Identification of the natural background of phosphorus in the Scheldt river using tidal marsh sediment cores

Elevated phosphate (PO4) concentrations can harm the ecological status in water by eutrophication. In the majority of surface waters in lowland regions such as Flanders (Belgium), the local PO4 levels exceed the limits defined by 15 environmental policy and fail to decrease, despite decreasing total phosphorus (P) emissions. In order to underpin the definition of currents limits, this study was set up to identify the pre-industrial background PO4 concentration in surface water of the Scheldt river, a tidal river in Flanders. We used the sedimentary records preserved in tidal marsh sediment cores as an archive for reconstructing historical changes in surface water PO4. For sediment samples at different depths below the sediment surface, we dated the time of sediment deposition and analysed the extractable sediment-P. The resulting time series of sediment-P was 20 linked to time series of measured surface water PO4 concentrations (data 1967-present). By combining the sediment-P and water-PO4 data, the sorption characteristics of the sediment could be described. Those sorption characteristics allowed us to estimate a pre-industrial background surface water PO4 levels, based on deeper sediment-P that stabilised at concentrations smaller than the modern. In three out of the four cores, the sediment-P peaked around 1980, coinciding with the peak in surface water PO4. The estimated pre-industrial (~1800) background PO4-concentration in the Scheldt river water was 62 [57; 66 25 (95%CI)] μg PO4-P/L. That concentration exceeds the previously estimated natural background values for lakes in Flanders (15-35 μg TP/L) and is about half of the prevailing limit in the Scheldt river (120 μg PO4-P/L). In the 1930s, river water concentrations were estimated at 140 [128; 148] μg PO4-P/L, already exceeding the current limit. The method developed here proved useful for reconstructing historical, background PO4 concentrations of a lowland tidal river. A similar approach can apply to other lowland tidal rivers to provide a scientific basis for local, catchment specific PO4 backgrounds. 30 https://doi.org/10.5194/bg-2021-100 Preprint. Discussion started: 20 April 2021 c © Author(s) 2021. CC BY 4.0 License.


Introduction
Elevated phosphorus (P) concentrations can harm the ecological status in water by promoting excess levels of primary production, i.e. eutrophication, which yields subsequent anoxia, blooms of toxic blue-green algae and generally affects biodiversity (Correll, 1998). Therefore, limiting P concentrations in the surface water is crucial to ensure a good ecological 35 status. Phosphorus in surface waters is present in solution (dissolved P) and in the suspended matter. The dissolved P can exist as inorganic phosphate (PO4), organic P forms, or PO4 sorbed to mineral colloids. The reactive P, determined by a colour reaction, is almost identical to PO4. The environmental limits are either expressed as reactive P (equated to PO4-P limits), as total P (TP) limits or both.
Many lowland regions, like Flanders (Belgium), do not achieve good water quality mainly due to the excess of nutrients. 40 Currently, the average PO4 in Flanders stabilised at 290 μg PO4-P/L, well above the river type-specific limits of 70-140 μg PO4-P/L (Smolders et al., 2017;VMM, 2018). Despite the current net-zero P-balance in agricultural soils, the situation did not improve since 2004. Therefore, the question arises when or, even, if these limits can be achieved. Natural background concentrations of lowland river PO4-P may be higher than those of upland rivers because of biogeochemical processes typical for such waters. Ferric iron-(Fe(III)) and aluminium-oxyhydroxides have a high affinity for PO4-anions and adsorb PO4, 45 limiting the P-solubility (Borggaard, 1990;Holtan et al., 1988). However, anoxic conditions lead to reductive dissolution of Fe(III) minerals. As a result, the associated P is released again to the overlying water when the sediment is strongly reduced (Baken et al., 2015). It is now well established that such conditions explain the typical summer peaks in PO4 and that regional differences in sediment Fe concentrations explain regional differences in surface water PO4 concentrations in Flanders (Smolders et al., 2017). 50 Since 2000, the European Union regulates surface water quality with the Water Framework Directive (WFD). The WFD does not prescribe limits for P in rivers and lakes but provides a framework for local regulations. The local maximum P concentrations are based on identifying pristine environments with minimal anthropogenic disturbance, i.e. the natural background (EU-Parliament, 2000). Flanders is densely populated, and it was not possible to locate pristine or reference lakes in the current environment. Instead, natural background TP concentrations for Flanders have been inferred from reference 55 lakes sampled in Central and Baltic states in Europe that were identified for the WFD (Cardoso et al., 2007). Based on that study, background TP concentrations in lakes of Flanders were set at 15-35 µg P/L, selected for lakes with representative depth and alkalinity. For rivers, no TP or PO4 natural background has been established.
The definition of the natural background has been subject to debate in many river basins (Matschullat et al., 2000;van Raaphorst et al., 2000). The natural background concentrations defined here are those concentrations found in the environment 60 in the absence of any human activity, reflecting only natural geochemical processes (Laane, 1992;Reimann and Garrett, 2005).
This definition implies that concentrations have to be estimated before the presence of human activity. This is not always feasible and, therefore, a pre-industrial background can be defined instead, which is inferred from samples dating before the industrial revolution (Reimann and Garrett, 2005). https://doi.org/10.5194/bg-2021-100 Preprint. Discussion started: 20 April 2021 c Author(s) 2021. CC BY 4.0 License.
The pre-industrial background is logically affected by anthropogenic processes. In Belgium, the industrial revolution started 65 around 1800 (Vanhaute, 2003). At that time, the Belgian population was about three million, and large scale agriculture dates back to the middle ages and the Roman period. However, the most substantial increase in nutrient emissions occurred after the 1950s due to sewer infrastructure, mineral fertilisers and P-loaded detergents (Billen et al., 2005;van Raaphorst et al., 2000).
Because P emissions mainly originate from point sources due to domestic loading, the increase of surface water P between 1950 and 1975 relates to the rise in population connected to sewer systems (Billen et al., 2005). Since 1985, the increase of 70 wastewater treatment allowed a significant improvement to the situation (Billen et al., 2005).
River floodplains provide sediment archives from which region-specific background concentrations could be derived. The sediments deposited onto floodplains bonded with surface water PO4 and are deposited at a predictable rate. Those sediments can serve as an archive for reconstructing historical P emissions trends and provide useful spatial and temporal information on historical P concentrations in adjacent water bodies (Birch et al., 2008). Boyle et al. (2015) used P profiles from lake sediments 75 in the UK to infer the historical evolution in population density in catchments over 10.000 years. Similarly, banded iron formations found in deep oceanic waters allowed to infer oceanic P concentrations from over two billion years ago (Bjerrum and Canfield, 2002).
In lowland rivers with tidal influence, also called estuaries such as the Scheldt estuary in Flanders, tidal marshes are the analogue of floodplains along fluvial rivers. Tidal marshes are vegetated ecosystems located along the tidal portion of rivers 80 or coastlines, which periodically flood during high tide and storm events. Due to their net accumulation of sediments, tidal marshes increase in elevation over time (Temmerman et al., 2003a). Tidal marshes directly adjacent to tidal rivers are regularly flooded during high tides, so river sediments and associated elements like P are deposited within the dense marsh vegetation (Friedrichs and Perry, 2001;De Swart and Zimmerman, 2009;Temmerman et al., 2004a). Researchers have used tidal marshes as sediment archives of deposited substances other than P, such as organic carbon (Van de Broek et al., 2019) and silicon 85 (Struyf et al., 2007).
However, it remains to be investigated to what extent P concentrations measured in tidal marsh sediment archives can be used to reconstruct historical changes in PO4 concentrations in the adjacent estuary. Therefore, in this manuscript, we test and evaluate a methodology to estimate the historical surface water PO4 concentrations. We described the tidal marsh sediment sorption characteristics by linking the P concentration in dated tidal marsh sediments to historical measurements of PO4 in the 90 Scheldt river water. Those sorption characteristics allowed us to estimate river water PO4 concentrations based on a sediment analysis. The underlying assumption is that P remains immobile in these sediments and that the historical trend of PO4 in the Scheldt river is reflected in this depth profile. Thereby, we argue that the sediment P-composition in deeper sediment layers of tidal marshes will reconstruct the historic PO4 concentration of adjacent water bodies. A database containing measurements of the PO4 concentration in the Scheldt river's surface water (1967-current) will verify this assumption. The hypothesis we test 95 is that the previously estimated natural background P in this major lowland river is larger than that what was estimated earlier for lakes (15-35 µg P/L) because these previous estimates did not account for the internal loading processes typical for lowland rivers.

Study area 100
Freshwater tidal marshes were sampled at four locations along the Scheldt river ( Fig. 1, Table S1). The Scheldt estuary is located in northern Belgium and the south-western Netherlands, where it flows into the North Sea. The river basin of the Scheldt covers a large part of Flanders (71%) and the adjacent region of Northern France; the total catchment area is approximately 22.000 km 2 . The population living in the river basin is about 10 million (Meire et al., 2005). The tidal wave extends from the mouth (Vlissingen) to 160 km upstream near Ghent, where sluices stop the tidal wave. The estuary's 105 freshwater tidal zone reaches from Ghent to Rupelmonde (Fig. 1). This research focused on freshwater tidal marshes, i.e. located in this freshwater tidal zone of the estuary. Sediment accreting in tidal marshes originates from the deposition of riverine suspended matter, including inorganic mineral sediment and organic matter (Callaway et al., 1996). For the period 1931-2002, the sediment accretion rates were 0.32-3.22 cm/y for freshwater tidal marshes (Temmerman et al., 2004b). We discriminate old and young tidal marshes, hereafter referred https://doi.org/10.5194/bg-2021-100 Preprint. Discussion started: 20 April 2021 c Author(s) 2021. CC BY 4.0 License.
to as marshes. Old marshes have a higher elevation compared to young marshes. As a general mechanism, young marsh surfaces accumulate sediments quickly and increase their elevation asymptotically up to an equilibrium level around the mean 115 high water level (MHWL) (Pethick, 1981;Temmerman et al., 2003a). Temmerman et al. (2003a) defined an old marsh as visible on topographic maps of Ferraris (1774 -1777), so it was formed before the 19 th century (Temmerman et al., 2003a).
Young marshes in the Scheldt estuary were formed more recently, by the natural establishment of pioneer marsh vegetation on formerly bare tidal mudflats, generally after 1944. During the last decades, the young marshes had a surface elevation below MHWL. Young marshes experienced more frequent inundations and therefore had larger sediment accretion rates than old 120 marshes. Old marshes built up slower, at a rate comparable to the rise of the local MHWL (Temmerman et al., 2004a).
This study analysed depth profiles of sediment cores originating from tidal marshes along the freshwater Scheldt river. The analysis contained two old and two young marshes, referred to as Old1, Old2, Young1 and Young2 (locations indicated in Fig.   1). In total, we analysed eight cores; three replicate cores for both sites Old1 and Young 1 and one core each for Old2 and Young2. 125

PO4 concentration in surface waters
The IMIS (Flanders Marine Institute) provided surface water phosphate-P (PO4-P) data measured colourimetrically on a  Undisturbed sediment cores were taken using a gauge auger (0.06 m diameter) at each sampling location. In the field, the cores were divided into subsamples with a 0.03 m interval. The sediment samples were dried at a maximum temperature of 50 °C 145 for 48 hours and crushed to a 2 mm grain size. Macroscopic vegetation residues were removed manually using tweezers (Van de Broek et al., 2018). Bulk density, grain size distribution and organic carbon (OC) content were analysed by Van de Broek et al. (2018). For detailed information, we refer to Van de Broek et al. (2016. https://doi.org/10.5194/bg-2021-100 Preprint. Discussion started: 20 April 2021 c Author(s) 2021. CC BY 4.0 License.

Sediment analysis
The dried sediment samples were analysed for oxalate-extractable P, Fe, Al and Mn (Pox, Feox, Alox, Mnox;Schwertmann, 150 1964). That acid oxalate extractant at pH = 3 targets poorly crystalline oxyhydroxides of Fe, Al and Mn and the associated P.
The poorly crystalline oxyhydroxides are the most reactive due to their large specific surface area (Hiemstra et al., 2010). A sub-sample of each sediment sample was taken, and the extraction was made over two hours at a solid-liquid ratio of 1 g in 50 ml in an end-over-end shaker at 20°C (26 rpm). The suspension was filtered through a 0.45 µm membrane filter (CHROMAFIL ® Xtra PET -45/25). The extract was diluted 20 times and measured by inductively coupled plasma optical emission 155 spectrometry (ICP-OES). The degree of P-saturation (DPS; %) was calculated as in Eq. (1). The DPS represents the ratio of extractable (Pox) to the P sorption capacity of the sediment. This P sorption capacity is estimated as half of the sum of oxalate Feox and Alox, because not all the Fe and Al in a soil is available for P sorption with Feox, Alox and Pox in molar units.
The DPS is expressed in percentage and can be interpreted as the ratio of sorption sites on the sediment occupied by P. Previous 160 research used the DPS to identify agricultural areas sensitive to phosphate leaching and showed a good correlation with pore water P concentrations (Breeuwsma et al., 1995;Lookman et al., 1995;Schoumans and Chardon, 2015;Schoumans and Groenendijk, 2000;van der Zee, 1988).

Age-depth model
The sediment analysis and the surface water PO4 data had to be linked by a corresponding date and location to fit a sorption 165 model. Therefore, an age-depth model was used to calculate the time since deposition of each sediment sample. Temmerman et al. (2004bTemmerman et al. ( , 2004a) developed a model (MARSED) to estimate sediment deposition rates and the resulting evolution of the sediment surface elevation in the tidal marshes of the Scheldt estuary. Hence, we could also use that model to determine the time since deposition of sediments throughout the sampled sediment profiles. The MARSED model simulates the tidal supply of suspended sediments and the settling of these sediments to the marsh surface during tidal inundation cycles integrated over 170 the years to decades. The model was calibrated and validated against measured data on sediment deposition rates on the Scheldt estuary tidal marshes from 1945 until 2002 (Temmerman et al. 2003;2004). The empirical data on sediment deposition rates were derived from radiometric and paleoenvironmental dating of sediment cores at the same locations as sampled for the present study (Temmerman et al., 2004a(Temmerman et al., , 2004b in the previous sentence), for the present study, we wanted to estimate the time of sediment deposition throughout the sediment 180 profiles as good as possible. Therefore, we rescaled the modelled marsh surface elevation by fitting it through the observed depth-age points from the GPS measured marsh surface elevation in 2016 and previously published radiometric and paleoenvironmental dating (Temmerman et al. 2003;2004). This rescaling procedure is explained in the Supporting Information (Fig. S1, S2, S3, S4). The observed age-depth points were available starting from 1958 for sampling site Old1, 1947 for Young1, 1963 for Old2 and 1968 for Young2 (Temmerman et al. 2004). 185 An approximate extrapolation procedure was used to estimate the time of deposition of sediment from depths below the oldest measured age-depth points (mentioned in previous sentence). This extrapolation procedure was only applied for old marshes, which were defined as marshes that existed at least since the end of the 18 th century (Temmerman et al. 2003a;2004). Two sediment cores originated from such old tidal marshes (Old1 and Old2). We know from observed age-depth points that old marshes reached equilibrium with the MHWL before 1944, and that they built up their elevation after 1944 at a rate comparable 190 to the rate of local MHWL rise (Temmerman et al., 2003a). Here, we assumed that between 1800 and 1944, these old marshes also accreted at a rate comparable to the MHWL rise. Historical tide gauge data of MHWL rise was available from 1901 for site Old1 and from 1930 for site Old2 (ScheldeMonitor Team and VNSC, 2020;Temmerman et al., 2003a) and a linear regression of the MHWL against time was used to estimate the marsh surface elevation before 1944 (Fig. S6, S7). However, the accuracy of the dating will be lower going further back in time. Such extrapolation to earlier dates is not appropriate for 195 young marshes, as they only formed after 1950 by pioneer vegetation establishment on formerly bare mudflats (Temmerman et al. 2003a;2004). Those mudflat sediment profiles do not have continuous sedimentary records as tidal marshes and are likely to be disturbed by erosion and sedimentation alternations (Belliard et al., 2019). Therefore, for the young marsh sampling locations, the sediment deposition time could not be extrapolated before the oldest available measured age-depth points, dating back to 1947 on location Young1 and 1968 on location Young2. 200 2.6 Relating surface water PO4 with sediment P: the sediment-water model The age-depth model and linear regression of MHWL provided a deposition year for each sediment sample. Through that agedepth relation the dataset of water PO4 between 1967-2016 was linked to the sediment DPS for each core. The resulting dataset contained all available surface water PO4 readings between 1967 and 2016 closest to the tidal marshes in Temse (n = 1932) and a corresponding DPS value of a sediment sample originating from one sediment core or, when available for Old1 and 205 Young1, a mean of the replicate sediment samples. This dataset allowed to fit a sorption model further termed the sedimentwater model. Schoumans and Groenendijk (2000) presented a Langmuir-type sorption model to predict PO4 concentration leaching from a soil layer based on the DPS Eq. (2).  Groenendijk, 2000;Warrinnier et al., 2018). The model relies on surface complexation between PO4 and Fe-, Aloxyhydroxides in the sediment, determined by a chemical equilibrium between solid (adsorbed) and dissolved PO4 phase (Warrinnier et al., 2019). We did not use the existing sorption models for soils directly, and instead, the parameter K, in Eq.
(2), was calibrated to the most recent Scheldt water data. So that, the K-value adapted to the geochemistry of the tidal marsh 215 sediments.
We explored 16 different scenarios to fit the sediment-water model Eq. (2). These scenarios illustrate the statistical uncertainty surrounding the estimated PO4 concentrations. The model was either fitted separately for each sediment core or on the combined cores Old1 and Young1 (SI.VI). Every sediment sample had between one and three replicates, depending on the depth and the core. We entered either the average value of these replicates or the individual replicate DPS values. One sediment 220 sample covered several deposition years, so multiple PO4 observations corresponded with each sediment sample. Again, the average of all corresponding PO4 readings was taken, or all available values were used separately, resulting in 16 models (Table S2). For each of these, the parameter estimation of Eq. (2)

Evaluation Model Performance
The predictions of the sediment-water model were evaluated by plotting the measured surface water PO4 against predicted PO4 between 2007 and 2016 (Fig. S10). Additionally, the actual over-predicted median ratio was calculated for data points between 2007 and 2016. The prediction of recent years is interesting to evaluate the model's performance because of two reasons. First, the most recent surface water PO4-concentrations are relatively low and more representative of background concentrations. 230 Second, the monitoring data have a high temporal resolution, and the age-depth model is more accurate at shallow depths.

History of surface water PO4 concentrations
The Scheldt PO4-concentrations varied greatly over the past decades, with the peak in surface water PO4-concentrations between 1975 and 1985 (Fig. 2). In Temse, the annual mean concentrations rose from 410 µg PO4-P/L in 1967 and peaked in 235 1980 with 1570 µg PO4-P/L. Between 1990 and 2003, a decrease followed the peak and over the last decade concentrations stabilised between 160 and 200 µg PO4-P /L in Temse. Current PO4-levels are a factor two lower than in 1967 and almost a factor ten lower than the peak in 1980 ( Fig. 2

Sediment cores
The Pox in the sediments ranged between 370 mg P/kg and 13,000 mg P/kg, while the DPS ranged between 13% and 94% (Table 1). In all soil cores starting at the surface, the DPS and Pox increased with depth and peaked at about 0.5 m depth (Fig.   S7, Fig. S8). In deeper (>1.0 m) sediment layers, Pox and DPS decreased and stabilised for Old1, Young1 and Young2 (Table   1). Overall, the Pox increased by, on average, a factor of 3.5 between the surface and the maximum concentrations (Fig. S8,  245 Table1). Based on the age-depth model these peak DPS were deposited between 1960 and 1985 in three of the four sediment cores (Fig. 2). Only the core Old2 reached the peak earlier (ca. 1940-1950). Most importantly for this work, DPS for Old1 showed an apparent stabilisation in deeper or older layers, which indicated undisturbed sediment layers (Fig. 3, Fig. S7).

Old1
Top 4 0 -9 2300 (± 2400) 36 (± 3) Within the first meter, Feox was stable in the three soil cores (Old1, Young1, Old2) with concentrations around 20,000 mg/kg, except for Young2 for which Feox was a factor two larger (Fig. S9). For Young1 and Young2 a decrease in Feox concentration occurred at depths > 1 m. For Old1, Feox showed a steady decline from 20,000 mg/kg at the surface to 10,000 mg/kg at the bottom of the profile (Fig. S9). The Alox concentrations showed a similar trend as the Pox concentrations, with an initial increase followed by a decrease with depth. The strong correlations of Alox and Feox with Pox (rAl = 0.73 and rFe = 0.65) illustrate the 260 positive effect of Fe and Al oxyhydroxides on P sorption.

Sediment core selection 270
Under the assumption that PO4 does not migrate, the tidal marsh sediment cores can provide an archive for river water PO4.
However, that assumption may be most violated at two locations, Old2 and Young2. Considering P-migration, evaluating the distance from a creek within the tidal marsh is crucial. That distance is essential for two reasons. First, within 10 to 20 m of the creeks, the groundwater table fluctuates largely with the tides (Van Putte et al., 2020), which can induce vertical Pmigration. Secondly, sediment accretion is more difficult to predict at closer distances to the creeks and can affect the age-275 depth relation (Temmerman et al., 2003b). The distance to the nearest creek within the tidal marshes are 21 m for Old1, 56 m for Young1, 35 m for Young2 and 5 m for Old2. Furthermore, the profile of Old2 had a peak of Pox at an earlier date (1950) than was expected from surface water data (1980), indicating P-migration (Fig. 2). Consequently, Old2 was not taken up for interpretation of the relation between DPS and PO4. For core Young2, deeper sediment layers had a larger DPS compared to the surface layers (Table 2). Additionally, the age estimation of sediments older than 1968 was not possible due to this tidal 280 marsh's young age, which hampers the interpretation of DPS values from deeper layers as representative for background values.
Furthermore, Feox concentrations were a factor two larger than the other cores (Fig. S9)  The two remaining soil cores, Old1 and Young1, originate from the same tidal marsh area named The Notelaer, located near the city of Temse (Fig. 1) and has been the topic of multiple studies on sediment accretion (Temmerman et al., 2004b(Temmerman et al., , 2003a and soil OC stocks (Van de Broek et al., 2018;Van De Broek et al., 2016). The sediment profiles of both sites Old1 and Young1 rise and fall in DPS comparable to dynamics in surface water PO4-concentrations (Fig. 2, Fig. 3). In deeper sediment layers, DPS and Pox stabilise below levels of recent deposits (Fig. S7, S8). These observations indicated a well-preserved Pox 290 and DPS profile, essential for the DPS-PO4 relation. Therefore, Old1 and Young1 are considered the best profiles for applying the sediment-water model and interpretation of background concentrations.    Table S2. Model 3b (in bold), is proposed as the most accurate one (see text).

Sediment-water model fit
The sediment-water model Eq.
(2) was fitted on DPS-PO4 data from the different sediment cores both individually and for a combination of cores (Table S2). Two observations were omitted because the DPS values were too large (0.99 -1.02) and produced artefacts in the results. The Nash-Sutcliffe model efficiency (E) of the models ranged between 0.04 and 0.85 depending on the input data (Table S2; Nash and Sutcliffe, 1970). The sediment-water model was fitted on the data of each 305 core separately and for the combination of the data from Old1 and Young1, as they came from the same tidal marsh location.
The models fitted on data from sites Old2 and Young2 were not considered as migration possibly affected those cores (crf. section 3.3).
The models fitted on an average DPS (across replicates) associated with individual PO4 readings were considered most suitable (Models 1b, 2b, 3b; Table S2). Models using unique DPS associated with single PO4-data duplicated or even triplicated the 310 PO4-data, artificially creating more degrees of freedom (Model 1c, 2c, 3c; Table S2). Using mean PO4-values artificially reduced the degrees of freedom, compromising the model predictions, by increasing RSE and widening confidence intervals (Models 1a, 2a, 3a; Table S2). The fitted parameter K (L/kg) ranged between 1.0×10 6 and 5.4×10 6 for the different input datasets, with the 95% confidence intervals ranging between 0.8×10 6 and 7.2×10 6 . The variation of parameter K for the different input datasets was larger than the individual confidence limits variation (Table S2). Thus, the uncertainty was more 315 pronounced due to the variability in sediment samples than due to the model fit.

Model performance
The sediment-water model performance was evaluated by comparing the actual by predicted PO4 concentrations over the last decade, as those PO4 concentrations are more comparable to the background. Therefore, we calculated the ratio of predicted over observed PO4 in the surface water of the Scheldt river between 2007 and 2016 and plotted actual by predicted plots (Fig.  320 S10; Table 2). A ratio of predicted over observed below one of indicates that the model underestimates ambient PO4concentrations (mean Temse [2007-2016 = 170 µg PO4-P/L). Model 1b overestimated recent observations, which is unwanted for calculating the background, and was therefore considered unsuitable. In contrast, half of the observations were underestimated by at least 25% for model 2b and by 11% for model 3b ( Table 2). The actual by predicted plots illustrate the same (Fig. S10). Based on these observations, model 3b was considered as the best model, although the residual standard error 325 (RSE) was lower for model 2b (Table S2). Model 3b predicted recent PO4-concentrations best, with median underestimation of only 11% ( Table 2). The selected model 3b succeeds to reconstruct the rise and fall in surface water PO4-concentrations based on the sediment characteristic DPS (Fig. 4).  (Fig. 2). Before 1930, modelled PO4-concentrations stabilise at levels below current observations (Fig. 4).

Estimating background PO4 concentrations in the Scheldt river 340
The deepest sediment layers are most suitable for predicting background PO4 concentrations of the Scheldt river water. These layers are the oldest and expected to have experienced the least impact of P additions from anthropogenic sources. The Old1 marsh site was appropriate for this purpose as it developed before 1774, which dates before the industrial revolution in Belgium. The average DPS for the bottom sediments, dated between 1800 and 1840, was 20% for core Old1 (Table 1 (Table 2).

Background vs ambient PO4 concentrations 350
This work presents a novel approach to reconstruct background surface water PO4 concentration in a tidal river using the DPS of adjacent tidal marsh sediments. The background concentration is essential in the context of developing local nutrient limits.
The predicted pre-industrial background concentration (62 µg PO4-P/L; Table 2) is about half of the current surface limit of the Scheldt (120 µg PO4-P/L; Flemish Government, 1995). Remarkably, the predicted background concentrations are about a factor two larger than the background estimates of lake waters for Flanders today (15-35 µg PO4-P /L; Cardoso et al., 2007), 355 suggesting that the internal loading related to summer anoxia in lowland rivers, contributes naturally to larger PO4 concentrations in lowland rivers (see introduction).
Our analysis suggests that the pre-industrial PO4 concentration was about three times lower than the current concentrations in the Scheldt. The mean PO4 concentration of the Scheldt in Temse was 170 µg PO4-P/L between 2007 and 2016. However, in the 1930s, the concentration was estimated at 140 µg PO4-P/L and larger than current limits, at a time before widespread 360 connection to sewer systems, P-loaded detergents, and application of mineral fertilisers.

Limitations of the model
Care needs to be taken with background extrapolations to ensure that post-depositional processes have not modified the biogeochemical patterns and that the area represents the area of interest (Reimann and Garrett, 2005). Several factors can obscure the reconstructed background concentrations. First, vertical migration of P can enrich deeper sediment layers, thereby 365 causing an overestimation of the background. The sediment profiles at the tidal marshes are almost permanently saturated, and the intrusion of P-rich groundwater could affect the P concentrations in the tidal marsh sediment. Moreover, periodic flooding occurs at an approximate range of 300-350 inundations per year, depending on the tidal marsh elevation (Temmerman et al., 2003b). These conditions could favour phosphorus migration due to reductive dissolution of Fe (oxy)hydroxides (Baken et al., 2015). We removed two cores with indications of PO4 migration to address the issue (Old2 and Young2). 370 These cores were identified by comparing the peak in the DPS age-profile with the known peak in surface water PO4 concentrations in the 1980s and by considering the distance from nearby creek ( Fig. 2; Fig. 1). Additionally, the DPS levels of the deepest sediment layers were compared with layers at the surface. For one core the surface layers had lower DPS levels than the deepest layers (Young 2). The two remaining cores (Old1, Young1) had lower DPS-levels in deeper sediment layers (Fig. S7). More importantly, the modelled peak in PO4 concentrations based on the cores Old1, Young1 were found within 375 two years of the monitored peak and had a similar magnitude (Fig. 4). The coinciding peaks illustrate little migration of PO4 in Old1 and Young1, thereby justifying these cores as an archive for water-PO4.
The limited migration is also logical: at the average DPS of 90 % in sediment showing at the peak, the sorption models predict that the solid-liquid P concentration ratio is 2900 L/kg, at the average K value of models of Table 2. That ratio can be converted to dimensionless retardation factor, representing the ratio of the distance migrated by the PO4 compared to the distance traveled 380 by percolating water, of 7500 at a bulk density (ρb) of 1.3 and porosity (θ) of 0.5. With a net vertical annual water percolation of about 2 meters, this corresponds to a net vertical P migration rate of 2.5 cm over 100 years, i.e. vanishingly small (calculation details not shown).
Secondly, there is uncertainty on the age-depth estimation of the sampled sediment profiles. We expect that the age-depth model is most reliable for the Young1 sediment core, as it is based on a fitting of a modelled age-depth relation to four observed 385 age-depth points, while we only had two observed age-depth points available for the other cores. Additionally, observed agedepth points were not older than 1944. Hence, extrapolation of the age-depth model to periods before the older available agedepth points is increasingly uncertain.

Pre-industrial and natural background values
The population increase between 1800 and 1930 can provide a first, and very crude estimate of the population-DPS relation in 390 the Scheldt basin. In 1800 the population in Belgium was around 3 million. In 1930, this number has more than doubled to 7 million (Vanhaute, 2003). A linear relation between both suggests that the DPS is 8% for the pre-anthropogenic pristine environment, corresponding with a PO4-concentration 19-41 µg PO4-P/L, i.e. close to what researchers have suggested for https://doi.org/10.5194/bg-2021-100 Preprint. Discussion started: 20 April 2021 c Author(s) 2021. CC BY 4.0 License. pristine lakes. Clearly, such predictions need to be corroborated with older sediment observations and other archaeological information. The Scheldt river is logically more aerated than several smaller lowland rivers where summer anoxia are naturally 395 more present, i.e. the pristine PO4-P values will be likely higher.

Conclusions
Our study illustrated that tidal marsh sediments can provide an estimate of pre-industrial background PO4-concentrations of the freshwater rivers like the Scheldt river. A sediment assessment can record and time-integrate environmental events, which provides useful spatial and temporal information. Our data estimated that the pre-industrial background concentration is about 400 half of the environmental limits set for surface waters in Flanders and neighbouring countries. Around 1930, the PO4 levels were only about 20% lower than today, which is a remarkably large concentration at af time before the massive application of mineral fertilisers, with lower population density and limited connection to sewer systems. The current PO4 concentrations decreased by a factor ten from the peak found about 40 years ago, reflecting wastewater treatment efforts and reduced diffuse P emission. It is also clear from this study that the pristine, pre-anthropogenic PO4-P concentrations in the Scheldt river are 405 well below the current ambient ones.

Data availability
The results of the sediment data analysis and age depth model are provided in the supplement as csv format. Results of Surface water data are available upon request at the The IMIS (Flanders Marine Institute).

Author contribution 410
FL, ES and PC designed the research. FL conducted the investigation process, and developed the methodology under supervision of ES. MVDB carried out the fieldwork and conceptualized the use of the samples. ST prodived the methodology for the age-depth model and software. TM validated the use of the surface water data. EVM and FL placed the results in perspective with historical data. All the authors contributed to discussion and data interpretations, review and editing of the work. 415