession rendering marshes unsuitable to most meadow birds, establishing new coastal habitats should always be accompanied by appropriate management in the form of extensive grazing.

The focus of our analysis has been to improve habitat loss of herbivorous birds and nesting waders requiring a vegetative state of short-turf costal salt marsh. This entails recurrent grazing of coastal habitats, which is only one end of the spectrum of management regimes, and may sometimes conflict with additional nature conservation values (Doody 2008b). Grazing is known to generate habitat heterogeneity and increase plant species richness (Bos et al. 2002; Wolters, Garbutt & Bakker 2005; Buttenschøn 2007), hence often facilitating many nature conservation goals besides bird usage. Nevertheless, deciding which management practice to implement in any given area should always be founded on a preceding assessment of nature conservation interests. In the case of prevailing reed beds replacing formerly grazed species-rich marsh communities, which define the vast majority of marshes included in this study, there should be little hesitancy to recommence proactive management. Thus, defining our suggested change in management within the framework of Doody (2008b) would mean a transition in vegetative state from state 4 (abandoned, formerly grazed) to state 2 (moderately grazed).

Stock level and stock type are critical to the impact of grazing on salt marsh habitats, and timing of release of livestock is important to prevent trampling of wader nests. Detailed descriptions of the different grazing schemes can be found in the study by Buttenschøn (2007) and Doody (2008b), but as a general guidance, cattle grazing at a density of ≈1 cow per hectare from late May/early June to late October is appropriate to secure suitable bird habitats (Norris et al. 1997; Doody 2008b; Vulink, van Erden & Drent 2010). Information about historical grazing regimes and stocking rates in currently unmanaged areas is generally lacking, and the large geographical scale included in our study excludes overall generalizations. However, we believe the vast majority would have been managed in a similar way up until 30–50 years ago, when most Danish marshes were stocked with cattle during summer.

All over Europe, many bird populations dependent on salt marsh habitats are currently decreasing at an unprecedented rate (Delany et al. 2009), and the ability to improve their main habitats by relatively simple means is a window of opportunity that should not be missed. Implementation of improved salt marsh management would benefit both herbivorous waterfowl and potential recovery of declining wader populations. Especially for the extremely site-faithful dunlins (Thorup 1999), re-establishment of a continuous network of managed salt marshes seems to be essential to establish links between the few remaining breeding populations, and an important action to counteract potential problems of inbreeding in isolated populations, which have already been recorded in Sweden (Blomqvist et al. 2010). Based on the high conservation status declared for these habitats, it is rather thought-provoking that so little is done to preserve them, and a statutory obligation to effectuate sufficient salt marsh management has yet to be satisfactorily implemented in national legislation. As salt marsh degradation is usually a reversible process (Vickery et al. 1997; Thorup 1998; Bos et al. 2002), all that is needed to significantly improve these areas is an environmentally friendly management scheme involving local land owners. As such, the simple measure of extensive grazing is an example of an initiative with low costs and high ecological benefits.

The high demands on elevation models and necessity of site-specific evaluation of management schemes restrict our analysis to salt marshes in Denmark. However, this study is the first to model the dynamics of sea level rise and changing management on a near-national level, and the conclusions reached should be equally applicable in coastal habitats across the globe, and certainly throughout Europe, where the very populations dealt with here rely on salt marsh habitat further down the flyway. Rising sea levels are affecting coastal areas world-wide, and examples of declines in salt marsh management are numerous in contemporary scientific literature (e.g. Norris et al. 1997). As such, implementing proper management actions should relieve climate change–induced coastal habitat loss throughout the Western Palearctic. This will benefit bird populations across the board from redshank Tringa totanus L. in the UK (Norris et al. 1997) and barnacle geese Branta leucopsis Bechstein in the Netherlands (Vulink, van Erden & Drent 2010) to large numbers of wintering waders in southern Spain (Perez-Hurtado, GossCustard & Garcia 1997).

In our model of sea level rise, we assume a non-existent inward displacement of salt marsh habitats. In natural landscapes, rising seawater would imply a landward movement of coastal habitats, resulting in a geographical shift without necessarily causing areal losses. Conversely, the modern landscape with high anthropogenic pressures and intensive land use effectively prevent salt marsh areas from expanding inward. Dikes, dunes, dams and elevated field boundaries border every piece of coastal land in modern European landscapes and therefore inevitably result in coastal squeeze. Hence, in the current landscape situation, loss of salt marsh will follow the trajectory outlined in our model, describing the situation where no countermeasures have been taken to prevent these declines.

Even with the recent advances in elevation models and updated knowledge on climate change dynamics, a confident prediction of areal salt marsh loss is still far from complete. Some studies argue that enhanced sedimentation and associated accretion rates might to some degree counterbalance the effects of minor sea level rise (Thom 1992; Roman et al. 1997; Bartholdy, Christiansen & Kunzendorf 2004). However, accretion processes are highly variable in both time and space (Cahon & Reed 1995; Roman et al. 1997) and dependent on a wide array of physical processes in the land–water interface. Thus, tidal amplitude, sediment supply, accumulation rates, salt marsh topography, plant composition, wave erosion and storm frequency all are important factors controlling accretion rates. A thorough quantification of all these parameters across several geographical areas is an impossible task, and the unpredictable nature of accretion rates renders them inappropriate for long-term modelling (Reed 1995; Kirwan & Murray 2008). We agree with Moeslund et al. (2011) that currently no data are available to fully address the influence of accretion on the spatial scale included in this study, and the results presented here are therefore an upper limit to areal losses, with no compensation from higher accretion rates (FitzGerald et al. 2008). For now, however, there seems to be a general consensus that sea level rise presently exceeds marsh accretion (Roman et al. 1997; Andersen, Svinth & Pejrup 2011). The most likely scenario is that in areas with frequent inundation, high supply of sediments and high rates of sedimentation (e.g. the Danish Wadden Sea), coastal squeeze and areal loss will be a minor issue (Klagenberg et al. 2008), while in sheltered, low-tidal areas (e.g. inland fjords such as Limfjorden and the Baltic Sea) sea level rise could result in far-reaching negative impacts (Vandenbruwaene et al. 2011). At any rate, historic data show that much higher sea levels and large-scale flooding of landmasses have occurred several times in the past (Klagenberg et al. 2008).

Acknowledgements

Thanks are due to Aalborg Municipality for kindly providing some of the management data, Jesper Moeslund for general discussions on the modelling procedure and Pat Thompson and one anonymous reviewer for very constructive comments on an earlier draft.

Ancillary