First steps in the ESE assessment were listing human activities and associated stakeholder groups in the region. We identified 29 stakeholders from science, governance (federal state ministry, state agency, authority and district representatives as well as local mayor), tourism, fisheries and nature protection NGOs. Additionally, we carried out an institutional mapping to get an overview about institutions related to lagoon and water quality issues. Further, we listed main ecosystem goods and services as well as economic drivers and identified relevant social and economic components. During the first stakeholder workshop on 21. Oct. 2015, the scientists introduced zebra mussel farming as potential measure for improving water transparency and the ecological status of the lagoon. After a discussion, the group agreed to pick up this issue for a detailed study.
In the Oder Lagoon region, water quality aspects were already addressed in earlier studies. Previous iterative attempts and the present approach can be visualized using the DPSIR method (Fig. 2). The earlier attempts showed that it is hard to tackle large areas with complex problems and a large amount of potential stakeholders and to come to concrete results in a limited time. Therefore, the present approach focused spatially on a relatively small area of about 150 km2, the south-western part of the Kleines Haff, including several coastal towns and the city of Ueckemünde. Additionally the focus shifted from environmental problems and possible solutions towards a local utilization of the lagoon, addressing aspects of the EU Blue Growth strategy (European Commission 2012). The 3rd DPSIR analysis of this study and its outcome (Fig. 2) remained controversial within the research group and other alternative DPSIR cycles existed, but the approach helped to guide discussions and to develop a joint view about the issue.
System design and scenarios
Schadach (2013) classified areas, which are suitable for mussel farming, based on 12 criteria (shipping routes, dredging areas, recreational boating, harbors, bathing sites, camping sites, nature protection core areas, fishery harbors, location of devices, sediments, water depth and flow velocity). Four priority area classes where distinguished, where 1 indicated most suitable areas and 4 least suitable areas (Fig. 3). This map focused on areas suitable for commercial farming. During our discussion process, we defined three alternative spatially explicit and realistic scenarios for mussel farming. Consequently, we introduced additional criteria for all scenarios, like a) sandy bottoms, b) sufficient flow velocity to avoid long-term accumulation of faeces and pseudo-faeces on the sediment surface and ensure sufficient phytoplankton transport to nourish mussels, c) historic mussel beds in the surrounding to ensure a high likelihood of mussel larvae in the water and natural settling on farms as well as d) no legal or spatial planning restrictions and e) no risk of pollution (organic contaminants, heavy metals).
Scenario 1 - commercial mussel farm
Objective was to establish a mussel farm that removes nutrients, increases water transparency, provides feed (fresh mussels and meal) in the most efficient way and can be maintained as a profitable business (Fig. 4). Additional pre-conditions were high phytoplankton concentrations as well as a low risk of damages due to drifting ice. We assumed 15 farm units of 2.25 ha each, covering an area of 34 ha and a harvested production of 1000 t of mussels per year. Further, we assumed a mussel biomass and annual yield of 1.5 kg/m3 water (bi-annual harvesting), as well as cultivation ropes down to 2 m in a water depth between 3 and 5 m. The general setup is shown in Fig. 11.
Scenario 2 - beach mussel farm to support bathing-tourism
Objective was to locate a mussel farm left and right of a beach, so that the prevailing coast-parallel currents flow through the farm and the transported waters are becoming more transparent. As a result, the bathing area would have an increased water transparency and would become more attractive for tourists. At the same time, near the mussel farm the emerged and submerged macrophyte areas would benefit from improved transparency, spread, stabilize the sediments and in turn further increase water transparency (Fig. 4). Present experiences show that macrophytes will not spread into the bathing area, because of mechanical damage by bathers. Pre-condition is a production density that excludes any negative impacts on water body and sediment (low oxygen concentrations) and locations with stable coast-parallel current pattern. We assumed a mussel farm producing 1500 t of mussels per year covering an area of 18 ha, with a mussel biomass of 5 kg/m3 water as well as cultivation ropes down to 1.7 m in a water depth between 2 and 3 m. Since mussels are harvested only after 2 years and in late autumn, the total biomass actively involved in filtration can be up to 3000 t.
Scenario 3 – Environmental mussel farm for nature restoration
Objective was to establish a mussel farm that induces a self-reinforcing cycle with increasing environmental quality, nature restoration and supports the implementation of the Water Framework Directive (Fig. 4). Idea was that increasing local water transparency enables the extension and restoration of submerged macrophyte belts, which in turn stabilize the sediments and further increase water transparency. Assumption was to establish several non-intensive (1.5 kg/m2 mussel biomass) farms covering a total area of 1.12 km2, spread over an area of about 4 km2 in shallow waters (< 2.5 m) near the south-eastern coast of the Kleines Haff. Very likely, these areas were still covered by macrophytes a century ago.
The system design step in the ESE includes the definition of administrative and virtual system boundaries, the development of scenarios, and the development of a conceptual model. The conceptual model links the simulation models with the scenarios and the framework conditions (Fig. 5). Its development was an iterative process including the identification of the state variables and processes needed to address the issue on one side and the availability of data, modelling and financial resources as well as expertise on the other side.
The next ESE steps were ‘system formulation’ (Table 1) and ‘system assessment’. The latter provided concrete model simulations, assessments and visualizations for each scenario as basis for stakeholder discussions.
System Formulation & Assessment: The commercial mussel farm
Our commercial scenario assumed that zebra mussels have a similar nutrient content and composition compared to blue mussels and can serve as an adequate substitute. Our comparative laboratory nutrition analysis of the two samples confirmed this. Zebra mussels (blue mussels) contained 8.4% (7.9%) protein in fresh weight, 0.8% (1.5%) fat, 0.14% (0.12%) phosphorus, 1.3% (1.3%) raw ash and 89.3% (84.1%) water. The sum of saturated fatty acids is 35.3% (29.2%), the sum of single unsaturated fatty acids 35.8% (25.1%) and sum of multiple unsaturated fat acids 28.6% (45.7%). In general, both species showed a comparable quality as feed. Some of the differences may be natural variability and may result from different locations, growing conditions and water contents. Our literature survey (Drews 2012) and recent results, obtained within DBU-project EBAMA, suggest that Baltic blue-mussel meal seem to be a very good substitute for fishmeal, this is especially true with respect to important ingredients, like vitamin E or fatty acid quantity and composition (Krost, pers. com.).
Revenue: In 2014, the total annual costs of the Zoo Osnabrück for altogether 68 t of protein-rich animal feed were about 118,000 € (e.g. fish, meat and living animals). For protein rich feed, the zoo paid an average price of 1.67 €/kg. In 2014, about 37,000 €/a were spent to buy altogether 28 t feed fish (sprat, smelt, roach, herring and mackerel). The zoo feeding experiments showed that mongooses (Mungos mungo) and the oriental small-clawed otters (Aonyx cinerea) immediately accepted zebra mussels as food, raccoons (Procyon lotor) even showed preference for zebra mussels and Arctic foxes (Vulpes lagopus) accepted it with some reluctance. The results allowed the extrapolation that fresh zebra mussels can substitute about 10% of all protein-rich feed in the zoo at costs of 11,270 €/a. The annual consumption could be 6.7 t in the entire zoo at acceptable costs of 1.68 €/kg. Fresh mussels turned out to have, compared to other feed, the important benefit that they occupy zoo-animals. Beyond a transportation distance of 1200 km, they exceed the value of the mussels and the revenue became zero (Fig. 6a).
Altogether 254 animal parks and zoos were identified within 300 km around the lagoon. The potential demand of fresh mussels estimated as 1.4 t/a (within 10 km), 4.5 t/a (10–50 km), 18.5 t/a (50–100 km), 94 t/a (100–150 km), 26 t/a (150–200 km), 58 t/a (200–250 km), 137 t/a (250–300 km). Altogether we calculated a potential market of 340 t/a within 300 km. The large zoos in Hamburg (31 t/a) and Berlin (48 t/a) were identified as most important single potential customers (Fig. 6a). The distance-dependent sales volume and the transportation combined gave an idea, how much fresh mussel biomass could be sold at a certain price. Above about 5000 tons, the revenue became negative (Fig. 6b).
In 2014, altogether 7841 t of aquaculture fish were produced in 32 closed and 40 open fish aquaculture systems within 300 km from the Oder Lagoon. 38% of the fishes were carp species and 17% trout species. Other important fish species were catfish, pikeperch and sturgeon. Most species are known to feed on mussels, at least temporary (Schadach 2013). Compared to zoological gardens the sale to and potential revenue from fish aquaculture is low. Existing protein-rich feed, like fishmeal, is still relatively cheap (below 1500 €/metric ton in 2017/2018; https://datacatalog.worldbank.org) and largely prevents the usage of fresh mussels as alternative.
The revenue functions did not consider the productions costs. Fig. 6c shows the production costs in comparison. The production costs are always higher than the possible revenue from sale. The cost and revenue functions meet at a production volume of 800 tons at a price of about 830 Euros/t. However, we have to assume that not the total production has a size and quality that is suitable for zoos. In our calculations, we did not take into that not the entire production but possibly only 80% can be sold at this price. A mussel production for fish meal above 5000 tons would be profitable even without other sale options (Fig. 6c). The compensation fee for removing nutrients alone would be able to cover the farming costs. It means that mussel farming is a cost-effective measure to remove nutrients. Therefore, it could receive its funding as a measure within the WFD.
Even in deeper water and with vertical cultivation ropes of 2 m length, a production of 5000 tons mussels would require about 1.7 km2 mussel cultivation area. In our approach, assuming farm units of 2.25 ha, about 74 units would be required. The units would have to be spread over the lagoon to ensure a sufficient nourishment. This seems not realistic.
The survey among 21 Austrian fish aquaculture companies showed that they would accept mussel meal as a replacement for fishmeal, preferably without shells. Because of the high acceptance, the similarities between fish and mussel meal and the possibility to store it, the market is practically not limited. Further, the surveyed feed producers considered mussel meal as highly interesting as long as a sufficient and continuous supply of mussels for the production could be ensured. The companies considered trading fresh mussels on a commercial large-scale basis as not realistic.
Independently from the produced amount, the production of zebra mussels seems to be no suitable business model. However, as soon as mussel farmers get a compensation for the removal of nutrients, the production of more than 100 tons would become profitable. This is shown in Fig. 6c (dotted red line), where the compensation for nutrient removal is included in the cost function.
System Formulation & Assessment: The beach mussel farm
The beach mussel scenario required reliable flow and transport model simulations. A comparison between data collected with an Acoustic Doppler Current Profiler, floating on the water surface, and model simulations (depth averaged data, integrated over 10 min) showed a very good agreement outside the coastal wind shelter, namely 1 km and more off the shoreline. Figure 7a-c presents near-shore flow data compared to model simulations for 3 different days with offshore wind situations. In general, the agreement between data and simulated flow velocities is good, but the direction sometimes differs. Wind shelter due to coastal vegetation was not taken into account by the model. Further, the floating current profiler was not able to collect data for the first few decimeters near the water surface. Therefore, in shallow waters the depth-averaged flow data might not be reliable. Despite these uncertainties, the model simulations seemed sufficiently reliable for our purpose.
During summer (May-Sept.), close to Ueckermünde beach, the model suggested an average flow parallel to the coast, mainly towards south-east, and dominating slow current velocities below 3 cm/s (Fig. 7d). Relatively steady coast-parallel slow current velocities were a suitable precondition for an effective water filtration by mussels and a potentially strong positive effect on water transparency in front of the beach.
In summer, the artificial sandy beach of Ueckermünde (800 m length and about 43,500 m2 area) is intensively used by visitors (Fig. 8a). The model simulation assumed up to 3000 tons of mussels, distributed beside both sides and in front of the beach (Fig. 8b). Because of mussel filtration activity, the model suggested an improved water transparency (Secchi depth) of up to 0.4–0.5 m (Fig. 8c). This means instead of 0.5–0.6 m the summerly Secchi depth would be about 1.0 m.
The model suggested that these 3000 tons of mussels cultivated in a relatively high density in a shallow area with a water depth below 2.5 m would cause about 20 days with oxygen concentrations below 1 mg/l near the bottom (Fig. 9a). Since this would be a risk for the environment, especially the benthic flora and fauna, and maybe counterproductive with respect to fostering bathing tourism, the cultivated mussel biomass had to be reduced to an acceptable level. According to the model, a mussel biomass of 1500 tons would avoid negative effects on bottom oxygen concentrations and would be environmentally sustainable, but would increase Secchi depth only by 0.19 m (Fig. 9a).
Although the spatial resolution of the ecological model was compared to earlier model applications refined to 150 m grid cells, the model is not able to represent the coastline smoothly. This means it does not allow to optimize the allocation of the mussel farms. Therefore, we can assume that in reality a spatially optimized farming could generate the predicted effects with a lower mussel biomass.
In 2016, about 121,000 tourist overnight stays were recorded for Ueckermünde (Statistisches Jahrbuch 2017, https://www.laiv-mv.de). Taking into account tourists overnight stays in the surrounding of the city and in private accommodations, as well as day tourists, we estimated a number of 300,000 tourist days per year. To cover the costs for mussel farming around the beach, it was assumed that a fee is added to the tourist tax and for day visitors is added to the parking fee at the beach. Ueckermünde beach has a large parking area of about 1.5 ha. According to an empirical survey, tourists are willing to spend 1 € per day additionally for an improved water transparency of 1 m (Hirschfeld, pers. com.). Based on these assumptions, an improved water transparency by 0.19 m could generate about 57,000 € per year for supporting a 1500 t mussel farm.
Further, 25% of the tourists stated that they would come more often if the water transparency (Secchi depth) would be 1 m better (Hirschfeld, pers. com.). On average, overnight tourists spend 73 € per day in Mecklenburg-Vorpommern including overnight costs and daily expenses (Statistisches Jahrbuch 2017, https://www.laiv-mv.de). Using this value and assuming a 1 m better water transparency, the increased number of tourists would generate an additional income of about 5.4 mill €/a. Assuming that 50% of the 7% added value tax for overnight stays is used to support mussel farming this would generate an additional income for mussel farmers of about 190,000 €. Based on these calculations, but assuming an increased water transparency of only 0.19 m, a 1500 t mussel farm could be supported with 36,000 €/a additionally.
The accumulated revenue and the costs (per t of mussels) depending on the cultivated mussel biomass (standing stock) are shown in Fig. 9b. An environmental fee together with additional income from increased tourism is not able to cover the costs for a mussel farm. Adding a financial compensation for removing nutrients to the other two source of income would result in annual revenues of altogether 309 €/t and costs of 178 €/t mussels, assuming a 1500 t mussel farm. In this case a farmer would make a profit of 131 €/t mussels per year. Theoretically, already a small farm with a production of 300 t could make a profit. The cost function only includes running costs for maintenance, operation and labor. Including investment and capital cost, the total costs would be about 20% higher. However, all calculations are hypothetical, depend on several assumption and include simplifications. The provided numbers can hardly be considered as reliable, but served as starting point for discussions with stakeholders.
System formulation & assessment: The environmental mussel farm
In the EU Water Framework Directive (European Commission 2000), macrophytes serve as biological quality elements for surface waters. An improved status of macrophytes of the lagoon, is an environmental policy objective, because it would indicate an improved ecological status. Scenario 3 was sub-divided and altogether three model simulations were carried out, with a total mussel biomass of 540 t, 1621 t and 5065 t (Fig. 10a). With 1.5 kg mussels per m2 water surface, the average mussel density was low. As consequence, the farms covered large areas of 0.36, 1.1 and 3.5 km2. The increase in Secchi depth was 4.6 cm or 7.2% (summer average over 4 simulated years), considering the large farm. The available light above the sediment would increase by 45%. Because of low mussel densities, all farm-sizes would not affect the oxygen conditions above the sediment. The number of days with oxygen concentrations below 1 mg/l in average would remain below 1 day per year. Fig. 10b shows that the area with improved growing conditions for macrophytes is about 5 times larger than the area covered by mussels. Prevailing currents transport the transparent water along the shoreline and therefore into areas where macrophytes potentially could grow. The cost function is comparable to the beach scenario (Fig. 9b). The running costs of all assumed farms could potentially be covered by a financial compensation for nutrient removal.
Two major questions remained: would macrophytes recover and re-settle areas with higher light availability at the bottom without additional supporting measures? If yes, what are the critical light conditions to initiate growth and a spatial spread? Our field surveys and a literature study proves that different submerse macrophyte species are still present in the western lagoon and single macrophyte stands were found in a water depth of up to 1.8 m. This is true for the eastern, Polish part as well (Brzeska et al. 2015). Recent studies by Nowak et al. (2008) and Blindow et al. (2016) documented that germinable diaspores of several species are present in the sediments of all observed German Baltic coastal water. Nowak et al. concluded that diaspores have the potential to restore macrophyte communities. This can happen even decades after the stands were lost. The average nutrient levels in the Kleines Haff (western Oder Lagoon) have declined from about 250 μmol/l total nitrogen (9 μmol/l total phosphorus) in the late 1980s to about 100 μmol/l total nitrogen (5 μmol/l total phosphorus) between 2010 and 2015. However, this had no significant effects on water transparency (about 0.6 m Secchi depth, summer average 2010–2015) and chl-a concentration (about 70 μg/l summer average 2010–2015). Obviously, the lagoon shows a hysteresis effect and does not react to nutrient load reductions. There seems to be a potential for an improvement and the restoration of macrophyte stands. A good likelihood exists, that mussel farms could initiate a local restoration and can be considered as a supportive measure in the WFD. However, a concrete experimental farm is required to collect information on the required light conditions at the bottom, recovery behavior, and additional side effects of the farm e.g. on currents, turbulence, resuspension and shading.
System Assessment & Implementation
On 9. Sept. 2017, on a second workshop with 18 local and regional stakeholders the three scenarios and the results of the economic and ecological modelling were presented and the computer-aided preference and planning tool was applied. The number of participants declined, compared to the first meeting, because of the more specific topic. The tool application was successful in initiating an active, guided discussion on the mussel farm scenarios, where every participant’s view was equally taken into account. With 63%, the environmental farm was favored, compared to 19% for beach mussel farm and 18% for the commercial mussel farm. In the given time frame, only one group managed to finish tool application fully. The composition of the group had influence on the results, but the tendency was similar. Despite favoring scenario 3, the attendees were positive about all scenarios and called for an experimental implementation. An experimental farm of at least 1 ha was jointly suggested to gather the data required for more reliable economic and ecological model simulation as well as the analysis of possible positive and negative side-effects of mussel farms.
To ensure that one of the scenario can be implemented, two questions had to be answered: a) are zebra mussel larvae naturally available at the locations where mussel farming is considered and b) is it possible to grow the mussels on material and in a technical stetup that was originally designed for blue mussels (Petersen et al. 2012, 2014). Experiments that addressed these questions were already carried out prior and during the SAF application process, to avoid a delay in case of a mussel farm implementation.
In many cases, blue mussel larvae are caught at suitable places and are entered into growing devices and cultivated in other places. In our scenarios, we assume that the mussel larvae attach themselves and grow naturally in our farms, to save costs and to reduce the environmental impact of farming. Our Dreissena molecular study suggests presence of larvae at all sampled locations from 24. May onward, (Fig. 2). Earlier, on 7. May, positive molecular signal was detected at 3 of 5 sampled stations. This indicates a good recruitment potential, confirms our visual observation that Dreissena occupies suitable habitats fast and in high numbers everywhere in the lagoon and it allows a natural spat collection on the cultivation facilities.
A study by Schulze-Böttcher (2014) in Lake Usedom, an enclosed bay connected to the Oder Lagoon, showed that zebra mussels easily grow on common artificial nets (Fig. 11) in a density of 21,600–31,700 ind./m2. Within 4 months, the mussels increased their shell length from 1 to 1.2 cm to 1.8–1.9 cm. However, competition with bryozoans, moss animals, reduced the mussel growth. Grov (2015) concluded for Lake Usedom, that bio-deposition of faeces and pseudofaeces had only a minor effect on bio-available nutrients and organic carbon. The experiences of the cultivation experiments in Lake Usedom can be transferred to other parts of the Oder Lagoon and in more exposed open parts of the lagoon, problems with bryozoans are less likely. Against this background, the implementation of a farm is a realistic option. However, it is known from literature that zebra mussels may alter their environment e.g. by increasing the availability of phosphate (Wojtal-Frankiewicz and Frankiewicz 2011) or by affecting sediments (e.g. Christensen et al. 2003; Pollet et al. 2015). In our approach, we used the modelled oxygen depletion as indicator for negative impacts on the environment. We adjusted the mussel cultivation density, so that no increase in the number of days with hypoxia above the sediment occured.
An unexpected, new problem that hampers the implementation of a mussel farm is the closely related quagga mussel (Dreissena bugensis) that was recently observed in the eastern Oder Lagoon (Woźniczka et al. 2016). This new species, indigenous to the Dnieper River drainage of Ukraine, is currently invading the Oder Lagoon. A cultivation of Dreissena polymorpha would automatically favor the development of Dreissena bugensis. This may cause legal and permission problems for mussel farms e.g. according to EU-Regulation No 1143/2014 on the prevention and management of the introduction and spread of invasive alien species (European Union 2014). However, the stakeholders were not concerned about the quagga mussel occurrence, because it resembles the zebra mussel in behavior and appearance, and is only slightly bigger. From a pragmatic point of view, the quagga mussel may even be more suitable for farming.
As consequence of the unclear legal situation in the Oder Lagoon, and to test the general concept, in 2017, an experimental mussel farm was established in neighboring Greifswald Bay. Here, higher salinity (about 7 PSU) only allows the cultivation of blue mussels. Aim is to test a low-cost, nature friendly farming concept for producing mussel meal (as animal feed) and for removing nutrients, similar to the scenarios for the Oder Lagoon. Here, a cultivation system is tested that can be lowered to avoid damage by ice during winter (Fig. 11).