The establishment of networks of Marine Protected Areas ( MPAs) is becoming a widely used approach to protect vulnerable habitats and species and promote resilience in marine ecosystems. European countries are currently working towards a network of marine protected areas under the auspices of the Oslo-Paris Commission ( OSPAR). The components of the OSPAR MPA network are intended to help protect, conserve and restore relevant habitats and species which are, or may be, adversely affected as a result of human activities. While there are various definitions for characterising a network, in the OSPAR context it is characterised by coherence in purpose and by the connectivity between its constituent parts. Connectivity is defined in the present study as the extent to which animal aggregations in different parts of a species range are linked by the exchange of larvae, juveniles or adults (Palumbi, 2003) although in the OSPAR context it also includes dependence of one habitat type on another for structural integrity (Roberts et al., 2003).
Scotland is currently developing its contribution to the OSPAR network of MPAs implemented through the Marine (Scotland) Bill 2010. Under this Bill, 33 Nature Conservation MPA proposals have been identified and proposed to Parliament, whilst a further four potential sites for MPAs remain to be fully assessed. If approved, these Nature Conservation MPAs will help complete an evolving MPA network in Scotland's seas that already includes 46 (with the potential for one more) Special Areas of Conservation, 45 seabird colony Special Protected Areas, 61 Sites of Specific Scientific Interest, and eight fisheries management areas. The Nature Conservation MPAs have been identified for features (the collective term for species, habitats and geology) that currently do not have sufficient protection or are of functional importance to the ecosystem. The Scottish MPA project follows OSPAR advice in considering sub-regions when addressing replication and connectivity. Scottish waters fall into four OSPAR sub-regions: Region I (Arctic waters), Region II (Greater North Sea), Region III (Celtic Seas), and Region V (Wider Atlantic).
Replication of features within and among sub-regions is necessary to spread risk against damaging events and long term change affecting individual MPAs. Risk of local extinction is generally higher in isolated aggregations with low connectivity (Hanski, 2004) and so the number of MPAs within a sub-region needs to reflect the scale of connectivity of species and life stages that are deemed to be priority marine features ( PMFs). MPAs also have the potential to offer a wider ecosystem benefit through the build-up of reproductive mass and spill over of individuals and/or export of offspring. This potential for spill over and export of larvae is, therefore, an important consideration in the location and replication of MPA sites designed to protect PMFs (Palumbi, 2003).
OSPAR accepts that information on connectivity between sites will emerge over time and suggests that in the absence of dispersal data, connectivity may be approximated by ensuring the MPA network is well distributed in space, reflecting the scale of its location. For example, the near-shore is generally dominated by finer scale processes than the offshore, and so MPAs in offshore regions should be larger and further apart than those in near-shore areas. Further, given the variety of PMF species and habitats that are being considered for protection by MPAs, it will never be possible to account for all scales of connectivity among PMF species in siting and replication. Ecological guidance for the Marine Conservation Zones ( MCZ) in England and Wales has largely followed OSPAR guidance in proposing that similar protected habitat should be separated, where possible, by no more than 40-80 km between MPA boundaries. This scale was derived from a simple model of PMF larval transport that focussed on residual tidal flow (Roberts et al., 2010). However, in Scottish waters, evenly distributing MPAs across sub-regions makes little sense because of the diversity of habitats from the deep sea in the far west to inshore fjordic sea lochs, as well as the largely unidirectional large scale circulation patterns in Scottish waters (Turrell, 1992; Figure 1). Therefore, it is important to account for the known patterns and regional variation in current flow regimes in recommending the level of replication within and among sub-regions.
Figure 1: General surface circulation pattern around Scotland. Red arrows are water of Atlantic origin and blue arrows are coastal currents.
Many PMF species are epi-benthic animals that have a planktonic larval phase, but are sessile or have limited mobility following settlement. Hence, an understanding of dispersal of PMF larvae is essential for considering export and connectivity. Hydrographic conditions, interacting with the potential movement (vertical or horizontal) of PMF larvae, determine larval transport, so a biophysical modelling approach can be used to estimate transport from spawning to settlement sites. Such an approach requires output from a hydrodynamic model, as observational data necessary to quantify spatially and temporally resolved three-dimensional currents are virtually impossible to acquire at the relevant broad range of scales, in addition to ecological information such as spawning time, mortality, larval behaviour, planktonic larval duration ( PLD) and settlement time window. Unfortunately, too little is known about the life cycle of PMF species to derive accurate species-specific transport estimates, particularly for the fireworks anemone, white cluster anemone, small brackish water snail and gravel sea cucumber, which were not included in this modelling exercise due to lack of sufficient biological information. Heart sea urchin was not modelled either, as it was only associated with a number of inshore MPAs not resolved by the hydrodynamic model (see below). Nevertheless, in general it was possible to generalize on the probable extent of PMF species transport from available evidence. In the following sections, evidence relevant to connectivity is given for those PMF species that have had the largest influence on the choice of MPAs (Table 1).
Identified invertebrate Priority Marine Features in Scottish territorial waters.
|Phylum||Priority Marine Feature ( PMF)||Species name|
|Cnidaria||Burrowing sea anemone||Arachnanthus sarsi|
|Fireworks anemone||Pachycerianthus multiplicatus|
|Northern sea fan||Swiftia pallida|
|Pink soft coral/sea fingers||Alcyonium hibernicum|
|Tall sea pen||Funiculina quadrangularis|
|White cluster anemone||Parazoanthus anguicomus|
|Mollusca||Fan mussel||Atrina pectinata/fragilis|
|flame shell||Limaria hians|
|Heart cockle||Glossus humanus|
|Horse mussel||Modiolus modiolus|
|Iceland cyprine/Ocean quahog||Arctica islandica|
|Native oyster||Ostrea edulis|
|Small brackish water snail||Hydrobia acuta neglecta|
|Crayfish/spiny lobster||Palinurus elephas|
|Echinodermata||Gravel sea cucumber||Neopentadactyla mixta|
|Heart sea urchin||Brissopsis lyrifera|
|Northern feather star||Leptometra celtica|
A number of cnidarians, vulnerable to towed bottom gears, are important PMF species. The tall sea pen ( Funiculina quadrangularis) is predominantly sessile, although attachment to soft sediments is temporary and so, if disturbed, they can drift into the currents and move location. This species spawns between October and January but the PLD of the plannular larvae is not known. The PLD and settlement competency period of a similar species, Dendronephytha hemprichi, is relatively long (65 days) and the larvae can actively swim (Dahan and Benayahu, 1997). In contrast, the plannular larvae of the northern sea fan, Swiftia pallida are thought to be lecithotrophic with a short pelagic larval duration, suggesting limited potential for larval dispersal (Hiscock et al., 2001). Sea fans are also sessile once settled, with a permanent attachment to the substrate.
The bivalve, Modiolus modiolus is adapted to live partially buried, attaching itself to both soft and hard substratums by byssal threads. Individuals are reported to release gametes throughout the year (Brown and Seed, 1977) with peaks of spawning in spring and early summer (Comely, 1978; Jasim and Brand; 1989), but localised environmental factors, particularly temperature, are exceedingly important in controlling the annual reproductive cycle of this species (Brown, 1984; Seed and Brown, 2004). There are various estimates of PLD for the planktonic veliger stage. For example, under ambient summer water temperatures in Strangford Lough (Northern Ireland), larval duration was found to take approximately 38 days, although a settlement experiment showed that swimming veligers were present in the water column almost two months after initial settling commenced (Roberts et al., 2011). The Ocean Quahog ( Arctica islandica) is a long-lived bivalve often living for more than a 100 years (Witbaard, 1997). Spawning is protracted, and varies with location. The settlement of larvae may occur over several months and is believed to occur throughout the adult distribution ranges. Duration of the larval phase is approximately 55 days post fertilisation for temperatures of 8.5-10°C and 32 days at 13°C (Lutz et al., 1982). Fan mussels ( Atrina fragilis) are burrowing bivalves which have a temporary attachment to the substrate, so dispersal of settled individuals is likely to be very limited (<1m). They have been reported to spawn in the summer although there are no verifiable records regarding spawning times or PLD in the primary literature. The native oyster ( Ostrea edulis) has been found to spawn during the summer months (mid-May to September), coincident with spring tides and the new or full moon (Yonge, 1960; Wilson and Simons, 1985). Reproductive development and spawning is dependent on temperature (Wilson and Simons, 1985), although the exact temperature that illicits spawning is likely to fluctuate with area and local adaptation (Korringa, 1952). After internal fertilization, eggs are incubated for seven to ten days before release into the plankton (Tyler-Walters, 2008a). The larvae are pelagic for 11-30 days (Bierne et al., 1998; Tyler-Walters, 2008a). Flame shell ( Limaria hians) can occur in large aggregations and can swim actively if disturbed, but dispersal by this means is unlikely to be significant compared to the larval stage. Spawning times vary with latitude but in Scottish waters they have been recorded from May-June with peak settlement from July-August (Trigg, 2009).
Maera loveni, is a mud-dwelling infaunal amphipod, which lives in depths of 20-400 m. It is a northern cold water species that has reached its southern limit in Scotland where it is sparsely distributed around the coast. They deposit their eggs within a brood pouch on the underside of the adult female's body. Amphipods have no larval stages; the eggs hatch within a few weeks directly into a juvenile form. Dispersion is limited to crawling, swimming, and "rafting" on algae. The adults are potentially capable of swimming in currents, apparently only doing so if disturbed (Highsmith, 1985), but their dispersal potential is not known. Spiny lobster, Palinurus elephas, spawn one clutch per year from around July to October (Ansell and Robb, 1977; Hunter, 1999). Females incubate the eggs for around nine months with the larvae (phyllosoma) hatching in early summer (Hunter, 1999). The PLD may be very long, one to six months (Mercer, 1973; Marin, 1985). After mating and egg-laying, individuals may undertake migrations to deeper water in Atlantic waters (Ansell and Robb, 1977; Hunter, 1999), although tagging studies in the Mediterranean also indicate that they can remain quite site attached (Follesca et al., 2008).
The northern feather star, Leptometra celtica, is a crinoid echinoderm. Reproduction is via the pinnules, which rupture and release sperm and eggs into the surrounding sea water (Barnes, 1982). The fertilised eggs hatch to release a free-swimming vitellaria larva, which does not feed and only lasts a few days before settlement and metamorphosis into the adult. In another feather star species, larvae settled between two and twelve days after hatching (Kohtsuka and Nakano, 2005). Adult feather stars are usually sedentary, attaching themselves to the substratum (such as sponges or corals) with flexible cirri, but they can crawl and swim by undulating their arms.
It is clear from available accounts that the larval phase will account for nearly all of the dispersal potential of the PMF benthic species and that uncertainties regarding PLD and spawning times will not allow accurate predictions of larval transport to be made. Therefore, in the present study, a relatively simple biophysical modelling approach that accounts for regional oceanographic variation and some degree of biological realism was used, as described below.
For each of the Priority Marine Feature species, we will present maps of the distribution of particles representing individuals at the end of their larval phase ( PLD) and maps showing presence of larvae during their settlement window over any proposed MPA visited by those larvae released from those relevant MPAs (see Methods section). In both cases, when spawning takes place over more than one "season", season-specific maps will be presented. We will also present colour matrices that show the relative connectivity between origin and destination MPAs, based on the percentage of all particles released at each origin MPA. These results have been obtained for all spawning seasons individually, in the case of species that spawn in more than a single season, but only results combined over all seasons will be shown here. Finally, as described in the Methods section below, we will provide Tables with summary statistics (mean and standard deviation) of the dispersion distance from each origin MPA, for each spawning season (distance in km along a direct line between the particle start and end positions). Some additional analysis of general patterns emerging from these results will be presented at the end of the Results section.