Missing native oyster (Ostrea edulis L.) beds in a European Marine Protected Area: should there be widespread restorative management?

49 Anthropogenic pressures on the marine environment have escalated and shellfish habitats 50 have declined around the world. Recently, Marine Protected Areas (MPAs) have rapidly 51 increased in number and extent, but in Europe as elsewhere, management baselines rarely 52 account for historical conditions. 53 In the present study, the Dornoch Firth protected area (NE Scotland) was investigated as 54 well as three adjacent inlets and 50 km of open coastline. The area has low levels of 55 industrial development, is sparsely populated and conservation features were previously 56 considered “Favourable”. The aim of the present study was to investigate the historical 57 presence of native oyster ( Ostrea edulis ) beds, a habitat that is now rare and of conservation 58 importance throughout Atlantic Europe. 59 Centred on the protected area, but also gathering broader information from East Scotland, 60 an interdisciplinary review was made of archaeological records, navigational charts, 61 historical maps, museum collections, land-use records, fisheries records, public online 62 databases and naturalists’ records. Intertidal and subtidal surveys were also conducted and 63 sample oyster shells were radiocarbon dated. 64 The present study shows that O. edulis occurred in the inlets and open coast areas of NE 65 Scotland, and specifically in the protected area: Probably since the end of the last glaciation 66 to the late 1800s when they were likely fished to extinction. 67 Habitat restoration in protected areas is an emerging global theme. However, marine habitat 68 restoration is highly unusual and European oyster restoration is presently confined to 69 remnant populations with a clear history of exploitation or dwindling associated fisheries. An 70 interdisciplinary review of baselines will probably show scope for O. edulis restoration in 71 many other European MPAs. 72


Introduction
The decline in the condition of the marine environment has been linked to the historical dependence of human societies on the sea as a source of food, natural resources and transport (Jackson et al., 2001;UNEP, 2006).As human populations have grown, so too have the pressures on coastal areas where people have preferred to settle (Ray and McCormick-Ray, 2004).These pressures include overfishing, physical damage to benthic habitats, pollution and eutrophication but overall, they are evident globally and are particularly intense in Atlantic Europe (Halpern et al., 2008).Amongst marine habitats, shellfish reefs are especially sensitive to anthropogenic disturbance (see Cook et al., 2013) and are one of the most "imperilled" marine ecosystems on Earth (Beck et al., 2009).
In the last 25 years there has been an increasing recognition of the importance of protecting the marine environment through the designation of Marine Protected Areas (MPAs); now a mainstream global management tool (UNEP- WCMC and IUCN, 2016).In Europe the environmental legislation that facilitates the creation of MPAs includes the Habitats Directive (92/43/EEC), the Marine Strategy Framework Directive (MSFD: 2008/56/EC) and related domestic legislation within member states (see Frost et al., 2016).In the UK, MPAs account for 16% of the sea area and, in keeping with other Atlantic European countries, just under half of these are Special Areas of Conservation (SACs) designated under the EU Habitats Directive (Frost et al., 2016).The condition and management plans of SACs are typically judged against the ecological condition of the site when the Habitats Directive was introduced (1992).This 'baseline' thus provides the benchmark against which changes are measured (cf Dayton et al., 1998).
Since baselines have usually been defined at the time a protected area was designated, it is plausible that they might not reflect past conditions when certain species and habitats were more common, biodiversity was higher or when locally extinct species were present.Indeed, it has previously been observed that conservation and Marine Spatial Planning usually make no reference to historical conditions, specifically those prior to the influence of large scale anthropogenic changes since the Industrial Revolution (Howarth et al., 2014).However, assessing historical conditions can be difficult because impacts often pre-date readily available documentation or any kind of scientific study.
The 'shifting baselines' paradigm was first referred to in the scientific literature to suggest that the concept of 'pristine' in the marine environment evolves across different generations of sea users (Pauly, 1995;Sáenz-Arroyo et al., 2005).As a result, the past-history of the sea can inform management policies (Roberts, 2007) which are otherwise based on baseline conditions derived from limited historical data or current conditions (Gatti et al., 2015).Although the shifting baselines concept has been commonly applied within a fisheries context (Ferretti et al., 2015), it is also applicable to habitats and whole ecosystems where losses in biodiversity are typically measured against very recent conditions with little regard to the 'millennia of habitat loss' (Airoldi and Beck, 2007).Recent work under the MSFD calls for careful consideration of environmental baseline and reference conditions (see OSPAR Commission, 2011) because, in the context of 'sustainable management', it is logical to consider what could be achieved as well as what is being achieved in terms of benefits to human society.Shellfish reefs, and oyster reefs in particular, have been severely impacted by overexploitation, disease and other human induced pressures throughout their global distribution and mostly in the late 1800's and early 1900's (Gatti et al., 2015;Howarth et al., 2014).In the USA declines of up to 88% of the American oyster (Crassostrea virginica) biomass have been documented in some waterbodies (e.g., Chesapeake Bay) with consequent losses in associated ecosystem services such as water filtration (zu Ermgassen et al., 2012).In Europe, the native oyster Ostrea edulis has been a documented source of food since at least Roman times and in the North Sea over-exploitation has led to the loss of extensive beds (Anon., 1886(Anon., , 1882)).
O. edulis is found in Atlantic Europe and North Africa from Morocco to Norway as well as the Black Sea and in the Mediterranean (Cano et al., 1997;Zaitsev and Alenxandrov, 1998).It usually inhabits relatively shallow sheltered estuaries but has been historically recorded in deeper waters of up to 80 m off the Channel Islands and Grimsby (UK) and the wider North Sea (Laing et al., 2005).Native O. edulis beds were also present in Scotland but the advent of industrialised trawling (Roberts, 2007) and the development of the railway system at the end of the 19 th century, facilitated a sharp increase in demand by urbanised markets and a corresponding extirpation in the Firth of Forth (Scotland, UK; Fig. 1A) where an oyster (O.edulis) bed was reported to cover an area of about 10 x 30 km and to produce up to 30 million oysters in some years (Olsen, 1883;Fulton, 1896;Thurstan et al., 2013).Oyster populations in the UK still exist in remote sheltered sealochs on the Scottish west coast, most notably Loch Ryan (University Marine Biological Station Millport, 2007), and in a few inlets and coastal waters of southern England and Wales (Woolmer et al., 2011).Overall, O. edulis habitats are endangered in Europe, losses and threats have been overlooked and there is scope for restoration (cf Airoldi and Beck, 2007).As a result, in areas with well documented historical O. edulis fisheries, as well as remnant fisheries and populations, the notion of localised restoration is gaining traction (Kennedy and Roberts, 1999;Gercken and Schmidt, 2014;Sawusdee et al., 2015;Smaal et al., 2015) and O. edulis beds are now identified as a priority marine habitat for protection in European MPAs (OSPAR Commission, 2011).However, the European Atlantic MPA network covers a substantial area and O. edulis populations may well have been ubiquitous in what is now the MPA network: over and above those places where the recent history of fisheries is apparent in the modern context.
The Dornoch Firth is a little-studied protected area located in the remote northeast coast of Scotland with apparently low fishing pressure and a sparse human population.Unlike many areas in the North Sea, the Dornoch Firth (Fig. 1) has been virtually unaffected by industrial development and is considered of high environmental quality (Mackay et al., 2004).The Dornoch Firth is a Special Area of Conservation (SAC) and a Special Protected Area (SPA) under European Directives and, as such, it is also part of the OSPAR network of MPAs.In addition, the intertidal zone bounding the Dornoch Firth is a Site of Special Scientific Interest (SSSI).Two Nature Conservation Orders banning collection of shellfish and other invertebrates (except for mussels) provide protection.Although an important fishery for oysters existed in the past in Scotland (notably the Firth of Forth; Thurstan et al., 2013), there is no mention of O. edulis in conservation objectives or management planning documents for any of the marine protected areas of North-East Scotland, including the Dornoch Firth (SNH, 2016).
The present study focuses on O. edulis and through an inter-disciplinary review and novel first-hand field investigation, aims to determine the likely historical presence of native O. edulis beds in the Dornoch Firth protected area and the three neighbouring inlets and open coast in northeast Scotland.The paper ultimately aims to indicate if oyster restoration could be relevant to the conservation objectives of the wider European Atlantic MPA network, with its long history of environmental degradation (Gilbert et al., 2014); especially where there are now no remnant populations or contemporary knowledge of historical fisheries.

Site details
The Dornoch Firth is an estuarine inlet in NE Scotland and, along with the Inverness, Beauly and Cromarty Firths, is part of the much larger Moray Firth system (Figs.1A, B; Hunter and Rendall, 1986).The Firth is a sheltered, semi-enclosed embayment of glacial origin, ca.23 km long and occupying 12 273 ha (Stapleton and Pethick, 1996).The main freshwater inputs are from relatively small rivers (Balls, 1994;SEPA, 2011).The subtidal zone is shallow, not exceeding 10 m, with the exception of the entrance channel between Gizzen Briggs and Whiteness Sands (Fig 1B) where 17 m depth has been recorded (UK Hydrographic Office, 1978).Minimum water temperatures are 5 to 6°C in February and in July to September temperatures range from 12°C to 14°C (Balls 1994;unpublished SEPA Shellfish Growing Waters data 2006to 2008).
The outer Moray Firth is a fully saline (>34 psu), open water system with a decreasing salinity gradient towards the inner firths (Balls, 1994).Most of the inner Dornoch Firth can be regarded as brackish, with salinities ranging from <9 psu in the uppermost section to 25-31 psu in the main channel.The outer sections east of the Dornoch Bridge are less influenced by freshwater inputs and conditions here are regarded as fully marine.
The waters in the Dornoch Firth are well mixed from tidal currents and occasional wave action from easterly winds.According to Balls (1994) oxygen saturation at or exceeding atmospheric levels was found throughout the Dornoch Firth whilst dissolved oxygen levels were consistently above 5 mg l -1 from 2006 to 2008 (7-12 mg l -1 in most sampling dates), particularly in the winter months (SEPA, 2009).

Searches
A variety of information sources were searched, including navigational charts, historical maps, museum collections (biological and archaeological), land-use records, fisheries records, statistical records, public online databases and naturalists' descriptions (see Appendix A Table A1).Searches for records of Ostrea edulis in Scotland (live specimens and oyster bed habitats) were conducted using the National Biodiversity Network online atlas (NBN, www.nbn.org.uk), the Marlin website (www.marlin.ac.uk), the Marine Environmental Data Network (MEDIN; www.oceannet.org)and the UK Marine Recorder (July 2014 snapshot; http://jncc.defra.gov.uk/page-1599).
The archaeological online tool 'Canmore' (http://canmore.org.uk),compiled and managed by Historic Environment Scotland, and the Highland Council Historical Environment Record (HER; http://her.highland.gov.uk/) were interrogated using "midden", "shell midden", "shells", "Ostrea" and "oyster" as search terms.UK grid references (OSGB1936) for possible midden sites were plotted onto terrain maps using the open source GIS software qGISv2.10.1 (QGIS Development Team, 2014) and three sites in the Dornoch Firth were visited to verify records (below).
Government reports, statistical records and grey literature in local and university libraries and the National Library of Scotland were systematically searched for references to oysters (see search terms above) in the Moray and Dornoch Firths (e.g., Anon. 1885Anon. -1886;;Groome, 1901Groome, , 1884Groome, -1885)).Similar online searches were performed using resources such as The Internet Archive (www.archive.org),Google Scholar and Google Books, where Public Domain 19 th century (and earlier) references to "oysters" and "oyster fisheries" were available (e.g., Blaeu, 1654;Dunbar et al., 1837;Eyton, 1858;Groome, 1884Groome, -1885 p76; p76;Lizars, 1850).The electronic versions of the Statistical Accounts of Scotland for the Dornoch and Tain parishes (available from EDINA http://edina.ac.uk/stat-acc-scot/ and Google Books) were consulted (Sinclair, 1791;Anon., 1845).Historical and/or current information on O. edulis landings were obtained from the UK's Marine Management Organisation (MMO), the Annual Reports of the Fisheries Board of Scotland (Anon., 1882) and the Scottish Sea Fisheries Statistics (Appendix A, Table A.1).
Published accounts were complemented by consulting experts from the universities of Aberdeen and York, inshore fisheries groups, local fishers and conservation and environment agencies (Scottish Environment Protection Agency, Scottish Natural Heritage).
Museum curators and independent researchers on palaeontology, bivalve ecology, archaeomalacology, palaeoanthropology, Iron Age and medieval archaeology and natural history, were also asked for accounts of oysters and oyster fisheries in the study area (see acknowledgements for list of consultees).Specimens were sought in collections of bivalves, archaeological finds and fossils held at The National Museum of Scotland Edinburgh (NMS), the Hunterian Museum in Glasgow, the Glasgow Museum and the Natural History Museum of London.Local museums and field societies in NE Scotland were also consulted including Elgin Museum, Inverness Museum and Art Gallery, Dornoch Historylinks Museum, Tain Museum and Tarbat Discovery Museum (Portmahomack).Overall, searches were focused on the Dornoch Firth and the wider Moray Firth but also captured contextual accounts from North East Scotland.

Fieldwork
Between 2014 and 2017 shell middens recorded in the Dornoch Firth area were visited (e.g., Ness of Portnaculter, Cuthill Links; Fig. 1C).In the same period, beach transects, subtidal SCUBA surveys, and drop-down video surveys were conducted throughout the Dornoch Firth and opportunistically in adjacent areas (see Fig. 1).The investigators walked from the upper littoral to the lower infralittoral zones, swam and searched for 45 mins on SCUBA surveys or towed a camera for 10 minutes on drop-down video surveys to record the presence and abundance of live O. edulis or their shells.Subtidal O. edulis shell densities were estimated using the standard semi quantitative SACFOR abundance scale (S=Superabundant; A=Abundant; F=Frequent; O=Occasional; R=Rare; Joint Nature Conservation Committee, 1997).All O. edulis shell collected during these surveys were catalogued and deposited in the National Museum of Scotland collections in Edinburgh.

Radiocarbon dating
A subsample of five oyster (O.edulis) shells collected during the intertidal and subtidal fieldwork was aged using radiocarbon dating techniques.Three oyster shells were used from the subtidal and two from the intertidal (see Fig. 1C).The shells were selected to be representative of the geographical distribution of the collection sites.Any encrusting fauna, such as barnacles and bryozoans found to be growing on the dead oyster shells were scraped off and a sub-sample of shell (~1 cm 2 ) next to the ligament was cut out using a Dremel saw.This incorporated multiple growth increments, which has been found to reduce the errors associated with intra-shell variability (Culleton et al., 2006;Rick et al., 2012) and is more consistently preserved in oyster shells than the comparatively fragile outer edges.The O. edulis shell sub-samples were rinsed in fresh water, air dried and sent to the 14 CHRONO Centre at Queen's University Belfast (QUB) for radiocarbon dating.Prior to analysis, the outer layer of shell was removed by etching with 1% hydrochloric acid (HCl) to ensure that any potential contaminant carbon from post-depositional exchange and/or adsorbed modern CO 2 was removed.The samples were then hydrolysed to CO 2 with phosphoric acid.The CO 2 was converted to graphite on an iron catalyst using the hydrogen reduction method (Vogel et al., 1987).The 14 C/ 12 C ratio and the 13 C/ 12 C ratio were then measured by accelerator mass spectrometry (AMS).The sample 14 C/ 12 C ratio was background corrected and normalised to the HOXII standard (SRM 4990C; National Institute of Standards and Technology).The radiocarbon ages were corrected for isotope fractionation using the AMS measured 13 C/ 12 C, which accounts for both natural and machine fractionation.
Radiocarbon dates of the oyster shells were calibrated using CALIB 7.0.4 with the Marine13 calibration curve (Reimer et al., 2013;Stuivert and Reimer, 1993), which includes a correction for the global surface ocean reservoir effect (ca. 400 years), and a local reservoir effect correction (ΔR) (−29±53 years; Russell et al. 2015).
No permits were required for the described study, which complied with all relevant regulations.Field studies did not involve endangered or protected species.

Results
Archaeological records indicate that the first known human inhabitants of the Dornoch Firth area can be traced to the Mesolithic Period after the last glaciations, when the shoreline had retreated to approximately its present position (Doody et al., 1996;Fletcher, 1998).Twentytwo native oyster shells from this period collected at the Cuthill Links midden (Canmore Site No. 85364; Figs.1C and 2A; Appendix A Table A.2) were located at the Dornoch Historylinks Museum (see Batey, 1993).The Glasgow Museum also has a collection of Mesolithic and Neolithic remains from the Meikle Ferry midden (Canmore site No. 14654; Fig. 1C; Appendix   A Table A.2) that are dated 3 000 to 6 000 BC (ca.5-8 000 BP) and consisted of man-made artefacts (e.g., arrowheads, flint and ash), animal remains, cockle, periwinkle and oyster shell (Batey, 1993).
Oyster, cockle and mussel shell were ubiquitous in middens associated with Mesolithic, Neolithic and Bronze Age settlements, dating from ca. 8 000 BP, e.g., Meikle Ferry, Rosskeen (near Invergordon) and Cuthill Links (Batey, 1993) through to the 1800s (Appendix A Table A.2; Figs.2A and 3), suggesting that O. edulis was common in coastal areas close to where they were found.In general, these middens consisted primarily of periwinkle and cockle shell with less mussel and oyster shell (Historic Environment Scotland, 2014; Fig. 2B).Neolithic chambered cairns and Bronze and Iron Age 'brochs' and hillforts have also been recorded in archaeological surveys on both sides of the Dornoch Firth at Tarlogie and Castle Corbert, some of them with associated shell middens (C.Hatherley pers.com.2014).
O. edulis shell has also been recorded (as 'oysters') from more recent historical periods at several archaeological sites excavated in the surrounding area.For example, excavations at an early medieval settlement at Portmahomack (Fig. 3) yielded oyster shell associated with middens from the early Christian monastic period (6-9 th centuries) through to civilian settlements in the 13-15 th centuries (Carver et al., 2016;Holmes, 2017).Two hundred and twenty oyster shells (along with other marine shells and animal bones) from late medieval and early modern times (16-18 th centuries) were found at the Brora saltworks (Fig. 2) in waste heaps from the construction and post-construction levels (Hambly, 2012).These foods are thought to have been opportunistically gathered from the shore by workers involved in the construction and operation of the salt houses.
The first written account of oysters near the Dornoch Firth was obtained from Joan Blaeu's Atlas of Scotland (Blaeu, 1654) from the 1600s which states that "oysters […] and other shellfish are abundant at the river-mouths and sea-cliffs [of the northern Moray Firth]".Blaeu further described the Moray Firth area as being "fished for turbot, ray, dogfish, flounder and also lobsters, oysters and mussels, among other species" (p110).Olsen's 1886 Piscatorial Atlas (Fig. 4; Olsen, 1883) indicates that, in the 19 th century, oyster beds were found along the southern shores of the wider Moray, to the east of Lossiemouth (Figs. 1 and 4) where there are also records of extensive oyster shell deposits (Duff, 1842).There are remains of ancient oyster shell beds in Munlochy, Nigg Bay and near Tain (Miller, 1872).In 1837 in the Bay of Nigg and Cromarty Firths (Fig. 1B), William Macgillivray, a geologist and naturalist, recalled "an oyster scalp above Balintraad and Invergordon" and "…found oysters and about thirty or forty other species of shellfish, all indicating salt water" (Dunbar et al., 1837, pp. 411-413).There are other published accounts of oyster shell in the area, suggesting that dense beds of oysters were present in the Cromarty and Moray Firths (University Marine Biological Station Millport, 2007).
The Statistical Accounts of Scotland (SAS) have one relevant entry for the period 1835-1845 (Vol.2, p 24) under Nigg Parish, where there were: "…few oysters of large size at the Cromarty Firth but they are seldom dredged and do not promise to multiply."There were no records for oyster fishing in the Dornoch or Tain Parishes in either SAS volumes (for 1791 and 1834-1845) or in the annual reports of the Fisheries Board of Scotland (FBS) from 1882 to 1938 but the 1883 FBS report suggested that Dornoch would be well suited to oyster cultivation.However, the Cromarty District which encompasses the southern shores of the Dornoch Firth, is listed as one of three districts contributing to Scotland's oyster fishery in 1883 and 1884 (Anon., 1884) the other two districts being Leith (encompassing the Firth of Forth, see Fig1B) and Ballantrae (encompassing Loch Ryan, on the west coast of Scotland).Thirty thousand oysters were recorded in fisheries landings in the Cromarty District in 1883 and 26 000 in 1884.No oyster landings were recorded in the Cromarty, or adjacent districts, after this time and there is no evidence of industrial processing in this region (that would account for any large shell deposits).Oyster landings have since been largely restricted to the west coast of Scotland with Loch Ryan accounting for most recorded Scottish oyster landings for the period of 1886-1938 and since.
A review of the state of Scottish oyster fisheries published in the 1888 FBS report also notes that between 1876 and 1878, the whole Moray Firth bed, inside Fort George (i.e., Beauly Firth; Fig. 1B), was found to have good quality oysters (Anon., 1882).This oyster bed was identified and subsequently fished to extinction by fishermen from Colchester, England (based ~800 km away), such that by 1885 the beds were too sparse for dredging to be financially viable (Anon., 1882;Young, 1888).
Additional records of live O. edulis from the Moray Firth area were collected in 1879 and 1881 during a dredging expedition in the Cromarty and Inverness Firths (Inverness Scientific Society, 1881, 1879); the specimens from which are now in the Inverness Museum collections.Some live specimens were collected from Findhorn (Moray Firth) in the early 1900s (Bell, 1920) but the most recent verified scientific record of living oysters within the Moray Firth were from the Beauly Firth, where specimens were collected in 1975 (McKay and Smith, 1979).However, there is also a non-specific reference to a hobby fishery in the Moray Firth Partnership Management Plan in 2011 (MFIFG, 2011).In addition to historical accounts, interviews with local naturalists, museum curators and fisheries experts indicated that (to their knowledge) no local oyster fishery has been based in or immediately around the Dornoch Firth in recent times (i.e., the last 50-100 years).
From forty five intertidal survey sites (Figs.1B, C), O. edulis shells were found at 5 locations in the west of the mid Dornoch: emergent from sediment in an area scoured by freshwater run-off from the Tarlogie Spring, and scattered among mussel shell, cobble and pebble in the upper infralittoral zone and along the high tide mark (Figs.1C and 2C).The shells were all intact and in good condition, and the majority had full sculpture.Out of a sample of 39 specimens that were randomly collected, 16 had the hinge ligament intact, with the remains of the periostracum still present in some shells and two with articulated valves (Figs.2D-F).Occasional (1-9 per 100m 2 ) oyster shells were also found on the beach west of Portmahomack (Fig. 1B); these had substantial wear, almost polished in appearance with little remnant sculpture (Fig. 2F).Opportunistic shore surveys (2017) of eleven further sites found 'common' oyster shells along the Cromarty Firth (Cromarty Harbour Bay and Alness Point) and the Beauly Firth (Ardersier) but not Loch Fleet (Fig. 1B).O. edulis shell was also found at nine out of 18 random dive surveys and none of the 34 drop-down video locations undertaken throughout the Dornoch Firth from the Gizzen Briggs to Ardmore (Fig. 1C).Rapid flow, shallow water and poor visibility precluded further SCUBA survey up-stream of this point.Shell abundances ranged from 'Common' (1-9 m -2 ) to 'Rare' (1 per >100 m 2 ) as per SACFOR scales (Joint Nature Conservation Committee, 1997).The five shells that were sub-sampled for radiocarbon dating were found to have calibrated ages of between 4 203 and 8 241 years before present (BP; Table 1).

Discussion
The present study shows that Ostrea edulis have been reported from shell middens in the Dornoch Firth, adjacent Firths, and the wider Moray Firth, from the Mesolithic Period to the Modern Ages (16 th -19 th centuries; Figs.1-4).Furthermore, shell remains are still present throughout the system: some carbon-dated between 4 000 and 8 000 years BP from the sites where they were most common in the present study.Literary sources and fishing records provided historical evidence of live O. edulis present in the same areas as recently as the 19 th century.Therefore, it can be concluded that a long-standing O. edulis population existed in the Dornoch Firth and the encompassing Moray Firth system (Cromarty and Beauly Firths, the coastline from Inverness to Lossiemouth; Figs.1A, B).The present study, however, found only a small number of records of wild O. edulis in north-east Scotland after the late 1800s: Limited fisheries statistics showed that oysters have not been landed on a commercial basis since 1884 and the accompanying reports further indicated that, as elsewhere in Europe and the world, oyster populations appear to have been extirpated (Beck et al., 2011;Cranfield et al., 2004;zu Ermgassen et al., 2012).
The fossil record indicates that oysters were absent in Scotland at the end of the Pliocene, appearing at the beginning of the Pleistocene only to become scarcer again as a result of sea temperature fluctuation and ice abrasion during the glaciations (Duff, 1842;Jamieson, 1866;Tooley and Smith, 2005) and freshwater influxes in the post-glacial period (Bell, 1920).The presence of oyster shell in cultural remains from subsequent Mesolithic and Neolithic human deposits in the Dornoch Firth and neighbouring areas (Figs.1-3; Appendix A Tables A.1 and A.2) suggests that oysters repopulated the marine habitats following the retreat of the glaciers as early as 10 000 years BP.This view is in keeping with archaeological opinion that, because O. edulis shells have been recorded in discard heaps for thousands of years, oysters were relatively common and likely to have been easily gathered from the foreshore nearby (Carver et al., 2016;Holmes, 2017).Shells found in the Tarlogie stream were carbon-dated at over 7 000 years BP (Figs. 1 and 2C-D; Table 2).None of these shells were fossilised, and 16 out of 39 specimens collected had remains of the hinge ligament, which suggested that these oysters might not have been very old.Freshwater can have a large radiocarbon offset from the atmosphere due to the input of ancient carbon from weathering sedimentary rocks in the watershed or remineralisation of terrestrial organic material (Ingram and Southon, 1996).Since the Tarlogie oysters (Table 1) will have been exposed to marine and fresh water influences during their lifetimes it is possible that the age, which is based on a fully marine calibration, is exaggerated (see Ingram and Southon, 1996).However, hinge ligament is known to be well preserved in archaeological sites under certain conditions such as waterlogged environments (Winder, 2017).Therefore, on balance, the age of these shells corroborates the age of shell middens around the Dornoch Firth (Fig. 3) and is more likely to be relatively accurate.Blaeu's (Blaeu, 1654) literary accounts of oysters might be considered problematic because the translation of the Atlas (originally written in Latin) refers, on several occasions, to freshwater mussels as "oysters".However, because oysters are also included in a description of commercial species in the coastal areas of Sutherland, Blaeu's records should not be disregarded.Despite evidence to the contrary in the present study, T.J. Olsen's 1883 maps (Olsen, 1883;Fig. 4) indicate oyster beds around Lossiemouth and not in the wider Moray Firth (i.e., Dornoch, Beauly, Inverness and Cromarty Firths).Nonetheless, Olsen's work was based on consultation with fishermen in SE England, ca.600km to the south of the Moray Firth.Consequently, it might be expected that these relatively small northern water bodies would not have been as well documented as those closer to home ports or to major human conurbations, cf the Firth of Forth (Thurstan et al., 2013).Indeed, from the timing, it seems likely that the Colchester oystermen moved up the coast to the Moray Firth area once stocks in the Firth of Forth became depleted (University Marine Biological Station Millport, 2007).Uncertainties over bed ownership contributed to the failure to regulate fishing in the Moray Firth in the 1880s and these roaming fishermen also appear to have targeted oysters laid by locals (Young, 1888).It is plausible that the exhaustion of the native oyster beds in both the Cromarty and Moray Firths occurred soon after (University Marine Biological Station Millport, 2007) and it seems possible that other small, undocumented (to fisheries managers) O. edulis beds, such as those in the Dornoch Firth, would have suffered the same fate as a result of harvesting, a scenario similar to that reported for C. virginica in the USA by Lenihan and Peterson (1998).
The conservation status of European Special Areas of Conservation (SAC) have commonly been determined against baselines at the time when conservation legislation (e.g., the 1992 EU 'Habitats Directive': 92/43/EEC) was implemented (Johnson, 2008;OSPAR Commission, 2012).However, these baselines do not necessarily represent 'reference conditions', i.e., "a state at which anthropogenic pressures are absent or negligible" (OSPAR Commission, 2011).In Scotland, as in most industrialised countries in Europe, it is highly improbable that examples of pristine marine ecosystems can be found and used as an environmental reference because most areas have experienced some form of decline in ecological condition (Halpern et al., 2008;Thurstan and Roberts, 2010;Thurstan et al., 2013).Even areas currently judged to be in 'Favourable Condition' (as per the EU Habitats Directive), such as the Dornoch Firth have been modified by anthropogenic influences.Therefore, the current conservation status of these SACs probably do not reflect their full ecological potential (Costello, 2014;van Leeuwen et al., 2012).
Across Europe, centuries of heavy fishing have affected large expanses of the seabed (Airoldi and Beck, 2007;Jackson et al., 2001;Worm et al., 2006) and the baseline (see Pauly, 1995) for the management of marine protected areas (including SACs) has undoubtedly shifted in many cases.A shifted baseline might be expected for a more developed system such as the Clyde or the Firth of Forth, where industrial and fishing impacts have been documented (Thurstan and Roberts, 2010;Thurstan et al., 2013).In the Dornoch Firth, however, the history of human pressures and historical conditions have hitherto been opaque at best and might have mistakenly been considered negligible (cf OSPAR Commission, 2011).Examination of historical data and novel survey for O. edulis (this study) has shown that present biological communities in the Dornoch Firth are unlikely to correspond with unimpacted reference communities.
Despite the EU Habitats Directive requiring Member States to "… maintain or restore the natural habitats and the populations of wild fauna and flora at a favourable status", there are few, if any of the ensuing marine protected areas (SACs) where pre1992 [restorative] baselines or targets have been considered.In the wider European maritime context, the Marine Strategy Framework Directive (MSFD; 2008/56/EC) now seeks to promote the sustainable use of seas and the conservation of marine habitats to achieve Good Environmental Status (GES).Biological diversity is one of the "Descriptors" of GES and because O. edulis beds are "special habitats" identified elsewhere under Community (EU Habitats Directive) and International (OSPAR) agreements, they should attract widespread management and monitoring attention under MSFD (Cochrane et al., 2010).The historical presence and extent of key habitats such as O. edulis beds therefore needs to be considered widely in reference conditions and baselines for sustainable management if the aims of the MSFD are to be achieved.
Whilst it is known from the recovery of non-extractive marine reserves throughout the world that there is scope for biodiversity increase from the point of designation (Costello, 2014), extinct or extirpated populations and habitats have received very little, if any, consideration in European Marine Protected Areas.Globally, the literature on habitat restoration has increased two orders of magnitude since 2015 (Nilsson et al., 2016) with many authors considering the practicalities and cost-benefits in forest, riparian and, to a lesser extent in marine (mostly tropical) habitats (e.g., Aradóttir et al., 2013;Crouzeilles et al., 2016;Fariñas-Franco and Roberts, 2014;Mayfield, 2016;Nunes et al., 2016).The restoration of the European oyster has started to attract attention (as did the American oyster, C. virginica, over the last 20 years; Brumbaugh et al., 2006) particularly where remnant populations, well documented historical fisheries and remnant small-scale fisheries have persisted (Gercken and Schmidt, 2014;Sawusdee et al., 2015;Smaal et al., 2015).What the present study illustrates, and what has not yet been considered, is whether there is a widespread case for restoration throughout the European Atlantic MPA network where populations are long since extirpated and records are not immediately available without a detailed interdisciplinary review.
The systematic reintroduction of O. edulis throughout its former distributional range in the European Atlantic Marine Protected Area network is an intriguing concept that would be a shift in conservation baselines and could re-instate benficial ecosystem services (see Coen et al., 2007;Kent et al., 2016).Existing data (Appendix A Table A.3) suggest that present day environmental conditions in the Dornoch Firth are generally suitable for restoration: high water quality, adequate substratum and hydrodynamic conditions and the potential for the site to be a sink for larvae (Appendix A Table A.2).In addition, there is a relatively low risk of physical anthropogenic impacts from fishing, partly because of the aforementioned protected status and legislation relating to the area but also because access from the outer Dornoch Firth is restricted by a military firing range five days per week.Despite the ancient right to fish blue mussels (Mytilus edulis) in the Dornoch, the centre of this fishery has historically been in the eastern mid-section of the site, whereas the centre of oyster shell deposits, and likely area for restoration, are geographically separated to the west of this part of the Dornoch (Fig. 1), in keeping with 'channel beds' formerly described from the subtidal flanks of tidal channels in the Wadden Sea (Hagmeier and Kändler, 1927;Möbius, 1877).
There are other practical challenges in the restoration of oyster populations.Source brood stock for restoration is one potential problem but O. edulis has persisted in aquaculture sites, albeit at low rates of production.Genotype selection and the availability of locally fit stock is probably not one of the challenges, however, because remnant O. edulis populations shows high levels of genetic similarity at the regional sea scale (Vera et al., 2016): not limiting supply options.The introduction of disease agents such as Marteilia refringens and Bonamia ostreae with restorative oyster brood-stock and/or oyster spat movements, however, is a significant consideration (Culloty and Novoa, 1999) and would require stringent biosecurity measures, especially because restorative activities can involve the movement of millions of oysters (Brumbaugh et al., 2008;Schulte et al., 2009).In Scotland, shellfish farms are registered, to help prevent disease outbreaks and allow for the control of damaging species, including those considered to be invasive, under The Aquatic Animal Health (Scotland) Regulations 2009 (see also EU Council Directive 2006/99/EC).The Aquaculture and Fisheries (Scotland) Act 2013 also makes provision for the monitoring and control of movement of species, equipment and water connected with species considered as damaging.Aquaculture movements have long been identified as vectors for invasive nonnative species and disease (Eno et al., 1997) therefore whatever the present regulatory systems are, the scale of restoration programmes would represent a potential step-change the movement of oysters from nursery and on-growing operations and therefore demand levels of biosecurity that are compatible with the conservation management of protected areas objectives in the Dornoch.
Reintroduction of oyster habitat is entirely compatible with contemporary conservation management policy and legislation.The Marine (Scotland) Act 2004, for example, has provision for oyster bed habitats as desired features in new marine protected areas (Tyler-Walters et al., 2012).However, restoration operations could significantly affect SACs and the habitats and species for which they are being managed (against 1992 baselines).Therefore, technically, restoration requires a Habitats Regulation Assessment (HRA) to evaluate impacts against present conservation objectives and conservation interests under the EU Habitats Directive.Designated protected sites that stem from the 1992 EU Habitats Directive, and with a 'sliding baseline' (above), might require their objectives to be revised if, ironically, restoration is to avoid being perceived as a conflict with the present-day protected features.

Conclusions
Overall, this study has shown that the native European oyster (Ostrea edulis) was historically present in the Dornoch Firth SAC and throughout the region and the conditions in the protected area remain within its current potential distributional range.As an area with low anthropogenic pressures and suitable environmental conditions for O. edulis, reintroduction could be successful.European oyster populations have been lost throughout their range (Airoldi and Beck, 2007;zu Ermgassen et al., 2012) but there has recently been a rapid international expansion of MPAs (Costello, 2014) and legislative drivers for sustainable marine management.Interdisciplinary review of protected area baselines will probably show scope for O. edulis restoration in many more protected areas, throughout the European MPA network in addition to areas that support the few remnant populations and fisheries.

Figure 1 -Fig. 1 .
Figure 1-3 are double column fitting.Figure 4 is single column fitting Figure 1 is colour Fig. 1.Study area.(A) Location of North East Scotland; (B) and (C) Map of the Dornoch, Moray and adjacent Firths indicating the main places relevant to the present study.Sites with O. edulis shell or records and all field sites investigated are shown.Letter codes for survey sites (diamonds) indicate SACFOR abundance estimates.Map composed using qGIS v.2.10.1.

Fig. 3 .
Fig. 3. Timeline summarising key O. edulis records from the Dornoch Firth and surrounding areas.Note that archaeological records have not been compiled for Moray Firth and East Scotland regions.*Radiocarbon dates obtained from the present study Fig. 4. Entry for O. edulis in Olsen's 1886 Piscatorial Atlas of the North Sea.Fishermen accounts at the time indicated abundant oyster beds (in orange) in east Scotland along the south shore of the Moray Firth (1), the Firth of Tay (2) and the Firth of Forth (3).