Introduction

Sediment geochemical data are crucial for assessing the function of sediments in the ecosystem. The German research project “SECOS” (The Service of Sediments in German Coastal Seas) started in 2013 and included a sediment geochemical mapping as one important tool for later compilation of a complete “Baltic Sea Atlas”, a related database to be used for GIS applications. For some larger parts of the Baltic Sea, sediment geochemistry has already been studied during past decades and sediment geochemical maps have been published by, for example, Emelyanov et al. (1995; Bornholm Basin), Borg and Jonsson (1996; Baltic Sea, trace metals), Emelyanov (2002; Gdansk Basin), Uscinowicz (2011; Polish Baltic Sea), Vallius (2007; Gulf of Finland), and Emelyanov (2012; Gotland Basin).

Modern web-GIS based maps are published at the homepages of the geological surveys of Sweden and Finland. The geochemical maps cover only the land areas and, unfortunately, not the marine areas. Moreover, most earlier geochemical mapping exercises were performed on bulk sediments, rather than on specific grain-size fractions.

For the present study in the south-western Baltic Sea, it was decided to separate the fine fraction (<63 μm, mud) before elemental analyses to search for geochemical signatures, thereby at least minimising the grain-size effect. Accompanied by the large spatial extent (13,600 km2) and coverage (between 600 and 800 data points) of data collection, this aspect proved pivotal in confidently earmarking geochemical “anomalies” in these surface sediments.

Regional setting

The present study and most of the investigated surface sediment samples are based on the German Baltic Sea floor mapping published in 2012 by the Federal agency of Shipping and Hydrography (BSH; Tauber 2012; nine sheets, scale 1:100,000). From the grain-size data of these samples, a classification resulted in the definition of six major sediment types for the south-western Baltic Sea (Fig. 1, Table 1).

Fig. 1
figure 1

Distribution map of major sediment types in the German Baltic Sea (see also Table 1). The blank sector in the western part of the study area (3% of total area) could not be sampled because it represented a restricted military zone

Full size image
Table 1 The six major sediment types characterising the German Baltic Sea floor (cf. also Fig. 1)
Full size table

The German sector of the Baltic Sea is characterised by large and mostly sandy shallow-water areas. Partly gravelly, so-called lag sediments occur in wave-exposed areas from where fine-grained material is removed (e.g. coastal zone, Falster-Rügen Bank and Oder Bank) to be deposited at greater depths in the form of organic-rich mud (e.g. Mecklenburg Bay and Arkona Basin). For land-derived material input, mainly the Oder River and its estuary in the eastern part of the study area is important. Some of the relevant sub-regions and geographical names are included in Fig. 2.

Fig. 2
figure 2

Surficial sediment sampling locations and water depths (bathymetry) of the western Baltic Sea

Full size image

Materials and methods

Sampling and sample treatment

Surface sediments (0–2 cm) were collected by grab sampler (sand, hard ground), box-corer or multi-corer (soft sediment) during several cruises in the period 2000–2013. The field work was performed in blocks designed for the different sub-areas of the German territorial waters, and dependent on ship time. The samples for each sub-area were collected within ca. 2 years. The sample locations are displayed in Fig. 2. The selection of samples for the present investigations took into account a regular coverage of the whole area and a sufficient representation of each of the major sediment types. Complete grain-size data (obtained by sieve- and laser-based analyses) were already available for all the samples (see Tauber 2012) before the geochemical analyses of the present study.

The fine fraction (<63 μm) of the bulk sediment was separated by means of a plastic mesh (to avoid metal contamination) with a mesh size of 63 μm. The original sediment samples were suspended in fresh water to dilute the salt, followed by ultrasonic treatment to destroy aggregates before sieving. The separated fine-fraction suspension was collected in glass beakers, centrifuged to remove the water (and salt) and then freeze-dried for later analyses.

The decision to separate the fine fraction for geochemical analyses follows the recommendation by Perttilä and Brügmann in 1992, developed and later used for sediment environmental monitoring programs (HELCOM 2007). The advantages of this method are that comparable results from different regions and sediment types are achieved, while grain-size effects are excluded (or at least minimised). Additionally, it avoids analytical problems at the detection limit for coarser samples (sand). The present study is the first to employ this method for such a large sector (13,600 km2) and data base (600–800 data points) for the Baltic Sea.

To make the present results comparable with former sediment geochemical mappings of other regions in the Baltic Sea, a recalculation of the element contents of the fine fraction “back” to the original bulk samples was done. Thus, the element contents of the bulk sediment were calculated from the fine-fraction analysis by proportional numerical “dilution” with the coarser part of the samples (weight% >63 μm), which was measured and available for all the samples. This approach seems acceptable if the elements of interest are concentrated (enriched) mostly in the fine fraction, rather than in the coarse fraction. The approach was successfully verified based on a selected subset of samples. For all elements evaluated here, more than 90% of the total element contents could be accounted for by the fine fraction.

Geochemical analyses

Total carbon (TC), N and S were measured directly on dry samples using a Euro EA from Hekatech; and a EA 1110 CHN from CE-Instruments (GC-based methods). Total inorganic carbon (TIC) was determined directly by H3PO4 removal of carbonates with a Multi EA 4000 (IR method) from Analytic Jena. Total organic carbon (TOC) was then calculated as TOC=TC–TIC.

A DMA 80 from Milestone Lab.-Systems was used for direct Hg analyses, heating the sample Hg in a gold trap, heating again, and measuring by AAS. Biogenic opal (SiO2) was determined after basic extraction: 50 mg of the dry sample was treated with 100 ml 1M NaOH for 40 min at 80 °C in shaking steel backers. The resulting suspensions were then centrifuged, filtrated, and prepared for Si analysis with ICP-OES. The method was developed and modified as proposed by Müller and Schneider (1993), and quality assurance was tested by including standard material from Conley (1998).

Total P and the elements As, Cr, Cu, Fe, Mn, Pb and Zn were analysed after 1N HCl extraction from dry samples after burning for 3 h at 550 °C by ICP-OES (iCAP 7000) and ICP-MS (iCAP Q), both from Thermo Scientific. This ash-extraction method was chosen because it is much easier to handle for large sample amounts compared to the total digestion methods, and it allows the detection of the majority of the element contents in relation to total digestion. Validation of the method is based on analyses of 26 samples from the present study area and two different standards using a total digestion method (strong acid mixture, and pressure bomb). For P (this element was the main reason for choosing the ash-extraction method), there is a linear regression of 1.043 x with R2=0.979 and a recovery of 98% ash P from total P. For the other elements, the recovery is Fe: 74%; Mn: 73%; Cr: 65%; Cu: 95%; Pb: 91%; Zn: 92%; As: 89%. Table 2 gives an overview of the analytical methods and quality control.

Table 2 Chemical analyses quality control
Full size table

GIS analysis and map construction

To construct area-wide maps of geochemical parameters, it is necessary to transfer the spatially irregularly distributed data points of laboratory analyses into regular grids. In a first step, the geographic coordinates of the sampling stations given in the geodetic reference system WGS84 were transformed into Cartesian coordinates of the geodetic reference system ETRS89 (UTM zone 33N) by using standardised routines of the software ArcGIS version 10.2. On the basis of these metric coordinates, equidistant grids of 250 m cell size were interpolated in a second step with the software SURFER version 12. Gridding was done by the kriging method using a semi-variogram model. This geostatistical method describing the spatial variance of measurements calculates an experimental variogram, which shows the variance of data points according to their distances in a specified radius. On the basis of this graph a model function was derived, which was used for the interpolation process to enhance the quality of the approximation.

To fit a model function for each geochemical parameter, different parametrisations such as function types, search radius variations and direction controlled calculation were tested. The search radius controls the number of data points used for the interpolation of grid points. Large radii smoothen local structures such as basins and sand bars, whereas smaller ones show such structures in greater detail. The interpolation of all geochemical parameters was done with a maximum of 60 km to include large-scale structures like the Arkona Basin in the north-eastern part of the investigation area in the spanned search radius. Larger radii lead to unwanted smoothing. Direction controlled interpolation was tested by elliptic search fields in west–east and north–south orientations to describe possible transport regimes and fluxes in sediments. Only in sub-areas, like the Oder Bank in the eastern part of the study area, were improvements visible; for the study area as a whole, no consistent direction control could be identified. All final interpolations were done without direction controlled search fields.

As function types are linear, quadratic, spherical and Gaussian functions were used, depending on the experimental variogram of each geochemical parameter. In a final step, the interpolated grids were processed in ArcGIS to develop individual classifications and labelling. Root mean square errors were calculated for each map (Figs. 3, 4, 5, 6) and the corresponding grid, and found to be in acceptable ranges.

Fig. 3
figure 3

Regional distribution of the fine fraction (<63 μm, weight% of total sediment). The fine fraction was separated for geochemical analyses

Full size image
Fig. 4
figure 4

Distribution of organic carbon (weight% TOC): a total sediment, b fine fraction. Note that the scales (legend) of the content ranges are different in a and b, which improves the visibility of the gradients

Full size image
Fig. 5
figure 5

Distribution of phosphorus (weight% P): a total sediment, b fine fraction. Note that the scales (legends) of the content ranges are different in a and b, which improves the visibility of the gradients

Full size image
Fig. 6
figure 6

Distribution of mercury (Hg, μg/kg): a total sediment, b fine fraction. Here a and b have the same scale (legends)

Full size image

Modelling

To relate sediment composition to local ecosystem productivity, net primary production rates were derived from a model simulation using the GETM hydrodynamic model (Burchard and Bolding 2002; Klingbeil and Burchard 2013), to which the latest version of IOW’s ecosystem model ERGOM (Radtke et al. 2012) was coupled via the Framework for Aquatic Biogeochemical Models (FABM; Bruggeman and Bolding 2014). In the present case, a setup with 600 m horizontal resolution and 40 vertically layers was used, as previously validated by Klingbeil et al. (2014) and Gräwe et al. (2015a), and which automatically adapts towards stratification (Hofmeister et al. 2010; Gräwe et al. 2015b). Nutrient load data for the different rivers were compiled from four different sources: the Baltic Environmental Database of the Baltic Nest Institute (http://nest.su.se/models/bed.htm), the data from HELCOM (2015), Lääne et al. (2002) and the German environmental authorities. Three-hourly forecasts of the German Weather Service (DWD), created by the COSMO/EU model, were used as atmospheric forcing. Net primary production was calculated as the difference between phytoplankton gross carbon uptake and respiration.

For validation of model-based primary production, the modelled surface chlorophyll values were compared to satellite-derived estimates provided by the Copernicus website (marine.copernicus.eu), which are calculated from ocean colour measurements following an algorithm developed by Doerffer (2014). These daily estimates were derived from multi-satellite data via the BalAlg algorithm (Pitarch et al. 2016) and cover the complete modelled area and period. In a point-to-point comparison, the data show a correlation of only 49.6%. This is due to inability of the model to predict the exact location of meso-scale structures present in the 1 km spaced observations. If the chlorophyll concentrations are averaged over time (to analyse only the spatial variability), then the correlation coefficient rises to 75.5%.

Results

Altogether about 850 fine-fraction samples were analysed for TOC, TIC, N, S, Hg and biogenic opal (SiO2) and, of these, about 600 samples were additionally analysed for As, Cr, Cu, Fe, Mn, P, Pb and Zn. All measured elements are reported as contents in relation to the dry weight of the sample material (mass per unit mass).

Table 3 gives an overview of statistical parameters of the analytical results. Table 4 shows a correlation matrix of the elements analysed in the fine sediment fraction. The regional distribution of the fine fraction is shown in Fig. 3 (cf. sediment types in Fig. 1: fine-grained mud in the basins, and coarse sediments in the shallower areas of banks, sand flats, and coastal zones). For the geochemical results, three selected constituents (TOC, P, Hg) are reported in Figs. 4, 5, and 6. Finally, the two scatter plots of Fig. 7 depict the correlation of primary production and chlorophyll a to the TOC contents of the fine fraction.

Table 3 Statistical parameters of sample analyses results (n number of samples)
Full size table
Table 4 Matrix of correlation coefficients for TOC, opal, and element contents in the sediment fine fraction
Full size table
Fig. 7
figure 7

Correlation of TOC content in the fine sediment fraction with a average net primary production derived from a numerical model for 2003, b chl-a concentration in the water column determined from satellite measurements (average of all validated measurements in the period 01 April 2015 to 01 April 2016)

Full size image

The fine fraction of sediments is generally considered to be the main carrier and “reactor” of environmentally relevant material fluxes and cycles because the coarse fraction commonly consists of chemically inert quartz grains, rock fragments and silicate minerals, as is the case in the present study area. The elements and other components analysed here can be allocated to three functional groups: (1) biogenically relevant (TOC, N, P, bio-opal), (2) at least partly redox-sensitive (Fe, Mn, P, S), and (3) strong anthropogenic impact (Hg, Pb, Cu, Zn, As). The carbonate content (TIC) of most samples was very low throughout the investigated area, and therefore can be considered as less important within the present context.

In these groups, the three selected constituents can be seen as representative of primary (and secondary) production (TOC, Fig. 4), eutrophication (P, Fig. 5), and anthropogenic impact (Hg, Fig. 6). In each of Figs. 4, 5 and 6, the top map shows the calculated content for the total (bulk) sediment, whereas the bottom map shows the distribution pattern for the analysed fine fraction. In order to better visualise the data and to improve the comparison of the spatial differences (patterns) in the maps, different scaling intervals were used for the total sediment and fine fraction for TOC (Fig. 4) and P (Fig. 5). For Hg (Fig. 6), the same scale was used. Some remarkable and partly unexpected distribution patterns are evident in the maps. Geochemically related elements show similar patterns, due to their dependency on material sources, pathways and cycles (see Discussion).

The correlation coefficients between the analysed elements (Table 4) supports these affinities. The correlation coefficients between most elements of similar properties and/or possible binding partners are highly positive. However, because the separation of the fine fraction is already a form of normalisation, the correlation coefficients are probably lower than expected. The coefficients obtained for total sediment contents (not shown) for all analysed elements are much higher (all above 0.9), being a direct function of the grain-size effect. For general information about the geochemical composition of mud sediments of the Baltic Sea, see also Emelyanov (2012).

The regional distribution of organic carbon (TOC), which mainly reflects the primary production and deposition of young or fresh organic matter (TOC/N ratios mostly ranging between 7 and 9, see Table 3), is illustrated in Fig. 4. Comparing the distribution maps of TOC contents for the total sediment (Fig. 4a) and the fine fraction (Fig. 4b) with that of mud (Fig. 3), some striking differences are observed. The distribution of TOC in the total sediment mirrors that of mud, indicating that TOC contents increase with increasing mud content. The opposite is true for TOC in the fine fraction, where the highest contents occur in areas of low mud contents. Very similar patterns were found for nitrogen and biogenic opal (not shown here).

Similar patterns are observed for total P (Fig. 5). The highest contents of P in the total sediment are positively correlated with the mud distribution (Fig. 5a), whereas the fine-fraction distribution shows a markedly different pattern (Fig. 5b). This applies especially to the south-eastern part of the study area (Pomeranian Bight), but interestingly not to the central part where P is low in both the fine fraction and the total sediment. This general trend may be explained by the fact that P is partly mobile and therefore mainly bound to iron oxy-hydroxide, which is redox-sensitive and can be mobilised by iron reduction due to organic matter degradation.

P geochemistry has been the focus of several studies during the past decade, especially for the Baltic Sea (e.g. Lukawska-Matuszewska and Bolalek 2008; Mort et al. 2010; Jilbert et al. 2011; Puttonen et al. 2014). Because it was not possible to analyse P binding forms by sequential extraction methods (SEDEX; Ruttenberg 1992) for all the samples of this study, a selected sample set (n=10) was analysed in a specialised laboratory in Utrecht, Holland. The SEDEX results are (given as averages in % rel. From total extracted P, which was 900 mg/kg dry weight): easily exchangeable P: 1.2%; Fe-bound P: 25%; authigenic Ca-P: 24%; detritic P: 24%; organically bound P: 26%. The easily exchangeable part is nevertheless important. Possible mobile P, driven by changing environmental conditions (e.g. redox), is the Fe-bound portion if sedimentary conditions switch from oxic to anoxic, with a smaller contribution from degradation of organic matter (organically bound P). This means that about ¼ of the total P in the sediments could be mobilised by short-term changes at the sediment surface. Longer-term changes—e.g. early diagenetic reactions or further mineralisation of organic matter in deeper sediment layers—should lead to strong changes in the different binding forms and the related concentrations or contents in the sediments. This, and additional pore-water measurements, are subject of ongoing and future investigations in the present study area. The strongly varying nature of bottom oxygen concentrations was recently summarised for the study area by Zettler et al. (2017).

In contrast to TOC and P, the distribution pattern of mercury (Hg) shows a more complex pattern (Fig. 6). Similarly to the former two constituents, Hg is positively related to mud content in the total sediment, although restricted to more confined areas (Fig. 6a). In the case of the mud fraction, Hg is partly positively and partly negatively related to mud content, depending on the location in the study area (Fig. 6b). The negative relation is particularly prominent in the southern Pomeranian Bight. This more complex distribution pattern of Hg can be explained by it being mainly an anthropogenic pollutant. Other anthropogenic sourced heavy metals and As (not shown here) are characterised by similar distribution patterns. Particularly, historical hotspots (dumping sites) are seen in the western Arkona Basin and the inner part of Mecklenburg Bay. For the sandy area of the Oder Bank and Pomeranian Bight, the impact of (old) polluted material from the Oder estuary is visible (see also Discussion).

Model results

The TOC content of the fine fraction (Fig. 4b) shows highest values in the Oder Lagoon and Pomeranian Bight. To assess whether these values originate from high local primary production, TOC was correlated with modelled primary production (Fig. 7a) and measured chl-a values (Fig. 7b). The comparison is restricted to a longitudinal corridor between 13.5°E and 14.3°E (blue crosses in Fig. 7) to characterise the transitional area between the Oder Bank and the Arkona Basin. Values east of 14.3°E are shown in orange to highlight the data from the shallow Oder Bank.

The comparison with both measured chl-a and modelled net primary production (NPP) shows essentially the same trend, but with low correlation. Whereas high local NPP is associated with elevated TOC contents in the fine fraction, the reverse is not always the case, i.e. in low-productivity areas the TOC contents can be high or low. High TOC contents can be observed even in areas of relatively low NPP.

Discussion

For further interpretation of the present results, it is necessary to first consider the relations between freshly deposited material, often called “fluffy layer”, and the 2 cm surface sediment samples used for this study. This concerns the age or time span represented in the analysed material, and possible impacts related to temporal variability and various sediment mixing processes. The fluffy layer likely results from primary production but is also affected by re-suspended and re-deposited (older) material from lateral transport driven by intensive hydrodynamics. For the eastern part of the study area, this has been investigated by Löffler et al. (2000), Kuhrts et al. (2004), and Tauber and Emeis (2005). Dating by radio-nuclides (e.g. 210Pb, 137Cs) is rather difficult or impossible even in the mud-filled basins of the study area (see Leipe et al. 2013; Moros et al. 2017). The 2 cm sediment surface samples would contain the fluffy layer material plus older material mixed by bioturbation, hydro-turbation and direct anthropogenic impact (ground fishery). This concerns both the fine-grained muddy sediments of the basins as well as the sands of the shallower banks.

There seems to be no way to restrict the time span to specific years or to separate what portion is old and what is young material. Therefore, the element maps shown in Figs. 4, 5 and 6 likely represent the average state resulting from variable depositional conditions of the past decades.

Kersten et al. (2005) used the short-life 234Th isotope for investigating storm-driven disturbances and transport in muddy sediments of the Mecklenburg Bight, and found a large area impacted by an old dumping site of contaminated material. For this example and the present study, the anthropogenically sourced impact/input of heavy metals (e.g. Hg) can be used as tracer for the reconstruction of transport pathways, and the time span is restricted by the pollution history. The deeper, mud-filled basins in the present study area can be regarded as depositional sites (albeit not undisturbed) in comparison to the sandy banks and coastal zones where the fine-fraction content is lower and the hydrodynamics is stronger.

The grain-size effect

A general comparison of the two maps (total sediment and fine fraction) for each of the elements discussed here shows evidence of the so-called grain-size effect. According to the expected transport energy, the fine-grained material (as major carrier of these elements) mainly deposits in the basins, whereas the coarser sediments (sands) of the shallower banks and coastal zones are depleted in fine-grained material. This simple sorting is driven by decreasing energy, but not necessarily leading to differences of element contents inside the fine fraction for areas of high or low proportions of this fine fraction. Indeed, in the present case local geochemical “anomalies” can plausibly be explained rather in terms of point source or concentration/enrichment effects rooted in chemical and/or biological reactions (see below).

Nutrients, productivity, benthic fauna

The eastern part of the study area, the Pomeranian Bight with the sand-covered Oder Bank in its centre, is a good example of a highly productive sector. In the south, the Oder River estuary represents an important source for nutrients (Voss and Struck 1997) and pollutants. The whole bight is characterised by high primary production from spring to autumn, besides being densely colonised by benthic organisms (Glockzin and Zettler 2008). At least three important species of benthic macrofauna (Mya arenaria, Macoma balthica, Cerastoderma glaucum) are present with remarkably high biomasses in sandy areas (Darr et al. 2014). The nutrient chain, starting from primary production (e.g. diatoms) to filter-feeding bivalves and eventually predators, is short (cf. shallow waters) and very intense. Huge flocks of sea-ducks overwinter in this internationally recognised bird resting site, feeding mainly on benthic macrofauna (Kube 1996).

The Pomeranian Bight and Oder Bank constitute a hotspot for biogeochemical material cycling. However, the bight is not a (long-term) depositional area for fine-grained material. Organic matter and fine particle-bound nutrients and pollutants are transported towards the adjacent basins (Arkona Basin, Bornholm Basin). Due to the high energy input, strong re-suspension occurs and the residence time for fluffy layer material has been estimated to be in the range of weeks to months (Christiansen et al. 2002; Emeis et al. 2002; Kuhrts et al. 2004). Nevertheless, some of the fine fraction in this sandy area is remarkably enriched in organic carbon and nutrients (P, bio-SiO2).

The correlation of TOC in the fine fraction with chl-a and net primary production (Fig. 7), however, shows that high local TOC contents do not necessarily originate from local production. While high local productivity values always correspond to TOC enrichment in the fine fraction, the opposite is not true, high TOC values being found also in the easternmost part of the Oder Bank. In this relatively exposed area, chl-a values and primary productivity are much lower compared to the immediate proximity of the coast or even the Oder Lagoon. One possible explanation could be lateral transport processes, the fine fraction deposited here possibly originating from even more heavily euthrophied sites close to the Oder River mouth from where it is transported by repeated sedimentation and re-suspension. The high Hg content in this area (Fig. 6b) supports this hypothesis. Nevertheless, the real causes of element enrichment in the fine sediment fraction of the shallow Oder Bank remain unclear. Probably the geochemical function of benthic fauna (e.g. filter feeders) is more important than one would suppose but this needs confirmation in further investigations (see also Gogina et al. 2017).

Pollution

Mercury (Hg) can be used as a strong indicator for anthropogenically sourced environmental pollution (Fig. 6). Two historical hotspots are clearly visible in both Hg distribution maps. The industrial waste dumping site in the Mecklenburg Bight dates back to the 1960s, and the hotspot in the Arkona Basin reflects post-World War II dumping activities. Both are situated in mud-filled basins and have already been pinpointed during former investigations (see Leipe et al. 2013). An interesting new result is that the distribution of Hg in the fine fraction of southern Pomeranian Bight sediments (Fig. 6b) reveals the ongoing transport of (old) contaminated material from the Oder Lagoon to the open Baltic Sea. The soft sediments of the Oder Lagoon act as a stopover of pollutants (and nutrients) from historical times when pollution of the river was at its maximum (ca. 1960–1990; see Leipe et al. 1998). It is mainly the wind-driven re-suspension of the surface sediment in the shallow lagoon (average depth 3.8 m) which regularly leads to material transport with the river discharge into the open Baltic Sea (Tejakusuma 2004).

Beldowski and Pempkowiak (2003) found similar results for the Gdansk Basin, Vistula River mouth region in Poland, with a transport chain from the river (source) to the depositional basins. They observed an additional impact of (chemical) remobilisation caused by changes in speciation of Hg, which could not be determined in the present study. An impoverishment or decreasing trend of Hg contents with increasing distance from hotspots or sources (chemical aureole) is clearly visible in the fine-fraction Hg map. However, a general decrease of pollution over, for example, the past two decades can be reconstructed only in well-dated undisturbed sediment cores (see Moros et al. 2017, and comments above). Elevated Hg contents have also been recorded in Kiel Bay, indicating contamination by local industry and harbour activities.

Conclusions

Further investigations should focus on estimating volume-related element concentrations (see Flemming and Delafontaine 2000, 2016; Leipe et al. 2011) considering sediment water content and dry bulk density based on the results of this study. This would allow to determine inventories and even accumulation rates, if dating is possible for selected sediment cores from the sub-areas discussed here. Questions related to material cycles and diagenetic processes would furthermore require the analyses of pore-water profiles and element speciation for P, Ca, Fe and trace metals (e.g. Hg; see Beldowski and Pempkowiak 2003, 2007).

For geochemical investigations of large areas with a heterogeneous sediment distribution, a separation of the fine fraction before analyses (similar to monitoring programs) is recommended and offers several advantages. The grain-size effect (“dilution” by sand) can be excluded and, additionally, geochemical “anomalies” become visible. In the present study area in the south-western Baltic Sea, the data reveal local hotspots of, for example, Hg of mainly anthropogenic origin (dumping sites, river discharge). Furthermore, and rather unexpectedly, there are enrichments of TOC and elements (e.g. P) in shallow-water areas (sandbanks) which could be related to local conditions involving, for example, dense benthic fauna or biogeochemical reactions at the sediment–water interface. The distribution of local or sub-regional primary production alone could not be confirmed as the main source. For further investigations of material sources, pathways and cycles, the deposition of biogeochemical and environmentally relevant elements, as well as (volume related) concentration-based maps, the presented examples of element maps are the first step. Including the original data behind these maps, modern web-GIS portals are surely the future of inter- and multidisciplinary science and a powerful interface for public use.