Biogenic habitats

Results of review of literature for assessment of habitat loss

Seagrass beds

Outer Hebrides

The Sound of Barra hosts extensive areas of seagrass beds. Harries et al. 2007 provide a coverage estimate of 360 ha based on 2006 survey work, of which 260 ha is mapped within the NatureScot GeMS database. Based on analysis of satellite imagery underpinned by ground-truthing an overall trend of thinning of the beds in the Sound of Eriskay was identified by Harries et al. 2007 between 2001 and 2006, although increase in seagrass density was noted locally. Hydrological changes resulting from the construction of the Eriskay causeway were identified as a possible driver of the change, although natural temporal fluctuation in density may also have been a significant factor.

Recent analysis of satellite imagery from 2005 to 2017 reveals that the process of eelgrass thinning is continuing (Figure a). Plant loss is particularly marked to the east of the Eriskay causeway (bed A in Figure a). Some redistribution of eelgrass appears to have taken place to the west of the causeway, although overall density appears lower, particularly within bed B (Figure a). Drivers for seagrass decline clearly extend beyond causeway construction as marked thinning and contraction of beds can also be identified up to 8 km to the south-west off Fiaray (Figure a) and Fuday. No quantitative assessments of eelgrass loss have been carried out but the visual impression from the satellite imagery is that several of the major beds have possibly lost, in the order of, 50% of their cover between the years 2001 and 2017. This figure includes both thinning and peripheral contraction of beds.

Figure a: Satellite imagery of the Sound of Eriskay (upper) and Sound of Fiaray (lower) in the Sound of Barra, with conspicuous seagrass patterning delimited by white lines. Main images from May 2017, with insets showing appearance of seagrass beds in January 2005. Seagrass cover is normally lower in winter time. © 2019 Google, Maxar Technologies and Getmapping plc.

Argyll

Complete eelgrass bed loss is known to have taken place recently in Loch Creran. Smith (1887) described the presence of a green sloping bank of Zostera marina extending for at least 'several hundred yards' in 1883, although its location within the loch was not stipulated. In 1989 Connor (1990) recorded a dense bed of Z. marina at around 1 m above chart datum along the northern shoreline of the lower basin. Moore et al. (1998) identified three small beds in 1998 along the northern shoreline of the lower basin, one of which is likely to have been the bed recorded by Connor (1990). The spatial extent of the beds was not measured, although it is probable that total extent was <1 ha. The last known record of Zostera marina in the loch is believed to be in 2002, the most developed bed of the three recorded by Moore et al. (1998). Subsequent surveys of the loch for eelgrass, the most recent in 2019, have failed to record any evidence of the continued presence of the species in the loch.

Unfortunately, the scale of eelgrass distribution in Argyll is quantitatively unknown. The GeMS database only includes coverage data for one bed amongst the Treshnish Islands; however, there is extensive presence of the habitat around Islay and in Loch Sween and Loch Craignish.

Other regions

Subtidal seagrass beds are known from all the other SMRs apart from North Coast and North East. Little is known about their extent (apart from Clyde) and data on temporal trends are insufficient to form a judgement on current status.

Flame shell beds

Clyde

Hall-Spencer & Moore (2000) reviewed the evidence for a decline in flame shell populations in the Clyde Sea area over the previous 40 years and implicated scallop dredging as a likely causative factor in their disappearance from a number of locations. Since 2011 the only bed to have been examined in the Clyde SMR is that at Otter Spit in Loch Fyne (Moore et al., 2013b). This may represent the only remaining bed in the Clyde area. A record of a bed south of Inchmarnock Island (GeMS database derived from Allen, Axelsson, Dewey & Clark, 2013) should be regarded as uncertain, as the record only relates to the finding of several flame shell nests which collectively represented nest coverage of the sea bed of <1%. In 2012 the extent of the Otter Spit bed was estimated to be 50 ha (Moore et al., 2013), however, its northern limit was found to have retreated from part of its former range. Based on the distribution of former records, bed loss is estimated at around 5 ha minimum (see Moore et al., 2013b). MPA monitoring here in 2015 indicated that no further habitat loss had taken place. Hall-Spencer & Moore (2000) report observing damage to the bed here by scallop dredging.

Argyll

The presence of a flame shell bed off Port Appin in Loch Linnhe, Argyll was first recorded in 1988 (Connor, 1989) and its distribution mapped by Moore, Harries & Trigg (2012) based on spot dive and diver transect records over the period 2001 - 2006 (Figure b). The bed covered an area of the order of 40 ha, although with some localised bare patches. As part of a broader programme for assessment of the distribution and condition of selected Priority Marine Features the bed was resurveyed in 2011 and was found to be extinct over most of its former range (Moore et al., 2012). It was restricted to three relict patches totalling 3 ha. A subsequent 2015 survey (Cook, 2016) failed to record the habitat on two of the three patches and this was confirmed by a 2018 survey (Heriot-Watt University, unpublished), which failed to observe the habitat on any of the three patches. The habitat was recorded to the north of its previously known distributional limit, although subsequent examination of the area suggests that extinction may now have occurred between 2018 and 2019.

Figure b: Contraction of flame shell bed at Port Appin, Loch Linnhe, showing historical extent of bed based on pre-2011 records and following a detailed 2011 survey (Moore et al., 2012). Point records also shown for post-2012 surveys.

The same 2011 survey (Moore et al., 2012) carried out a comprehensive examination of the west coast of Shuna Island in order to validate a 1989 flame shell bed record (Connor (1989)). No evidence of the continued existence of the bed was found.

Flame shell beds were first mapped in Loch Creran in 2012 (Moore et al., 2013b) occupying a total area of 18.5 ha and split between two locations. However, as part of condition assessment work in 2018 (Moore et al., 2019) an additional bed was located and the measure of total extent of the habitat in the loch increased to 35 ha. The habitat was found to be in good condition; however, the increase in extent was likely to be due to the increase in spatial spread of survey sites rather than expansion of the habitat.

West Highlands

As in the case of Loch Creran measures of bed extent in Loch Alsh have increased from 75 ha in 2012 (Moore et al., 2013b) to 105 ha in 2016 (Moore et al., 2018) probably due to wider sample coverage rather than habitat expansion. Survey work in Loch Carron in 2017, following the damage to a flame shell bed by scallop dredging, recorded a total coverage of the habitat of 214 ha, with an estimated 0.4% (0.9 ha) exhibiting evidence of degradation (Moore et al., 2018). The flame shell population was able to persist in much of the damaged areas and particularly where the bed was well-developed, and it is considered that total habitat loss in Loch Carron is unlikely to be greater than 1%, although possibly significantly greater locally.

Since 2011 potential changes in other West Highland beds have been recorded in the Sound of Mull and off Wester Ross. A bed off Rubha an Ridire at the eastern entrance to the Sound of Mull was recorded in 1989 (Connor, 1989) but re-examination of the location in 2011 failed to record it (Moore et al., 2012). This was noted as being a small bed and so may have been missed. It appears likely that there has been some contraction of the small bed at Sruth Lagaidh Narrows in Loch Broom prior to 2011 (Moore, Harries, Trigg, Porter & Lyndon, 2011). Bed extent was assessed in 2010 as 7 ha. A survey of the site in 2017 found no firm evidence for a temporal change in bed extent since 2010, although there was an apparent improvement in habitat condition, in the form of an increase in the thickness and luxuriance of the byssal turf, which overtopped the bound pebbles and shells in 2017, but not in 2010 (Moore, 2019).

Recent SNH and Marine Scotland dropdown video surveys (2018 & 2019, unpublished) have identified the presence of several previously unrecorded flame shell beds in the region of the Inner Sound to the west and north of Loch Carron. The surveys were undertaken as part of the EMFF Project Engaging the Fishing Industry in Marine Environmental Survey and Monitoring.

Orkney Islands

Flame shell beds have been recently recorded at several locations in Scapa Flow (Hood, 2016; unpublished 2019 SNH/MS EMFF dropdown video survey). However, their extent, as well as temporal patterns in condition, are currently unknown.

Other regions

Flame shell beds are unrecorded in all other regions.

Serpulid aggregations

Argyll

In Loch Creran the serpulid worm tube aggregations form bushy, well-developed reefs up to 75 cm in height and 1-2 m in diameter. The extent of the serpulid reef habitat was assessed in 2005 as 108 ha (Moore et al., 2006). Most of the habitat appeared in good condition, albeit there was localised damage from demersal fishing, finfish and mussel farming, vessel moorings and organic pollution, although cessation of the organic discharge had led to some recovery of the habitat. Condition monitoring in 2017 to 19 based on diver, sidescan and single beam sonar surveying revealed a marked deterioration in the status of the habitat (Moore et al., 2019). There was widespread fragmentation of the reefs with loss of most of the larger structures (Figure c). A reef condition metric based on measures of reef canopy height, percentage of broken reef, tube occupancy by Serpula vermicularis, and the presence of pristine reefs, was determined from 40 diver transects passing through the habitat. The score ranged from 1 (pristine) to 8 (scattered broken reef rubble fragments supporting few serpulid worms). The mean score was 5.8 with 58% of transects providing the lowest scores of 7 - 8 (Figure d). No equivalent data are available for the 2005 survey, but based on considerable experience of the area, the mean score is likely to have been within the range 1 - 2. Thirty-five per cent of the transects in 2017 to 19 produced condition scores of 8 and this can be considered indicative of the scale of habitat loss.

Figure c: Screen grabs from video taken along the same transect in Loch Creran in 2005 (left) and 2014 (right) showing temporal degradation of the serpulid habitat.

Figure d: Frequency of Loch Creran diver survey transects examined in 2017 - 2019 allocated to serpulid reef condition scores (1 = undegraded, 8 = largely unpopulated reef rubble).

The cause of habitat deterioration is uncertain; however, it is considered most likely that it is part of a long-term natural cycle of degradation without compensatory replenishment (see Moore et al. (2019) for further details).

West Highlands

In Loch Teacuis, an arm of Loch Sunart, serpulid reefs were first recorded in 2006 (Mercer, Howson & Moore, 2007), with a habitat coverage of 20 ha, within which reef coverage was around 4% (Dodd, Baxter & Hughes, 2009). By 2015 living reefs had become largely extinct, although scattered reef fragments, largely unoccupied by worms, were still present (Kamphausen, 2015). The cause of loss is unknown. The action of storms on reef structures, perhaps exacerbated by the presence of colonising kelp plants, was suggested as a possible factor (Kamphausen, 2015).

Recent SNH survey work has revealed small clusters of Serpula vermicularis of up to around 50 tubes on cobbles and boulders in Loch Ailort (Moore, 2019). No temporal data are available for this site and the extent of the feature is unknown, although it is likely to be small.

Other regions

Serpulid aggregations are unrecorded in all other regions.

Horse mussel beds

Knowledge of the spatial extent of Scottish horse mussel beds is insufficient to assess the quantitative significance of recorded habitat degradation. Many known beds are not delimited within the GeMS database. There are several recorded examples of loss of extent or condition of beds.

Argyll

In the Argyll region declines in horse mussel density were recorded on three beds in Loch Creran (with also extent reduction at one of them) between 2005 (Moore et al., 2006) and 2018 (Moore et al., 2019). The scale of bed loss is unknown, although the previous extent of the bed is likely to have been small, based on the limited sampling intensity in 2005 (Moore et al., 2006).

West Highland

There is no firm evidence that the areal extent of the habitat has declined in the West Highland region over the assessment period. There has been a steady temporal decline in density of a bed in Loch Alsh. Measurements of mean percentage cover of the seabed by live horse mussel shells have reduced from 29% in 1999 (Mair, Moore, Kingston & Harries, 2000) to 16% in 2004 (Emu Ltd., 2006), 11% in 2007 (Marine Bio-images, 2007) and 7% in 2012 (Moore et al., 2013b). In the same region, there is also an indication (though not backed by quantitative evidence) that horse mussel density has declined in Strome Narrows in Loch Carron (Moore et al., 2018). In both Loch Alsh and Loch Carron it is possible that competition with flame shells has played a role in the reduction in horse mussel density.

Moray Firth

MPA monitoring of the largest horse mussel bed in Scottish waters off Noss Head found the distribution of the bed in 2017 to be consistent with earlier mapping of the feature based on records from 2010 to 2013 (Scottish Natural Heritage, 2014; Moore, 2019).

Other regions

Horse mussel beds are also found in Forth and Tay, Orkney Islands, Shetland Isles, Outer Hebrides and Clyde. No data on temporal extent trends are available.

Blue mussel beds

Moray Firth

Both intertidal and subtidal mussel beds are present in the Dornoch Firth and have supported a fishery for hundreds of years. In 2010 scattered subtidal beds occupied an area of around 218 ha (Bromham, 2010), with mapping provided by Cook, Cordingley, Woolmer & Sanderson, 2016) (Figure d). A subsequent survey in 2015-16 (Cook et al., 2016) recorded only two relict patches occupying a total area of 1 ha (Figure e). The loss of habitat was considered to possibly involve factors such as poor recruitment success and increased levels of post settlement mortality.

Figure e: Distribution of subtidal blue mussel beds in the Dornoch Firth in 2010 and 2016.

Other regions

Subtidal blue mussel beds are known from all other regions apart from Orkney Islands, North Coast and North East. Extent data are available for the large Tay bed (Forth and Tay Region) within the GeMS database, but no temporal trend data are available for any of the beds.

Maerl beds

Habitat loss

Due to a lack of historical and current measures of maerl bed extent it is difficult to assess the degree of habitat loss in Scottish waters. Inverse correlations between fishing intensity and maerl distribution around Arran (Clyde) (Moore, 2014a) and the Summer Isles (West Highlands) (Moore, 2014b) are indicative of habitat loss, and deterioration of the habitat is known to have taken place in Loch Ewe (West Highlands) (Moore, 2014b). Quantitative assessment of the current level of habitat degradation is generally impossible. However, in the Clyde region a 17.5 ha bed last recorded in 1999 (Hall-Spencer & Moore, 2000) was found to have disappeared by 2012 (Moore et al., 2013b), representing around 9% of the known extent of the habitat in the Clyde region. Of the three mapped beds in Argyll, temporal distributional data are only available for Taynish Rapids in Loch Sween. A bed extent of 2.7 ha was measured in 2012 (Moore et al., 2013a). Digitisation of a sketch of this bed following a 1984 survey by Lumb (1986) indicates coverage of 1.8 ha. No temporal comparisons of extent can be made for the remaining Scottish Marine Regions.

Habitat condition

Figure f shows Azti Marine Biotic Index (AMBI) scores for 29 Scottish maerl beds from six of the Scottish Marine Regions for which there are infaunal core sample data since 2011 and includes 12 sites where earlier data are also available. All sites fall within the good or high ecological status categories, indicating the presence of undisturbed or slightly disturbed communities. As there is uncertainty regarding the accuracy of the category boundaries for these maerl habitats, all sites are best regarded as simply showing little or no evidence of disturbance.

For the 12 sites where temporal trend data are available (West Highlands and Outer Hebrides), the mean AMBI is effectively static (1.47 pre-2011, 1.42 post-2011). The highest AMBI values relate to sites where anthropogenic disturbance is unexpected. One temporal trend that may be linked to human disturbance is the marked drop in AMBI (hence drop in disturbance) at site ML02 in Loch Laxford (Figure f). The bed is adjacent to a mussel farm, presumably responsible for the high density of mussel shells and dominance by the pollution-tolerant polychaete, Pseudopolydora paucibranchiata, in 2009 (Moore, Harries, Porter & Lyndon, 2010) but unrecorded in 2015 (Moore et al., 2017).

Figure f: Mean Azti Marine Biotic Index (AMBI) scores for infauna of 29 maerl beds sampled from 2011, with additional historical data for 12 of them. Ecological status categories of water bodies for the Water Framework Directive also shown (from Muxika, Borja & Bonne, 2005).

Extent assessments for all habitats

The baseline extent estimates for each habitat type are shown in Table a and losses recorded in 2011 to 2018 in Table b. The baseline estimates were derived from historical studies spanning a broad time period, but all predate the measures of extent recorded between 2011 and 2018 that have been used to assess habitat loss. The extent loss indicator was not used for the 2011 Atlas (Baxter et at., 2011) but as losses are assessed against historical baseline values, they can be considered to indicate temporal trends.

The AMBI scores have not been integrated with the habitat loss results to form a multi-metric indicator, but this would not have influenced the assessment outcomes. The scores do, however, indicate that the condition of the maerl beds examined from the six regions do not provide significant cause for concern, insofar as habitat condition is characterised by the infaunal community.