Academia.eduAcademia.edu

Figure 5 – uploaded by Jan Stafleu

See full PDFdownloadDownload figure
Table Ill. Lithostratigraphic overview of the main resource-relevant geological layers in the Quaternary. In the possible to model consistently three resource-relevant layers: Pleistocene, lower and upper Holocene (Hademenos et al., in press). Important Pleistocene deposits in the BPNS are concentrated in paleovalleys. The near-coastal Ostend Valley has the most voluminous infill (Figure 10). In the sNPNS, the thickness of the Pleistocene deposits increases from southwest to northeast, and toward the axis of the lowstand Rhine-Meuse river linking Rotterdam Harbour to the Hinder Banks area (location, see Fig. 1). As these Pleistocene units were formed during times of high and low sea level, they include muddy as well as gravelly sediments with strongly varying thicknesses and large spatial heterogeneity. Locally, extraction is possible, but it needs to be guided by geological knowledge from seismics and boreholes. The lower-Holocene units are characterized by low-energy deposits accumulated in tidal settings. Typically, these are highly variable in composition, with abundant fine- grained material. They are unsuitable for extraction. Only the upper-Holocene units, deposited under the combined action of waves and tidal currents, are recommended for extraction, especially in the BPNS; they are relatively homogeneous and contain only small percentages of silt to clay fractions. Table IV provides an overview of the main lithoclasses per unit, together with typical admixtures. Figure 10 shows a typical succession of the three layers on the BPNS. The distribution and thickness of the three layers, as well as their lithoclass and admixture distributions, can be queried in the TILES DSS.

Figure 10 Ill. Lithostratigraphic overview of the main resource-relevant geological layers in the Quaternary. In the possible to model consistently three resource-relevant layers: Pleistocene, lower and upper Holocene (Hademenos et al., in press). Important Pleistocene deposits in the BPNS are concentrated in paleovalleys. The near-coastal Ostend Valley has the most voluminous infill (Figure 10). In the sNPNS, the thickness of the Pleistocene deposits increases from southwest to northeast, and toward the axis of the lowstand Rhine-Meuse river linking Rotterdam Harbour to the Hinder Banks area (location, see Fig. 1). As these Pleistocene units were formed during times of high and low sea level, they include muddy as well as gravelly sediments with strongly varying thicknesses and large spatial heterogeneity. Locally, extraction is possible, but it needs to be guided by geological knowledge from seismics and boreholes. The lower-Holocene units are characterized by low-energy deposits accumulated in tidal settings. Typically, these are highly variable in composition, with abundant fine- grained material. They are unsuitable for extraction. Only the upper-Holocene units, deposited under the combined action of waves and tidal currents, are recommended for extraction, especially in the BPNS; they are relatively homogeneous and contain only small percentages of silt to clay fractions. Table IV provides an overview of the main lithoclasses per unit, together with typical admixtures. Figure 10 shows a typical succession of the three layers on the BPNS. The distribution and thickness of the three layers, as well as their lithoclass and admixture distributions, can be queried in the TILES DSS.

Related Figures (41)

Transnational and Integrated Long-term Marine Exploitation Strategies VAN LANCKER VERA, FRANCKEN FREDERIC, KAPEL MICHEL, KINT LARS, TERSELEER NATHAN,
Transnational and Integrated Long-term Marine Exploitation Strategies VAN LANCKER VERA, FRANCKEN FREDERIC, KAPEL MICHEL, KINT LARS, TERSELEER NATHAN,
Transnational and Integrated Long-term Marine Exploitation Strategies Contract - BR/121/A2
Transnational and Integrated Long-term Marine Exploitation Strategies Contract - BR/121/A2
Figure 1. The Belgian and southern Netherlands part of the North Sea, a typical sandbank environment in the southern part of the North Sea, northwestern Europe. The different sandbank systems are: (1) Coastal Banks; (2) Flemish Banks; (3)(6) Zeeland Banks; (4) Hinder Banks; (5) Ebb-tidal-delta shoals; (7) Transition into sandwave fields.
Figure 1. The Belgian and southern Netherlands part of the North Sea, a typical sandbank environment in the southern part of the North Sea, northwestern Europe. The different sandbank systems are: (1) Coastal Banks; (2) Flemish Banks; (3)(6) Zeeland Banks; (4) Hinder Banks; (5) Ebb-tidal-delta shoals; (7) Transition into sandwave fields.
Figure 3. The TILES approach in defining the resource potential or reserve. Three information pillars provide input for a DSS on sustainable sand and gravel extraction: resource availability, sustainable exploitation, and resource constraints. Integration of cross-domain information yields sustainable extraction thresholds to be used in scenario analyses that will help to ensure that extracted volumes stay within sustainability limits, a prerequisite within Europe’s MSFD. contemporary socio-economic context and with minimal environmental impact. Figure 2. User criteria to define the concept of resource to reserve. The geological resource is the maximum availability. The reserve is constrained to what is most suitable and certain for extraction within the contemporary socio-economic context and with minimal environmental impact.
Figure 3. The TILES approach in defining the resource potential or reserve. Three information pillars provide input for a DSS on sustainable sand and gravel extraction: resource availability, sustainable exploitation, and resource constraints. Integration of cross-domain information yields sustainable extraction thresholds to be used in scenario analyses that will help to ensure that extracted volumes stay within sustainability limits, a prerequisite within Europe’s MSFD. contemporary socio-economic context and with minimal environmental impact. Figure 2. User criteria to define the concept of resource to reserve. The geological resource is the maximum availability. The reserve is constrained to what is most suitable and certain for extraction within the contemporary socio-economic context and with minimal environmental impact.
Figure 4. Overview of lithological description data, associated metadata, and the codes that were derived to facilitate the development of geological models. Borehole data were standardized following European guidelines (e.g. EU-FP7 Geo-Seas for geologica and geophysical data; EU-FP7 SeaDataNet for oceanographic data). Main lithology (e.g. gravel, sanc clay), Wentworth classes (e.g. fine, medium, coarse sand) and (adjusted) Folk (e.g. sandy muc gravelly sand, muddy gravel) were mapped consistently for the BPNS and sNPNS. Secondar components (admixtures) were also added to the database (e.g. shell content, percentage fines (clay silt, or < 63 um) and/or gravel; organic and glauconite content) since they may render an aggregat less suitable for industrial applications. All terms used in the conversion are linked in the metadata t INSPIRE-compliant vocabulary organized in different hierarchical levels. The standardized formattin; allows further integration into European geological data portals (e.g. EMODnet-Geology, Europea Geological Data Infrastructure EGDI). Importantly, all data were subjected to extensive qualit control, and lithological descriptions were coded for easy querying and data extraction.
Figure 4. Overview of lithological description data, associated metadata, and the codes that were derived to facilitate the development of geological models. Borehole data were standardized following European guidelines (e.g. EU-FP7 Geo-Seas for geologica and geophysical data; EU-FP7 SeaDataNet for oceanographic data). Main lithology (e.g. gravel, sanc clay), Wentworth classes (e.g. fine, medium, coarse sand) and (adjusted) Folk (e.g. sandy muc gravelly sand, muddy gravel) were mapped consistently for the BPNS and sNPNS. Secondar components (admixtures) were also added to the database (e.g. shell content, percentage fines (clay silt, or < 63 um) and/or gravel; organic and glauconite content) since they may render an aggregat less suitable for industrial applications. All terms used in the conversion are linked in the metadata t INSPIRE-compliant vocabulary organized in different hierarchical levels. The standardized formattin; allows further integration into European geological data portals (e.g. EMODnet-Geology, Europea Geological Data Infrastructure EGDI). Importantly, all data were subjected to extensive qualit control, and lithological descriptions were coded for easy querying and data extraction.
using the dominant sediment type of the interval described. *DGPS: differential global positioning system.
using the dominant sediment type of the interval described. *DGPS: differential global positioning system.
Figure 5. Seismic (left) and borehole (right) data available for the BPNS and sNPNS. Data quality and data-related uncertainty Data uncertainty relates to vintage (time of sampling, relevant because the seabed changes through time), positioning, sampling technique and analytical procedures (van Heteren & Van Lancker, 2015). To quantify these uncertainties metadata were scored following newly developed protocols (Table 1). Thus, data quality and other uncertainty factors can be incorporated into resource calculations. Scores from 0 to 100 were applied to generate maps that are compatible with the probability maps on geological characteristics; more simple scores of 0 to 5 were assigned to each metadata field to be provided in geological log sheets from the newly developed geological data portal (see Dissemination and Valorization).
Figure 5. Seismic (left) and borehole (right) data available for the BPNS and sNPNS. Data quality and data-related uncertainty Data uncertainty relates to vintage (time of sampling, relevant because the seabed changes through time), positioning, sampling technique and analytical procedures (van Heteren & Van Lancker, 2015). To quantify these uncertainties metadata were scored following newly developed protocols (Table 1). Thus, data quality and other uncertainty factors can be incorporated into resource calculations. Scores from 0 to 100 were applied to generate maps that are compatible with the probability maps on geological characteristics; more simple scores of 0 to 5 were assigned to each metadata field to be provided in geological log sheets from the newly developed geological data portal (see Dissemination and Valorization).
Uncertainty was also quantified for the seismic data. The analogue and digital seismic data, collectec during the past 45 years, were catalogued and assigned with a quality indicator following Europear guidelines (EU FP7 Geo-Seas) with each (navigation) line being assigned a value ranging from 0 (nc data) to 5 (good data). This provides insight into the uncertainty associated with seismics-based laye! models covering specific areas and enables the selection of the highest quality lines for future work.
Uncertainty was also quantified for the seismic data. The analogue and digital seismic data, collectec during the past 45 years, were catalogued and assigned with a quality indicator following Europear guidelines (EU FP7 Geo-Seas) with each (navigation) line being assigned a value ranging from 0 (nc data) to 5 (good data). This provides insight into the uncertainty associated with seismics-based laye! models covering specific areas and enables the selection of the highest quality lines for future work.
Table Il. Attributes of the voxel model, presently at a resolution of 200*200*1 m. Modelling admixtures constraining resource suitability
Table Il. Attributes of the voxel model, presently at a resolution of 200*200*1 m. Modelling admixtures constraining resource suitability
Figure 6. Coupling between the 3D voxel model (left) and the suite of environmental impact models (right). The voxel model provides percentages (%) of the different lithological classes (Litho), which are translated into concentrations (C) of the different sediment fractions to initialize the seabed of the numerical model. Properties of the sediment fractions (e.g. particle size) are provided as well to parameterize the numerical model. Owing to computational constraints, the different lithological classes of the voxel model must be grouped into broader sediment fractions in the numerical model (i.e. f < c). Also, highly similar superimposed voxels should be merged before feeding into the numerical model (i.e. k < i). The sediment-transport and seabed-morphology modules include the possibility to simulate different sediment fractions and seabed layers. Relevant information contained in the voxel model can be transferred to the numerical model suite (Figure 6): various voxel attributes are used to define initial conditions (as a measure of expected sediment concentrations and erodibility) and to distil the most suitable parameters (e.g. particle size, settling velocity, cohesion, or any other relevant parameter) that characterize the different sediment fractions and stratigraphic units in the seabed sediment and shallow subsurface.
Figure 6. Coupling between the 3D voxel model (left) and the suite of environmental impact models (right). The voxel model provides percentages (%) of the different lithological classes (Litho), which are translated into concentrations (C) of the different sediment fractions to initialize the seabed of the numerical model. Properties of the sediment fractions (e.g. particle size) are provided as well to parameterize the numerical model. Owing to computational constraints, the different lithological classes of the voxel model must be grouped into broader sediment fractions in the numerical model (i.e. f < c). Also, highly similar superimposed voxels should be merged before feeding into the numerical model (i.e. k < i). The sediment-transport and seabed-morphology modules include the possibility to simulate different sediment fractions and seabed layers. Relevant information contained in the voxel model can be transferred to the numerical model suite (Figure 6): various voxel attributes are used to define initial conditions (as a measure of expected sediment concentrations and erodibility) and to distil the most suitable parameters (e.g. particle size, settling velocity, cohesion, or any other relevant parameter) that characterize the different sediment fractions and stratigraphic units in the seabed sediment and shallow subsurface.
Figure 7. Projections of how queries on the 3D voxel model are represented on 2D maps.
Figure 7. Projections of how queries on the 3D voxel model are represented on 2D maps.
Resource suitability of geological layers a The BCS is a sediment-depleted shallow shelf environment comprising a series of sandbanks. Its location on the edge of a structural high, the London-Brabant Massif, explains the dominance of recycling and redistribution processes that has created a complex and discontinuous Quaternary sediment cover. This thin cover is considered the most viable resource, since the underlying Paleogene strata include an abundance of compacted clays. Towards the border with the sNPNS, alternating sequences of silt and clay are the most common, with sizable units of silty sand, muddy sand and even calcareous sandstone beds (Le Bot et al., 2003). In the sNPNS, located in a structural low (the North Sea Basin), the Quaternary cover thickens rapidly from southwest to northeast. For the BPNS a new Top-Paleogene surface was published (De Clercq et al., 2016). Its geomorphology is characterized by planation surfaces, bounded by scarps and slope breaks, paleovalleys, and elongated depressions (Liu et al., 1992; Mathys, 2009; De Clercq et al., 2016; De Clercq, 2018). These have been infilled or smoothed during the Quaternary. In Figure 10, the Paleogene is undifferentiated. A subcrop map of the Paleogene, with the successions as described above, can be downloaded via http://www.emodnet-geology.eu. The BCS is a sediment-depleted shallow shelf environment comprising a series of sandbanks. Its location on the edge of a structural high, the London-Brabant Massif, explains the dominance of
Resource suitability of geological layers a The BCS is a sediment-depleted shallow shelf environment comprising a series of sandbanks. Its location on the edge of a structural high, the London-Brabant Massif, explains the dominance of recycling and redistribution processes that has created a complex and discontinuous Quaternary sediment cover. This thin cover is considered the most viable resource, since the underlying Paleogene strata include an abundance of compacted clays. Towards the border with the sNPNS, alternating sequences of silt and clay are the most common, with sizable units of silty sand, muddy sand and even calcareous sandstone beds (Le Bot et al., 2003). In the sNPNS, located in a structural low (the North Sea Basin), the Quaternary cover thickens rapidly from southwest to northeast. For the BPNS a new Top-Paleogene surface was published (De Clercq et al., 2016). Its geomorphology is characterized by planation surfaces, bounded by scarps and slope breaks, paleovalleys, and elongated depressions (Liu et al., 1992; Mathys, 2009; De Clercq et al., 2016; De Clercq, 2018). These have been infilled or smoothed during the Quaternary. In Figure 10, the Paleogene is undifferentiated. A subcrop map of the Paleogene, with the successions as described above, can be downloaded via http://www.emodnet-geology.eu. The BCS is a sediment-depleted shallow shelf environment comprising a series of sandbanks. Its location on the edge of a structural high, the London-Brabant Massif, explains the dominance of
Figure 9. Signature of the Dutch seismo-stratigraphic units (see Table III for correlation with lithostratigraphic units) together with typical borehole sections. Only the upper-Holocene units deposited under the combined action of tides and waves are recommended for sand extraction. They are relatively homogeneous, as depicted in E and G. Types D and F are variable in composition and may contain sizable admixtures of fines; sands are to be extracted with caution.
Figure 9. Signature of the Dutch seismo-stratigraphic units (see Table III for correlation with lithostratigraphic units) together with typical borehole sections. Only the upper-Holocene units deposited under the combined action of tides and waves are recommended for sand extraction. They are relatively homogeneous, as depicted in E and G. Types D and F are variable in composition and may contain sizable admixtures of fines; sands are to be extracted with caution.
= = Ss vy Figure 10. Main peclogical units on the BPNS: Paleogene (here undifferentiated), Pleistocene, lower and upper Holocene. The depth of the Top-Palaeogene was re-mapped (De Clercq et al., 2016). The Middle and Offshore Scarps, two important geomorphological features, were used to split the model into regions with similar ithological characteristics. Sandbanks have a unique internal structure and may be composed of different ayers with specific grain-size distributions and amounts of shell fragments, gravel, clay or other material Hademenos et al., in press). Table IV. Lithoclass variation and possible admixtures for the main Quaternary units: Pleistocene, including reworked transgressive lag (red), lower Holocene (orange), upper Holocene (vellow). Table IV. Lithoclass variation and possible admixtures for the main Quaternary units: Pleistocene, including
= = Ss vy Figure 10. Main peclogical units on the BPNS: Paleogene (here undifferentiated), Pleistocene, lower and upper Holocene. The depth of the Top-Palaeogene was re-mapped (De Clercq et al., 2016). The Middle and Offshore Scarps, two important geomorphological features, were used to split the model into regions with similar ithological characteristics. Sandbanks have a unique internal structure and may be composed of different ayers with specific grain-size distributions and amounts of shell fragments, gravel, clay or other material Hademenos et al., in press). Table IV. Lithoclass variation and possible admixtures for the main Quaternary units: Pleistocene, including reworked transgressive lag (red), lower Holocene (orange), upper Holocene (vellow). Table IV. Lithoclass variation and possible admixtures for the main Quaternary units: Pleistocene, including
Figure 11. Geological units and their main lithology (Holocene units not mapped for the Dutch onshore area; hence, the top Pleistocene is visible). The base of exploitable resources is the top of the clay-rich Paleogene (green). Despite extensive reworking of sediments during the Quaternary, Pleistocene sands outcrop, e.g., in some of the deeper swales between sandbanks and as infill of some of the paleochannels close to shore. Lower-Holocene sands are very patchily distributed, with the higher chance of them outcropping in swales or other bathymetric lows. Upper-Holocene sands are abundantly present. Figure 11 to 17 show the results for the entire TILES area, from the Belgian-French border to Rotterdam, the Netherlands. See captions for a summary of the main findings. For the interpretation of the legend of the probability maps as displayed in Figure 11 to 17, we refer to the IPCC Climate Report: occurrences with values < 0.1: very unlikely; 0.1 - 0.33: unlikely; 0.33 - 0.66: about as likely as not; 0.66 - 0.9: likely; and > 0.9: very likely. Rotterdam, the Netherlands. See captions for a summary of the main findings. For the interpretation
Figure 11. Geological units and their main lithology (Holocene units not mapped for the Dutch onshore area; hence, the top Pleistocene is visible). The base of exploitable resources is the top of the clay-rich Paleogene (green). Despite extensive reworking of sediments during the Quaternary, Pleistocene sands outcrop, e.g., in some of the deeper swales between sandbanks and as infill of some of the paleochannels close to shore. Lower-Holocene sands are very patchily distributed, with the higher chance of them outcropping in swales or other bathymetric lows. Upper-Holocene sands are abundantly present. Figure 11 to 17 show the results for the entire TILES area, from the Belgian-French border to Rotterdam, the Netherlands. See captions for a summary of the main findings. For the interpretation of the legend of the probability maps as displayed in Figure 11 to 17, we refer to the IPCC Climate Report: occurrences with values < 0.1: very unlikely; 0.1 - 0.33: unlikely; 0.33 - 0.66: about as likely as not; 0.66 - 0.9: likely; and > 0.9: very likely. Rotterdam, the Netherlands. See captions for a summary of the main findings. For the interpretation
Figure 12. Fence diagram of the geological units and their main lithology. It illustrates the relative scarcity of sands on the BPNS at the shoulder of the London-Brabant Massif, and accentuates the strong increase of Quaternary sand thickness in the sedimentary basin of the Netherlands. Figure 12. Fence diagram of the geological units and their main lithology. It illustrates the relative scarcity of
Figure 12. Fence diagram of the geological units and their main lithology. It illustrates the relative scarcity of sands on the BPNS at the shoulder of the London-Brabant Massif, and accentuates the strong increase of Quaternary sand thickness in the sedimentary basin of the Netherlands. Figure 12. Fence diagram of the geological units and their main lithology. It illustrates the relative scarcity of
Figure 13. Aggregate quality in terms of most likely lithological class. It confirms the predominance of fine sands in the coastal zone, shifting towards medium sands in the offshore region. Clays and silt are abundantly present in the eastern Belgian coastal zone.
Figure 13. Aggregate quality in terms of most likely lithological class. It confirms the predominance of fine sands in the coastal zone, shifting towards medium sands in the offshore region. Clays and silt are abundantly present in the eastern Belgian coastal zone.
Figure 14. Entropy or model uncertainty of the prediction of the lithoclasses in the Holocene. Entropy not calculated for Pleistocene deposits, here depicted in light grey.
Figure 14. Entropy or model uncertainty of the prediction of the lithoclasses in the Holocene. Entropy not calculated for Pleistocene deposits, here depicted in light grey.
Figure 15. Aggregate quality in terms of the probability that a voxel contains fine sand. Such output provides more in-depth information on where a particular lithological class can be found with higher certainty (yellow to red areas).
Figure 15. Aggregate quality in terms of the probability that a voxel contains fine sand. Such output provides more in-depth information on where a particular lithological class can be found with higher certainty (yellow to red areas).
Figure 16. Aggregate quality in terms of the probability that a voxel contains medium sand. Areas farthest offshore have the highest certainty that this important resource is present (yellow to red areas).
Figure 16. Aggregate quality in terms of the probability that a voxel contains medium sand. Areas farthest offshore have the highest certainty that this important resource is present (yellow to red areas).
Figure 17. Aggregate quality combining lithostratigraphy and lithoclasses. Here, an example of the probability that a Pleistocene voxel contains medium sand. It illustrates the influence of the Rhine-Meuse fluvial system gradually diminishing from east to west. On the BPNS these deposits have been largely eroded and/or reworked.
Figure 17. Aggregate quality combining lithostratigraphy and lithoclasses. Here, an example of the probability that a Pleistocene voxel contains medium sand. It illustrates the influence of the Rhine-Meuse fluvial system gradually diminishing from east to west. On the BPNS these deposits have been largely eroded and/or reworked.
Figure 18. Density of the lithological descriptions used to construct the Belgian part of the voxel model. Data density is lowest in the far western and northern offshore areas.
Figure 18. Density of the lithological descriptions used to construct the Belgian part of the voxel model. Data density is lowest in the far western and northern offshore areas.
Figure 19. Data density, categorized in core lengths, superimposed on geological heterogeneity (shown in purple). Using the TILES DSS, it is here queried for the top 5 meters of the model. Heterogeneity, or varying lithology, is highest in-between sandbanks. It is a proxy of how different in nature the voxels are in a chosen voxel stack, hence over a user-defined depth interval.
Figure 19. Data density, categorized in core lengths, superimposed on geological heterogeneity (shown in purple). Using the TILES DSS, it is here queried for the top 5 meters of the model. Heterogeneity, or varying lithology, is highest in-between sandbanks. It is a proxy of how different in nature the voxels are in a chosen voxel stack, hence over a user-defined depth interval.
Figure 20. Thicknesses of the main geological layers. See Figure 21 for their lithological variation
Figure 20. Thicknesses of the main geological layers. See Figure 21 for their lithological variation
Figure 21. Volumes of lithological classes in the different units of the Quaternary on the BPNS. Note that no unit has a unique lithology. Upper-Holocene sediments have the highest chance of being medium to coarse sands. Overall in the Quaternary, fine sands are predominant (49 %), followed by medium sands (36 %), clay-silt (9 %), and coarse sand and gravel (7 %).
Figure 21. Volumes of lithological classes in the different units of the Quaternary on the BPNS. Note that no unit has a unique lithology. Upper-Holocene sediments have the highest chance of being medium to coarse sands. Overall in the Quaternary, fine sands are predominant (49 %), followed by medium sands (36 %), clay-silt (9 %), and coarse sand and gravel (7 %).
Table V. TILES DSS search criteria that can be applied to estimate volumes for most optimal aggregate extraction. Here for aggregate sector 20D, Oostdijck sandbank (location, Fig. 1; see also Fig. 22).
Table V. TILES DSS search criteria that can be applied to estimate volumes for most optimal aggregate extraction. Here for aggregate sector 20D, Oostdijck sandbank (location, Fig. 1; see also Fig. 22).
‘igure 22. An example of a query in the TILES DSS using different qualitative resource criteria, here for the Jostdijck aggregate sector. Notice the resource volume difference as extra criteria are added to the query. -ach filter has a different effect on each lithoclass. The entropy filter reduces more the medium and coarse ands than the data quality criteria on the borehole density and positional quality. This is because the amount »f samples containing these lithoclasses are smaller than for the fine sand lithoclass, hence the entropies are ligher. With this, it is important to highlight that statistics have a big effect on the model. Figure 22. An example of a query in the TILES DSS using different qualitative resource criteria, here for the *Experience-based, from calculating the mean of the entropy distribution of various areas in the BPNS ( means no uncertainty, whereas 1 occurs when all lithological classes have the same probability). **British Standard for ‘Aggregates for concrete’: BS EN 12620.
‘igure 22. An example of a query in the TILES DSS using different qualitative resource criteria, here for the Jostdijck aggregate sector. Notice the resource volume difference as extra criteria are added to the query. -ach filter has a different effect on each lithoclass. The entropy filter reduces more the medium and coarse ands than the data quality criteria on the borehole density and positional quality. This is because the amount »f samples containing these lithoclasses are smaller than for the fine sand lithoclass, hence the entropies are ligher. With this, it is important to highlight that statistics have a big effect on the model. Figure 22. An example of a query in the TILES DSS using different qualitative resource criteria, here for the *Experience-based, from calculating the mean of the entropy distribution of various areas in the BPNS ( means no uncertainty, whereas 1 occurs when all lithological classes have the same probability). **British Standard for ‘Aggregates for concrete’: BS EN 12620.
Figure 23. Simulated residual currents in the Flemish Banks: vector plot of the residual currents on top of the bathymetry (left), and tidal ellipse showing the dominant SW-NE axis (right). The scenario investigates the behaviour of the system over a spring-neap cycle under average meteorological forcing (winds) and waves. Figure 23 shows the simulated currents in the Flemish Banks area: they are strongest along the SW-NE axis, leading to the characteristic ellipse of residual currents (Figure 23, right).
Figure 23. Simulated residual currents in the Flemish Banks: vector plot of the residual currents on top of the bathymetry (left), and tidal ellipse showing the dominant SW-NE axis (right). The scenario investigates the behaviour of the system over a spring-neap cycle under average meteorological forcing (winds) and waves. Figure 23 shows the simulated currents in the Flemish Banks area: they are strongest along the SW-NE axis, leading to the characteristic ellipse of residual currents (Figure 23, right).
Figure 24. Simulated sediment-transport ellipse for the different sediment fractions and the overall sediment transport. From left to right (a-e): Clay + Silt, Fine Sand, Medium Sand, Coarse Sand + Gravel, and overall transport.
Figure 24. Simulated sediment-transport ellipse for the different sediment fractions and the overall sediment transport. From left to right (a-e): Clay + Silt, Fine Sand, Medium Sand, Coarse Sand + Gravel, and overall transport.
The simulated processes related to sediment transport (bedload, suspended load, erosion, deposition) thus lead to different results as a function of the distinct parameterization between sediment fractions. As a consequence, the simulated qualitative and quantitative changes of the seabed itself are different. To illustrate this, a hypothetical extraction scenario is used. In this scenario, ~310 10° m? sediment are removed from the original bathymetry (Figure 25). This total removal corresponds to “60% of the total volumes available for extraction as defined by FPS Economy, Continental Shelf Service based on the geological surfaces generated in TILES (Degrendele et al., 2017). At a present rate of extraction (~3 10° m?/y), this would correspond to ~100 years of extraction. The extraction in this scenario extends the current extraction spatial pattern until the total quantity is removed. total quantity is removed.
The simulated processes related to sediment transport (bedload, suspended load, erosion, deposition) thus lead to different results as a function of the distinct parameterization between sediment fractions. As a consequence, the simulated qualitative and quantitative changes of the seabed itself are different. To illustrate this, a hypothetical extraction scenario is used. In this scenario, ~310 10° m? sediment are removed from the original bathymetry (Figure 25). This total removal corresponds to “60% of the total volumes available for extraction as defined by FPS Economy, Continental Shelf Service based on the geological surfaces generated in TILES (Degrendele et al., 2017). At a present rate of extraction (~3 10° m?/y), this would correspond to ~100 years of extraction. The extraction in this scenario extends the current extraction spatial pattern until the total quantity is removed. total quantity is removed.
Figure 26. Simulated deposition (blue, positive bathymetric change) and erosion (red, negative bathymetric change) pattern for the three scenarios: simulation with one average sediment fraction homogeneous over the whole area (left); simulation with multiple sediment fractions after coupling with the voxel model (center); and difference map between the results for the two simulations (right), highlighting the importance of the geological information exploited in the multi-fractions simulation.
Figure 26. Simulated deposition (blue, positive bathymetric change) and erosion (red, negative bathymetric change) pattern for the three scenarios: simulation with one average sediment fraction homogeneous over the whole area (left); simulation with multiple sediment fractions after coupling with the voxel model (center); and difference map between the results for the two simulations (right), highlighting the importance of the geological information exploited in the multi-fractions simulation.
Figure 27. Relationship between the measured depth difference (MBES data) and the depth difference attributed to aggregate extraction (EMS data) in the six areas over the investigated period in this study. To a large extent, the measured depth difference is explained by the depth difference due to extraction (Degrendele et al. (2014). Using corrected and additional data from other extracted and monitored areas, the slope of the regression line is closer to 1. For legend of the areas, see Table VI. On average, bathymetric changes (MBES data) between successive campaigns in monitored areas were observed to correspond to the depth differences expected from human extraction (EMS data Roche et al., 2011, 2013; Degrendele et al., 2014). This is illustrated for the monitored areas of Zone 2 (Flemish Banks) in Figure 27: in this approach, each area is binned into sectors of extractior intensity (depth difference due to extraction from 0 to 1, 1 to 2, 2 to 3, ... 10? m?/ha) within whic the mean observed difference in bathymetry is compared to the mean expected decrease in deptt due to extraction (see Roche et al., 2011). When the potential systematic error affecting MBES bathymetric measurements is corrected and when additional data from other extracted anc monitored areas are used, the volumes of sand extracted estimated from EMS data are in very higt accordance with those deduced from MBES surveys (Debese et al., submitted). accordance with those deduced from MBES surveys (Debese et al., submitted).
Figure 27. Relationship between the measured depth difference (MBES data) and the depth difference attributed to aggregate extraction (EMS data) in the six areas over the investigated period in this study. To a large extent, the measured depth difference is explained by the depth difference due to extraction (Degrendele et al. (2014). Using corrected and additional data from other extracted and monitored areas, the slope of the regression line is closer to 1. For legend of the areas, see Table VI. On average, bathymetric changes (MBES data) between successive campaigns in monitored areas were observed to correspond to the depth differences expected from human extraction (EMS data Roche et al., 2011, 2013; Degrendele et al., 2014). This is illustrated for the monitored areas of Zone 2 (Flemish Banks) in Figure 27: in this approach, each area is binned into sectors of extractior intensity (depth difference due to extraction from 0 to 1, 1 to 2, 2 to 3, ... 10? m?/ha) within whic the mean observed difference in bathymetry is compared to the mean expected decrease in deptt due to extraction (see Roche et al., 2011). When the potential systematic error affecting MBES bathymetric measurements is corrected and when additional data from other extracted anc monitored areas are used, the volumes of sand extracted estimated from EMS data are in very higt accordance with those deduced from MBES surveys (Debese et al., submitted). accordance with those deduced from MBES surveys (Debese et al., submitted).
Table VI. Monitoring campaigns used in this study for the six areas in the Flemish Banks. BRMA/BRMB/BRMC: Buiten Ratel, Monitoring area A/B/C; KBMA/KBMB: Kwinte Bank, Monitoring area A/B; ODMA: Oostdijck sandbank, Monitoring area A. For location of the sandbanks, see Fig. 1. When a map of the bathymetric difference is produced between two campaigns, two main pattern are dominating the resulting signal (see Terseleer et al., 2016): the deepening of the seabed due tc human extraction, which is sometimes locally concentrated and intense (up to > 5 m deepening); anc the natural migration of the very large bedforms covering the sandbanks, which results in successiv increase and decrease in bathymetry in accordance with the changing position of the bedform crest and troughs. The remaining variability, grouped here as residual variability, is an indication of the behaviour of the seabed beyond these two main patterns and can provide information in terms o long-term depletion/regeneration rates. In the scope of the TILES project, a protocol was set up t assess the residual variability of the seabed beyond the two main sources of seabed change (migration of very large bedforms, human extraction). It is summarized here (see Terseleer et al. 2016, for details).
Table VI. Monitoring campaigns used in this study for the six areas in the Flemish Banks. BRMA/BRMB/BRMC: Buiten Ratel, Monitoring area A/B/C; KBMA/KBMB: Kwinte Bank, Monitoring area A/B; ODMA: Oostdijck sandbank, Monitoring area A. For location of the sandbanks, see Fig. 1. When a map of the bathymetric difference is produced between two campaigns, two main pattern are dominating the resulting signal (see Terseleer et al., 2016): the deepening of the seabed due tc human extraction, which is sometimes locally concentrated and intense (up to > 5 m deepening); anc the natural migration of the very large bedforms covering the sandbanks, which results in successiv increase and decrease in bathymetry in accordance with the changing position of the bedform crest and troughs. The remaining variability, grouped here as residual variability, is an indication of the behaviour of the seabed beyond these two main patterns and can provide information in terms o long-term depletion/regeneration rates. In the scope of the TILES project, a protocol was set up t assess the residual variability of the seabed beyond the two main sources of seabed change (migration of very large bedforms, human extraction). It is summarized here (see Terseleer et al. 2016, for details).
Figure 28. Depletion (negative seabed change, in red) and regeneration (positive seabed changes, in blue) rates in the six areas, beyond dune migration and sand extraction (see text for details).
Figure 28. Depletion (negative seabed change, in red) and regeneration (positive seabed changes, in blue) rates in the six areas, beyond dune migration and sand extraction (see text for details).
Figure 29 (a-b). Density plots and boxplots (right) of the depletion/regeneration rates for each area. the six areas, and the boxplots in Figure 29b summarize the values. Figure 28 shows that the elimination of the dune migration is not perfect. Indeed, dune migration is not homogeneous over the dune fields of each area, and successive bands of accretion and erosion are locally visible where the migration was not fully compensated by the optimized overall dune migration. Yet, this effect is reduced and the remaining variability is considered to reflect the residual behaviour of the seabed. Figure 29a shows the density plot of these depletion/regeneration rates for the six areas, and the boxplots in Figure 29b summarize the values.
Figure 29 (a-b). Density plots and boxplots (right) of the depletion/regeneration rates for each area. the six areas, and the boxplots in Figure 29b summarize the values. Figure 28 shows that the elimination of the dune migration is not perfect. Indeed, dune migration is not homogeneous over the dune fields of each area, and successive bands of accretion and erosion are locally visible where the migration was not fully compensated by the optimized overall dune migration. Yet, this effect is reduced and the remaining variability is considered to reflect the residual behaviour of the seabed. Figure 29a shows the density plot of these depletion/regeneration rates for the six areas, and the boxplots in Figure 29b summarize the values.
Table VII. Percentile 10 (p10) and 90 (p90) of the depletion/repletion rates for each area. The median depletion/repletion rate is close to 0 m yr (no bathymetric change), but is more negative for BRMC. This area has a skewed distribution towards negative bathymetric changes (Figure 30). Interestingly, the distribution of locations experiencing erosion or deposition in BRMC shows a clear spatial pattern (Terseleer et al., 2016): erosional areas are concentrated on the crests of the very large dunes covering the sandbank (mostly at the northern corner of the area; Figure 29c), and accretion marks the extraction pit (along an axis perpendicular to the sand-wave crests, in the southeastern part of the area). Figure 30 shows the relationship between the depletion/repletion rate and the extraction rate: in areas experiencing a non-negligible extraction (> 0.1 m yr’, i.e. BRMA, BRMC, KBMB, and ODMA), a positive relationship exists between the depletion/repletion rate and the extraction rate. In other words, and as more clearly visible for BRMC (which experienced the largest extraction), the locally most heavily extracted locations tend to experience residual accretion. Across the six areas, the depletion/regeneration rates vary from -1.21 to 1.24 m yr” (in BRMC), but 90% of the surface within these areas experience a depletion/regeneration rate between -0.21 and 0.12 m yr”. Table VII shows the percentile 10 and 90 for each area: given that each value represent one pixel from Figure 29 (corresponding to 1 m2), these values encompass 80 % of the surface of each area.
Table VII. Percentile 10 (p10) and 90 (p90) of the depletion/repletion rates for each area. The median depletion/repletion rate is close to 0 m yr (no bathymetric change), but is more negative for BRMC. This area has a skewed distribution towards negative bathymetric changes (Figure 30). Interestingly, the distribution of locations experiencing erosion or deposition in BRMC shows a clear spatial pattern (Terseleer et al., 2016): erosional areas are concentrated on the crests of the very large dunes covering the sandbank (mostly at the northern corner of the area; Figure 29c), and accretion marks the extraction pit (along an axis perpendicular to the sand-wave crests, in the southeastern part of the area). Figure 30 shows the relationship between the depletion/repletion rate and the extraction rate: in areas experiencing a non-negligible extraction (> 0.1 m yr’, i.e. BRMA, BRMC, KBMB, and ODMA), a positive relationship exists between the depletion/repletion rate and the extraction rate. In other words, and as more clearly visible for BRMC (which experienced the largest extraction), the locally most heavily extracted locations tend to experience residual accretion. Across the six areas, the depletion/regeneration rates vary from -1.21 to 1.24 m yr” (in BRMC), but 90% of the surface within these areas experience a depletion/regeneration rate between -0.21 and 0.12 m yr”. Table VII shows the percentile 10 and 90 for each area: given that each value represent one pixel from Figure 29 (corresponding to 1 m2), these values encompass 80 % of the surface of each area.
Figure 30. Depletion/regeneration rates vs extraction rates in the six investigated areas. The smoothing curves illustrate the global trend of the relationship (they are produced with a generalized additive model using cubic regression splines). Two observations are made here: first, the regeneration rate is reduced compared to the extraction rate (on average, about one order of magnitude lower); second, the positive regeneration rates in some locations are accompanied by negative depletion rates elsewhere, leading to a global balance in bathymetric change within each area which is close to zero (or slightly negative in the case of the most heavily extracted area, BRMC). Altogether, these observations seems to indicate that the large- scale regeneration potential after human extraction is very limited. Moreover, these regeneration vs depletion dynamics suggest a possible flattening of the seabed, driven by local sediment redistribution. This can negatively affect the longer-term diversity in these areas: while a loca biodiversity increase has been observed at a small scale after disturbance by human extraction (De Backer et al., 2014), a longer-term and larger-scale seabed homogenization would negatively affect the habitat potential and biodiversity in areas of marine aggregate extraction. These observations can be exploited in decision making about the management of exploited areas and deserve further investigation: in the next step, the analysis is to be generalized to all monitored areas and actualized with the most recent data in order to further investigate the processes Two observations are made here: first, the regeneration rate is reduced compared to the extraction
Figure 30. Depletion/regeneration rates vs extraction rates in the six investigated areas. The smoothing curves illustrate the global trend of the relationship (they are produced with a generalized additive model using cubic regression splines). Two observations are made here: first, the regeneration rate is reduced compared to the extraction rate (on average, about one order of magnitude lower); second, the positive regeneration rates in some locations are accompanied by negative depletion rates elsewhere, leading to a global balance in bathymetric change within each area which is close to zero (or slightly negative in the case of the most heavily extracted area, BRMC). Altogether, these observations seems to indicate that the large- scale regeneration potential after human extraction is very limited. Moreover, these regeneration vs depletion dynamics suggest a possible flattening of the seabed, driven by local sediment redistribution. This can negatively affect the longer-term diversity in these areas: while a loca biodiversity increase has been observed at a small scale after disturbance by human extraction (De Backer et al., 2014), a longer-term and larger-scale seabed homogenization would negatively affect the habitat potential and biodiversity in areas of marine aggregate extraction. These observations can be exploited in decision making about the management of exploited areas and deserve further investigation: in the next step, the analysis is to be generalized to all monitored areas and actualized with the most recent data in order to further investigate the processes Two observations are made here: first, the regeneration rate is reduced compared to the extraction
he TILES DSS, available at http://www.bmdc.be/tiles-dss/, allows resource qualities and quantitie: o be visualised within set uncertainty ranges that are deemed acceptable. Information can be etrieved for any location (user-defined polygon as in Figure 31), but also within predefined area: uch as those designated in the Belgian marine spatial plan. A user can make a series of tailor-made uitability maps that assist in resource assessments. Cross-sections are an extra functionality lowing easy exploration of sandbank architectures. They are also invaluable for outreach anc »ducational purposes. Figure 31 illustrates the sandbank architecture of the Middelkerke Bank yrobably being the best-studied sandbank in the world. Figure 31 shows the most likely lithologica lass with associated volumes, supplemented with the location of the boreholes used to create the ubsurface model. A lacquer peel shows an excellent correspondence with the modelled lithoclasse: or the corresponding location. The alternation of different qualities of sand, clay and shell i: llustrative of how complex a sandbank body can be. The cross-section, as derived via the TILES DSS eflects this complexity and additionally provides insight into the spatial distribution of the ithological units. Figure 31. TILES DSS extract of the Middelkerke sandbank (BPNS), lithological classes with volumes, and lithostratigraphy. A cross-section is made along some vibrocore locations. Right: picture of core 120 showing the heterogeneity of the Quaternary layers.
he TILES DSS, available at http://www.bmdc.be/tiles-dss/, allows resource qualities and quantitie: o be visualised within set uncertainty ranges that are deemed acceptable. Information can be etrieved for any location (user-defined polygon as in Figure 31), but also within predefined area: uch as those designated in the Belgian marine spatial plan. A user can make a series of tailor-made uitability maps that assist in resource assessments. Cross-sections are an extra functionality lowing easy exploration of sandbank architectures. They are also invaluable for outreach anc »ducational purposes. Figure 31 illustrates the sandbank architecture of the Middelkerke Bank yrobably being the best-studied sandbank in the world. Figure 31 shows the most likely lithologica lass with associated volumes, supplemented with the location of the boreholes used to create the ubsurface model. A lacquer peel shows an excellent correspondence with the modelled lithoclasse: or the corresponding location. The alternation of different qualities of sand, clay and shell i: llustrative of how complex a sandbank body can be. The cross-section, as derived via the TILES DSS eflects this complexity and additionally provides insight into the spatial distribution of the ithological units. Figure 31. TILES DSS extract of the Middelkerke sandbank (BPNS), lithological classes with volumes, and lithostratigraphy. A cross-section is made along some vibrocore locations. Right: picture of core 120 showing the heterogeneity of the Quaternary layers.
Figure 32. Prototype of a variable-size voxel model, here of a sandbank. The colour indicates suitability for extraction of medium sands (greyish: highest suitability; green: lowest suitability). Voxel size provides a broad indication of data uncertainty. Small voxel sizes at the surface are indicative of the highly detailed bathymetrical data layer and of samples describing the active layer; increasing voxel size in the subsurface is a function of decreasing borehole density with depth. Figure 32. Prototype of a variable-size voxel model, here of a sandbank. The colour indicates suitability for
Figure 32. Prototype of a variable-size voxel model, here of a sandbank. The colour indicates suitability for extraction of medium sands (greyish: highest suitability; green: lowest suitability). Voxel size provides a broad indication of data uncertainty. Small voxel sizes at the surface are indicative of the highly detailed bathymetrical data layer and of samples describing the active layer; increasing voxel size in the subsurface is a function of decreasing borehole density with depth. Figure 32. Prototype of a variable-size voxel model, here of a sandbank. The colour indicates suitability for
Related topics:
Environmental SustainabilityVoxel Models
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved