- This line was added.
- This line was removed.
- Formatting was changed.
Ecological landscaping of sand extraction sites, NL
The goal of the pilot experiment is to test whether ecosystem-based landscaping enhances biodiversity after cessation of the sand extraction activities. Present sand extraction policy aims at restoration of the original pre-dredging habitat. This is, however, a limiting approach. Flat seabeds tend to be ecologically less valuable than seabeds with meso-scale bed forms such as tidal ridges, shore face connected ridges and sand waves. Such bed forms provide habitat to a larger range of species assemblages.
In a sand extraction site, seabed morphology and sediment composition is changed by removing a part of the seabed. Depending on the granted sand extraction license, this may lead to a significant increase of the water depth. Returning to the pre-dredging morphology may take decades or more. In recent years, it has become increasingly clear that seabed heterogeneity (i.e. bed forms) constitutes habitats allowing for more biodiversity and more biomass. The pilot project described herein has applied this knowledge to a large marine sand extraction area off the coast of South Holland, the Netherlands. Two large-scale bed forms were left behind after extraction, to test whether these will accelerate ecological recovery and enhance biodiversity and biomass in the dredged area.
The project encompasses the design, organisation and realization of two ecosystem-based landscaped sand bars in a large-scale and deep sand extraction site. Next to the ecosystem-based landscaping, ecosystem-based design rules for future sand extraction sites which optimize the balance between impacted surface area, sand yield, costs and ecological effects are developed.
To safeguard the supply of sand with sand extraction strategies based on ecosystem-based design rules which optimize the balance between impacted surface area, sand yield, costs and ecological effects. Ecosystem-based landscaping will be used to accelerate recovery and to boost habitat heterogeneity and biodiversity.
Restoration of habitats
With the traditional state of development, allocation of permits and related monitoring for sand extraction sites tend to focus on the rate of habitat recovery in the areas. It is often unclear, however, what is meant by the term recovery. The basic goal is re-establishment of a species assemblage like the one existing before dredging. This can only be attained if the original habitat, especially bed topography and sediment composition, are restored (Boyd et al. 2003). This is more likely to occur in case of shallow (2m) extractions, which give rise to limited geomorphological changes. Yet natural fluctuations may lead to the establishment of a community that differs from the original one (Van Dalfsen and Essink 2001).
Development of habitats
If the environmental conditions after dredging have changed significantly, a different stable state may develop with a different species assemblage. Seiderer and Newell (1999) suggest that recovery can be interpreted as the establishment of sufficient species diversity following cessation of dredging, which would allow the biological resources to progress towards diverse equilibrium assemblages. It is therefore advisable to use 'development of habitats' instead of 'restoration of habitats'. This allows the creation of (new) habitats that are tailored to the new circumstances and thus have a much higher chance of reaching a stable ecological state. The creation of bed forms and/or combinations of sediment characteristics yielding geomorphological gradients in a sand extraction site will enhance the conditions necessary for habitat diversity. In turn, habitat diversity may result in high biodiversity as a variety of species will be able to settle within the area.
Planning and design
The Planning and Design works for the Maasvlakte Sand Extraction pilot consisted of the following phase and aspects:
- Initiation phase during which the various options have been developed and tested for their basic feasibility;
- The pre-feasibility phase in which the scientific hypothesis for the proposed solution was elaborated and substantiated;
- Feasibility phase during which in broad consultation the objectives and procedures for the pilot were established; it also included the selection of the test location.
It is obvious that such pilot, set within the national coastal policy and as part of a commercially running construction project only could have been made with due attention to governance aspects.
The increased demand for marine aggregates can lead to a growing impact on the seabed and the organisms that depend on it (macrozoobenthos and demersal fish). Traditionally, these impacts are considered to be negative, though inevitable. Therefore, permitting authorities prescribe that impacts need to be minimized in time and space (traditionally, in the Netherlands permits did not allow for extraction deeper than 2 m below the undisturbed seabed; recently, based on new understanding permits allow for deeper sand extraction, up to 20 m below the seabed).
In the pre-feasibility phase, a literature study supported the idea that an extraction site can be designed and operated in such a way that a good ecological status (species assemblage, high biomass and species richness) can be achieved after extraction. By taking the potential ecological response into account both during the early stages of the design phase and when determining the operational methods, enriched ecological conditions of the sand extraction site could be reached. This is especially the case if the seabed in the sand extraction site is landscaped, e.g. by creating bedforms such as large-scale sand waves or tidal ridges.
Fig. 2. Relation between numbers of species and the location on tidal ridges (tops/crests, swales or troughs or the slopes (van Dijk et at. 2012).
A typical species for the crests of sand waves is the crustacean Urothoe elegans and for troughs the polychaete Spiophanes bombyx . A typical species for the shoreface-connected ridge area is the polychaete Nephtys cirrosa . The swale of the shoreface-connected ridge is typified by the polychaete Eteone longa (Baptist et al. 2006). The study of van Dijk et al. (2012) shows that marine habitats over two tidal ridges in the North Sea vary from low-density/low-diversity communities on the well-sorted sandy crests of the ridges to high-density/high-diversity communities in the poorly sorted muddy, gravelly sediments in the adjacent troughs (see adjacent figure Van Dijk et al. 2012).
Fig. 3. Hypothetical benthic fauna distribution at sand waves.
The troughs of sand waves which are characterised by fine sediment and higher mud content, biodiversity often appeared to harbour more species and a higher biomass (Fig. 3). In turn, the macrozoobenthic communities are an important food source to other species such as flatfish. Clearly, natural physical gradients due to the presence of bedforms have a positive effect on the overall biodiversity and biomass. Landscaping an extraction site to increase the variation in physical parameters such as morphology and hydrodynamic conditions is therefore expected to have a similar effect on the ecosystem in the area.
A desk study was carried out to determine if seabed landscaping was possible within the Dutch legislation. The study identified and analysed the opportunities and obstacles in the existing judicial and licensing framework related to sand extraction in Dutch coastal waters. Policy and legal requirements were assessed by the extent to which, and the way in which, they stimulate or impede the development of seabed landscaping, considering technical dredging aspects as well as ecological opportunities.
For the Maasvlakte 2 sand extraction site a license for sand extraction was already granted. Discussions with permitting authorities and stakeholders led to the conclusion that the landscaping could be integrated in the contractor’s working plan. If the design of the landscape remained within the general requirements and boundaries of the permit, no objections were to be expected (Lulofs, 2010).
The next step was to prepare the design of the landscaping. This was done according to the principles of Eco-Dynamic Development and Design (EDD):
After defining the ecological and physical baseline conditions , the planning and design phase concentrated on optimising the landscaping design for the sand extraction site. It followed the relatively simple approach described below (Rijks, 2011):
The following stakeholders were asked for their input, based on their role in the project:
Choice of Pilot Location
The Maasvlakte 2 sand extraction area lies about 20 km offshore from the Port of Rotterdam (PoR) (Fig. 4, no. 1). The site was chosen in consultation with the project owner, the permit authority and the contractor.
Fig. 4. Location of the pilot site Maasvlakte 2 in front of the Port of Rotterdam (PoR) denoted with 1.
The project team consisted of marine ecologists, seabed morphologists, hydrodynamical and contracting experts from consultancy agencies, contractors, applied research institutes, universities and governmental authorities. Their expertise was used during regular brainstorming sessions and an international group of experts were asked for their verdict concerning the project. Furthermore, the early involvement of stakeholders in the project team was essential for the success and the swiftness of the design-phase. The stakeholders represented all aspects of the project, from technical aspects to contracting and governance. They contributed to the project through regular meetings with the project team throughout the project development.
Trailing suction hopper dredgers
(picture: Daan Rijks, Boskalis)
After the planning phase, the actual dredging and creation of the sand bars started. The dredging activities mainly took place during slack tide. Normal dredging occurs in line with the tidal current and filling of the hopper occurred when sailing from the northern to the southern part and vice versa. Changes in the dredging direction can induce constraints with regards to lateral deviation of the trailing arms and the drag heads.
Execution of the works
The following steps were taken:
The pilot experiment took place in the sand extraction site used for the development of the Port of Rotterdam enlargement ‘Maasvlakte 2’ (MV2). Between 2008 and 2012, 220 million m3 was extracted from a 20-m deep area south of the Euromaasgeul shipping lane. A large northern extraction site is separated by an exclusion area consisting of clay and a southern smaller extraction site.
3D representation of the parallel ecosystem-based sand bar (left) and from the oblique ecosystem-based sand bar from an eastern direction (right) (picture: Maarten de Jong).
One sand bar, parallel to the tidal current, was dredged out the seabed in 2010 in the north-western part of the northern sand extraction site and one sand bar oblique to the tidal current in the south-western part in 2011. Several trailing suction hopper dredgers (TSHDs) were involved, systematically following each other to deepen the projected trough. The sand bars, resulting in between the two troughs, are 700 m long and the crests are 70 m wide at a water depth of 30 m. The water depth of troughs surrounding the parallel sand bar is 40 to 44 m. The troughs of the oblique sand bar are at 32 m water depth. These dimensions are not different to those of natural sand ridges.
Box core sampling for monitoring
Around 2009, knowledge on the relationship between ecology and the seabed was predominantly based on expert judgement and a few scattered data sets. This project offered a unique opportunity to gather field data. During design and preparation of the project a solid monitoring plan was drafted, making use of on data collected in a baseline study. During the recolonization phase after the sand extraction operation and the development of ecosystem-based landscaped sand bars much data from short term monitoring campaigns was gathered. Next to monitoring results about short-term effects, medium and long-term data are currently collected during the recolonization measurements of the Port of Rotterdam (PoR).
Monitoring in the Maasvlakte 2 (MV2) sand extraction sited focused on the general effects and the landscaped sand bars. General effects were studied in a deep area without landscaping and the effects of landscaping were studied at two landscaped sand bars .
The monitoring was meant to provide answers to the following questions:
Monitoring of the MV2 sand extraction site initially copied the PoR monitoring campaign to ensure inter compatibility of data. The 2006 and 2008 baseline data gathered by PoR was used as a reference data on the ecological, morphological and hydrodynamic conditions before the large-scale and deep sand extraction operation for the construction of MV2. It is important that before any intervention is made, consensus is obtained among all parties involved on this reference situation, so as to avoid discussion on it after the system has been disturbed. Getting insight in temporal variability in baseline data is important to separate the ecological effects due to the interventions and autonomous changes.
Specific monitoring of the effect of the landscaped sand bars started within one month after they had been created. It should be noted that immediately after dredging there is hardly any macrozoobenthos left in the sand extraction site. This means that initial monitoring can be carried out quickly and with a minimum of samples. Later, when the dredged and ecological landscaped seabed are being colonized, sampling can be intensified, according to the principles of Adaptive Monitoring Strategies.
To assess the situation prior to the large-scale and deep sand extraction for the construction of MV2, data were collected by NIOZ and IMARES in the framework of the baseline study for the Environmental Impact Assessment (EIA). The EIA was commissioned by the Port of Rotterdam (PoR) and investigates (ecological) effects of the Maasvlakte 2 project. A set of 470 box corer and bottom sledge samples from the baseline monitoring campaign of PoR was selected from data collected in April-June 2006 and 2008 at 235 locations in a 2500 km2 study area off Rotterdam.
235 Sampling locations of the Port of Rotterdam (PoR) baseline study in 2006 and 2008.
The box corer had a surface area of 0.084 m2 and a maximum penetration depth of 30 cm. Samples were wet-sieved using a 1 mm mesh sieve and the residue was stored in jars with a seawater solution of 6% buffered formaldehyde (Perdon & Kaag 2006, Craeymaeersch & Escaravage 2010). The bottom sledge was equipped with a 5-mm mesh cage of 10 cm, penetrating up to 10 cm into the seabed. On average, a surface area of 15 m2 was sampled during each sledge haul of around 150 m length. Specimens were identified up to species level when possible. Infaunal ash-free dry weight biomass (g AFDW m-2) was analysed by means of loss on ignition (2 days at 80°C followed by 2 hours at 580°C). Wet weight of epifauna without sampled fish (g m-2 WW) was directly measured. Biomass of razorblade Ensis spp. was determined by using regression equations based on previous IMARES field surveys (Craeymeersch & van der Land 1998).
Sediment samples from the upper 5 cm were collected from untreated boxcorer samples and kept frozen until analysis. Sediment samples were analysed with a Malvern Mastersizer 2000 particle size analyser. Percentile sediment grain size (D10, D50, D90) and sediment grain size distribution among the different classes; clay (< 4 μm), silt (4-64 μm), mud (<63 μm), very fine sand (63 μm-125 μm), fine sand (125 μm-250 μm), medium sand (250 μm-500 μm) and coarse sand (> 500 μm) were measured as percentage of total volume (vol%). Sediment sorting (D90/D10) was determined, sediment organic matter (SOM) was analysed by means of loss on ignition (freeze-dried sediment samples were placed for 2 hrs at 580°C) and values of SOM were calculated as percentage of sediment mass (mass%).
A monitoring campaign was designed and aligned with the survey campaign of PoR to maximize inter compatibility and amount of field data output and to reduce costs. During the campaign focus was placed on the short-term effects in the sand extraction site on macrozoobenthos (infauna and epifauna), demersal fish, sediment characteristics, sedimentation-erosion and hydrodynamics.
Used monitoring techniques: top from left to right: a box corer for infauna sampling, sediment sampling from a box corer sample, sieving of sediment to extract infauna: bottom from left to right: bottom sledge to sample epifauna, 4.5 m wide beam trawl to sample demersal fish and multibeam measurements for bathymetry and changes in sedimentation-erosion.
Macrozoobenthos short-term monitoring campaign
To ensure comparable data, sampling was carried out using identical protocols as during the baseline study (De Jong et al., 2015a; Perdon and Kaag, 2006) and during the recolonization study of the Environmental Impact Assessment (EIA) of the Port of Rotterdam for the construction of Maasvlakte 2 (runs from 2014-2018) . A box corer with a surface area of 0.077 m2 was used to sample sediment and infauna, organisms larger than 1 mm and mainly living in the seabed. A bottom sledge was used to sample macrobenthic in- and epifauna with a size range of 0.5 – 10 cm. Bottom sledge samples are hereafter called epifauna (EP) beacause the largest part is epifauna although infauna was collected as well (De Jong et al., 2015b).
Sampling with the box corer was executed by the Royal Netherlands Institute for Sea Research (NIOZ) on 29–30th June 2010, 2–5th May 2011 and 23–25th April 2012. In 2010 and 2011, 45 and in 2012, 64 box corer samples were collected. To reach a higher spatial resolution in- and outside the extraction site in 2012, a subsample of the box core sample was analysed which reduced the sampled surface area to 0.015 m2. Four samples were collected in the deep parts of the extraction site, 14 samples in the reference area (near and far field) and four samples in the shipping lane area. No maintenance dredging was executed in the shipping lane according to Rijkswaterstaat. Specimens were identified up to species level when possible and ash-free dry weight biomass (g AFDW m-2) was analysed by means of loss on ignition, 2 days at 80 °C followed by 2 h at 520 °C.
Sampling with the bottom sledge for epifauna was conducted by the Institute for Marine Resources & Ecosystem Studies (IMARES Wageningen UR) on 7–8th July 2010, 14–15th June 2011 and 6–7th June 2012. The bottom sledge was equipped with a 5-mm mesh cage. On average, a surface area of 15 m2 was sampled during each sledge haul of approximately 150 m length, 10 cm width and a maximum penetration depth of 10 cm. In 2010 and 2011, 26 and in 2012, 32 bottom sledge samples were collected. In 2010, 11 bottom sledge samples were collected in the reference area. In 2012, three samples were collected in the shipping lane area of the Port of Rotterdam and one sample in the reference area. Specimens were identified up to species level when possible. Wet weight of epifauna was directly measured after sorting (g m-2 WW). Biomass of Atlantic jackknife clam Ensis sp. was determined by using regression equations based on previous IMARES field surveys. Sampling locations are visualised in figure 11 and 12.
Samples from the upper 5 cm of sediment were collected from untreated boxcorer samples and kept frozen until analysis. Sediment samples were freeze dried, homogenised and analysed with a Malvern Mastersizer 2000 particle size analyser. Percentile sediment grain size (D10, D50, D90) and sediment grain size distribution among the different classes; clay (<4 mm), silt (4-63 mm), mud (<63 mm), very fine sand (63 mm-125 mm), fine sand (125 mm-250 mm), medium sand (250 mm-500 mm) and coarse sand (>500 mm) were measured as percentage of total volume. Sediment organic matter (SOM) was analysed in 2012 by means of loss on ignition as percentage of sediment mass (freeze-dried sediment samples were placed for 2 hours at 520 °C) (De Jong et al., 2014).
Modelling of abiotic variables
The sand extraction site and surrounding area have complex hydrodynamic conditions. Due to the region of fresh water input (ROFI) from the Rhine River, periods of strong haline stratification, up- and downwelling, wind-driven flow, baroclinic cross-shore flows and wind and wave-induced mixing occur frequently. This may lead to considerable fluctuations in salinity and bottom shear stress. Delft 3D DD ZUNO has been used, a hydrodynamical model of the southern North Sea consisting of a coarse curvilinear horizontal grid with two grid refinement towards the Dutch coast through domain decomposition (DD). The nested model grid covers an area of about 7.5 km by 7.0 km and the horizontal grid size is about 45 m by 38 m for the research area.
To generate appropriate boundary conditions for the nested model (Maasvlakte 2 sand extraction site), four open boundaries were used with tangential velocity (Tonnon et al., 2013). Twelve vertical σ-layers were specified, the relative thickness of these has been chosen in such way that near-bed and near-surface vertical gradients were better resolved. From top to bottom, these layers represent respectively 4.0%, 5.6%, 7.8%, 10.8%, 10.9%, 10.9%, 10.9%, 10.9%, 10.8%, 7.8%, 5.6%, and 4.0% of the water depth. The bathymetry of the nested model was interpolated from multibeam measurements performed by the dredging companies in October 2010. The model was forced with measured meteorological and riverine discharge data for the specific period. In order to keep the calculation time manageable, one single spring-neap cycle with relatively high river discharges was used for validation. During December to April, the largest riverine discharges can be expected and stratification is more likely to occur. The period between 2 March and 17 March 2007 was simulated using a time step of 15 seconds. Mean and maximum values of bed shear stress (N m-2) and near-bed salinity (ppt) were modelled.
A commercial fishing vessel was used, the Jan Maria, GO 29, with a length of 23 m, less than 300 horsepower and equipped with a standard commercial 4.5 m beam trawl. The beam trawl was equipped with four tickler chains, five flip-up ropes and diamond mesh size of 80 mm, which was applied at 4 knots fishing speed. The ship’s GPS-system logged the position of the sampling locations and water depth was determined with the ship's depth sounder. The maximum haul distance was one nautical mile in the reference area. Shorter hauls were planned within the sand extraction site; at the landscaped sand bars, hauls of approximately 700 m length were applied. Some of the hauls ended before the planned end coordinates because of difficulties with fishing inside the sand extraction site due to large changes in seabed topography and sediment composition.
In surrounding reference areas, fishing direction was generally perpendicular to the direction of naturally occurring seabed patters to ensure heterogeneous sampling of crests and troughs of sand waves. In the sand extraction site, fishing direction was generally parallel to the seabed structures to enable comparisons between the different locations. Sampling was made in the reference area, at the slope of the sand extraction site, two locations in the deep parts of the extraction site i.e. the south-east and north-west, in the troughs and at the crests of the sand bars.
Fish tracks and macrozoobenthos samples (picture: Maarten de Jong).
Sediment and modelled hydrodynamical variables
An increase in mud content was observed after the cessation of sand extraction. Locations with a high mud value also have a high sediment organic matter content (SOM). Modelled time-averaged bed shear stress is high at the crests of natural occurring sand waves in the reference areas. The northern edge of the northern sand extraction site shows the highest bed shear stress values. The southern part of the northern sand extraction site has the lowest bed shear stress values. Differences in bed shear stress are also visible at the ecosystem-based sand bars.
A selection of abiotic variables at infauna sample locations: median grain size (micron) and mud and sediment organic matter (2012) (%) and salinity (g/kg) and bed shear stress (N/m2) for the year 2010 (De Jong et al., 2015b) .
Modelled bed shear stress (N/m2) in the Maasvlakte2 sand extraction site (Tonnon et al., 2013) .
Two years after the cessation of sand extraction, macrozoobenthic biomass increased five-fold in the deepest areas. Species composition changed significantly and white furrow shell (Abra alba ) became abundant. Several sediment characteristics also changed significantly in the deepest parts. Macrozoobenthic species composition and biomass significantly correlated with time after cessation of sand extraction, sediment and hydrographical characteristics. Ecosystem-based landscaped sand bars were found to be effective in influencing sediment characteristics and macrozoobenthic assemblage. Significant changes in epifauna occurred in deepest parts in 2012 which coincided with the highest sedimentation rate.
Characteristics of infaunal recolonization (De Jong et al., 2015b) .
Characteristics of epifaunal recolonization (De Jong et al., 2015b) .
Characteristics of macrozoobenthos per location (De Jong et al., 2015b) .
Demersal fish species
One and two years after cessation, a significant 20-fold increase in demersal fish biomass was observed in deep parts of the extraction site. The average demersal fish wet weight biomass in the reference areas was 20.9 kg WW ha-1 whereas in the deep areas of the MV2 sand extraction site biomass increased up to 522 kg WW ha-. The most abundant fish species in the extraction site is plaice (Pleuronectes platessa) whereas in the reference areas, dab (Limanda limanda) is most abundant.
Demersal fish biomass in- and outside the sand extraction site in 2010, 2011 and 2012. Values are proportional to the radius of the circles in the bubble plot with maximum values converted to bubbles with 1000 m radius. The highest biomass value was found in 2012 at the trough of the oblique sand bar (522 kg WW ha-1).
In the troughs of a landscaped sand bar however, a significant drop in biomass down to reference levels and a significant change in species assemblage was observed two years after cessation. The fish assemblage at the crests of the sand bars differed significantly from the troughs with tub gurnard (Chelidonichthys lucerna) being a typical species of the crests. This is a first indication of the applicability of landscaping techniques to induce heterogeneity of the seabed although it remains difficult to draw a strong conclusion due the lack of replication in the experiment. An ecological equilibrium is not reached after 2 years since biotic and abiotic variables are still adapting.
Significant differences in demersal fish species assemblages in the sand extraction site were associated with variables such as water depth, median grain size, fraction of very fine sand, biomass of white furrow shell (Abra alba) and time after the cessation of sand extraction. Large quantities of undigested crushed white furrow shell fragments were found in all stomachs and intestines of plaice, indicating that it is an important prey item.
nMDS ordination with demersal fish samples and significant associations with variables. Continuous variables are depicted with arrows and categorical variable time after the cessation, only with text (Tref: reference in black, Trecent in white, T1 in light grey and T2: dark grey. Sub-locations are denoted with symbols, reference as black bullets, slope of the reference area as large open circles, crests of sand bars as squares, trough of sand bars as diamonds, deep (SE) as point-up triangles and deep (NW) as point-down triangles. In 2012, the significant association of the ordination and infaunal white furrow shell biomass is denoted with an arrow and surface plot to show the non-linear property of the relationship. Stress of all ordinations was below 0.07.
During the MV2 sand extraction project, many new insights were gained concerning technical and ecological aspects of the design and the organisation of large scale, deep sand extraction projects and ecosystem-based landscaping. The project also generated a broad discussion amongst the various stakeholders on how changing physical conditions can trigger the development of new ecological habitats.
An important lesson learned is that ecosystem-based landscaping in sand extraction sites only make sense if:
- the sand extraction volume and area are large enough for ecosystem-based landscaping to have maximum added value;
- Ecosystem-based landscaping can be used to increase ecological value (de Jong et al., 2015; De Jong et al., 2016, 2014) ; and
- Ecosystem-based landscaping can be carried out during the extraction process without additional equipment mobilisation and with minimal interference to the overall sand extraction production process.
Overall, it became clear that it is still too early to prescribe landscaping to other sand extraction projects, even if they meet the above conditions. The present pilot experiment is still on-going and, although the concept appears to be positive, its added ecological value remains scientifically to be proven.
Lessons on project development
The most important lessons learned are:
Lessons on physical and technical aspects
Several lessons were learned concerning physical parameters and the technical realization:
Lessons on the ecological response
Lessons on governance
The project team worked closely together with the stakeholders, thus learning several lessons concerning governance issues:
Lessons on realization
Several important lessons were learned during the execution of the dredging works, concerning the workability of the design and the consequences of creating a specific landscape as compared with a case without predefined bedforms:
Lessons on monitoring
EcoShape pilot project
Related Building Solutions