TEMPORAL AND SPATIAL TRENDS OF HEAVY METAL LEAKAGE FROM ACID SULFATE SOILS

TEMPORAL AND SPATIAL TRENDS OF HEAVY METAL

LEAKAGE FROM ACID SULFATE SOILS

Leakage of Ni, Zn, Cu and Fe to freshwater and marine sediments, North-eastern Sweden

Bella Blomkvist

Temporal and Spatial Trends of Heavy Metal Leakage from Acid Sulfate Soils

Leakage of Ni, Zn, Cu and Fe to freshwater and marine sediments, North-eastern Sweden

Temporala och Spatiala Trender i Läckaget av Tungmetaller från Sura Sulfatjordar

Läckage av Ni, Zn, Cu och Fe till sjö- och havssediment i nordöstra Sverige

Bella Blomkvist

Abstract

Acid Sulfate soils (AS) are recognized for being a source of acidity in freshwaters in America, Europe, Australia, and Asia. Apart from the problematic acid leachate, AS soil serves as a possible source of toxic heavy metals in freshwaters, a problem which has received far less

attention than problems related to acid leachate. This study assessed to what extent heavy metals enriched in sediments from the area around lake Persöfjärden (North-eastern Sweden) could be attributed to export from AS soils. I found that: i) Nickel (Ni), Zink (Zn), Copper (Cu) and Iron (Fe) in the sediments had a partial likely origin from AS soils; ii) vertical variations in Ni, Zn, Cu and Fe concentrations suggested events of increased metal transport in the past; and iii) there is a tendency that Zn and Ni assumed from AS soils have caused elevated metal concentration in marine sediment deposited in Persöfjärden. Inferred sediment rates suggest that elevated metal concentrations occurred in sediment deposited during the 18-19^th and 20^th century. I argue that these periods correspond to periods of increased drainage of the AS soils during past agricultural activities in the catchment.

Table of contents

1. Introduction

1.1 2

2. Methods

2.1

Area of investigation

2.2

Sampling

2.2.1 Sampling equipment 5

2.3

Analysis

2.3.1 LOI and XRF analysis 6

2.4

Statistics and data interpretation

3. Results

3.1

Sediment geochemistry

3.2

Historical events in sulfide weathering and denudation

Spatial gradient in sulfide weathering product

4. Discussion

4.1

Origin of Ni, Zn, Cu, and Fe in Persöfjärden

Historical land-use and its effect on element leakage

The spatial gradient of Ni and Zn

5. Conclusions

6. Acknowledgments

7. References

Appendix 1

Appendix 2

Appendix 3

1. Introduction

Sulfate soils are fine grained soils containing iron sulfides. The soils hold the potential to create problematic acidious conditions affecting freshwaters, causing problems in coastal and tropical regions, most commonly in America, Australia, Asia and Europe (Dent and Pons 1995, Proske et al. 2014). When drained, sulfate soils becomes very acidic, and the surroundings become affected by the low pH and heavy metal leakage (Sukitprapanon et al. 2017, Åström and Björklund 1997, Nystrand et al. 2016). The acidity, and sequent heavy metal leakage, affects biota by altering habitats, creating toxic environments (Nystrand et al. 2016). Sulfate soils are formed when organic matter is broken down and reduced into reactive sulfates and trivalent iron, which in turn can form the iron sulfides, often pyrite or iron monosulfate (Becher et al. 2019). As the sulfate soil is exposed to oxic conditions through drainage, the iron sulfides in the soils are oxidized, releasing hydrogen ions (Becher et al. 2019, Boman et al. 2010). Areas affected by heavier drought, because of climate change, might suffer from heavier draining and AS soil impact in the future (Saarinen and Kløve 2012, Fitzpatrick et al. 2017). The soils with insufficient buffer capacity become acidic - creating acid sulfate (AS) soils. As the metals are mobilised in the AS soils, they follow the runoff to rivers, lakes, and marine estuaries (Lax 2005, Erixon 2009, Nordmyr et al. 2008, Nystrand et al. 2013, Nystrand et al. 2016, Job et al. 2018, Ivarsson and Hansson 1995). The extent of oxidation, and subsequently the extent and composition of metals leaching varies with local conditions as runoff and meteorological conditions or sediment

proportions (Nordmyr et al. 2006, Toivonen and Österholm 2011, Dent and Pons 1995, Virtanen et al. 2017).

In Sweden and Finland, these soils are generally found where the land was formerly covered by sea or lake, where prerequisites for anoxic conditions during the sedimentation of the soils could develop (Sohlenius and Öborn 2004). Today the soils are found mainly on landmass, formerly covered by the Littorina Sea, and readily in post-glacial clay and silts (Sohlenius et al. 2015).

Initially studies in Scandinavia have focused on the acidity problematics of AS soils, but during the last decade focus has shifted, with studies showing the effects of AS soils on heavy metal leakages. Many studies exploring the heavy metal leakage of Scandinavian AS soils has had a

“snapshot” resolution, discussing present effects on soils and freshwaters - leaving more to study and discuss about long time, centennial-scale, leakage of AS soils affected by human land-use.

Sweden has a history of early intense draining of both forests and agricultural lands to create larger areas of, and more productive forests and arable lands. Drainage of agriculture lands evolved and intensified during the 19th century. By digging open canals and ditches, as well as placing covered pipes, to lower lakes and subsoils water, more suitable lands were freed.

Drainage of forests and wetlands intensified at the end of the 19th century to culminate during the 1930’s, mainly through digging of open ditches. The perception of drainage of forests and wetlands changed in the 1990’s when awareness rose of the leakage of nutrients the

encroachments caused, as well as the awareness of the importance of wetland ecology (Wesström et al. 2017). Because of the general quality of sulfate soils in Scandinavia, fine grained and often located in the lower coastal landscapes, the soils are often cultivated, and thus oxidized creating the gnarly AS soils (Toivonen and Österholm 2011, Becher et al. 2019).

When comparing metal concentration between clay soil in Sweden containing sulphates to soil lacking them, the concentrations of metals is not higher in the soil containing sulphates, but they hold bigger potential of mobilised metals (Sohlenius and Öborn 2004). This is visible in the increased elemental concentration of certain metals in the runoff waters from the AS soils (Sohlenius and Öborn 2004). Nickel (Ni), Manganese (Mn), Cobolt (Co), Zink (Zn) (Åström and Björklund 1997, Sohlenius and Öborn 2004), and Cadmium (Cd) (Sohlenius and Öborn 2004), are some of the elements that becomes mobilised in AS soil due to oxidation (Sohlenius and Öborn 2004), mainly because of the intensified weathering of the lowered pH-level (Sohlenius and Öborn 2004). Fe (Sohlenius et al. 2009) and Cu (Sohlenius and Öborn 2004) are mobilized to a lower extent, as they partly bind to oxides formed in the chemical processes of AS soil.

This thesis is written as a contribution to the project KolArctic: Geo-Bio Hazards in the Arctic Region, as requested by Geological Survey of Sweden (SGU). The project aims to serve as an exchange of knowledge surrounding environmentally hazardous soils between Sweden, Finland, Norway, and Russia. This thesis studies the effect of such AS soil on the sediments of lake Persöfjärden, and the adjacent marine bay Brändöfjärden, to which the lake drains, on a temporal and spatial perspective.

1.1 Aim

Persöfjärden in north-eastern Sweden is a good example of a lake which may have been highly affected by leakage of metals from its surrounding AS soil (Byrsten and Sandberg 2004, Erixon 2009). This thesis, in contrast to previous studies, focuses on the long-term (centennial

timescale) effects of metal leaching from AS soil affected by human land-use. Here I will address the following research questions:

1. Is the predominant origin of heavy metals (Ni, Zn, Cu and Fe) in the Lake Persöfjärden sediment derived from sulfide weathering or silicate weathering?

2. Is there any indication in the sediment geochemistry record that historical land-uses and their soil drainage activities have increased sulfide weathering and denudation of heavy metals?

3. Is there a spatial gradient in the marine system where sulfide weathering products become stronger to the near-shore source area?

2. Methods

Areas investigated includes lake Persöfjärden and bay Brändöfjärden. Sediments of seven locations were investigated for elemental concentration through X-ray fluorescence analysis (XRF), and for organic content through loss on ignition-analysis (LOI).

2.1 Area of investigation

Persöfjärden (65.77952°N 22.09013°Ö) is a lake in the northern Swedish coastal landscape. The lake was formed by the isostatic rebound, dividing the lake from the sea in about 1750 A.D.

(Byrsten and Sandberg 2004). The lake is a part of the Altersundet catchment, and the Outlet of Persöfjärden sub-catchment. The sub catchment area is 53,43 sq.km, with a land-usage

consisting of approximately 48% forests, 12% agricultural land and 6% mire and wetlands (SMHI). Persöfjärden has two outlets, draining into Furufjärden and Brändöfjärden. The outlet to Brändöfjärden consist of a widened canal and acts as the major outlet from Persöfjärden.

Figure 1. Overview of sample sites in lake Persöfjärden and bay Brändöfjärden. Map data from Lantmäteriet. With markings for sample local 2-8.

The main soils in the immediate surroundings of Persöfjärden consist mainly of clay-silt, till, and glaciofluvial sediments. The soils surrounding the canal outlet to Brändöfjärden consist mainly of clay-silt, while the soils surrounding Brändöfjärden consists mainly of till. According to SGU’s map of Acidic sulfate soils (SGU, kartvisaren, figure 2) the area in immediate connection to Persöfjärden and its outlet Altersundet is mainly active acidic sulfate soil on potentially acidic sulfate soils (estimation, described in Becher et al. 2019).

Figure 2. Estimated and measured AS soils (SGU), where mapped dark blue areas are estimated non-acidic sulfate soils, red areas are estimated active acid sulfate soils upon potentially acid sulfate soils, and green areas are estimated potentially acid sulfate soil. Points of each colour are measured soils.

Persöfjärden was in the 1930’s lowered in an effort to dry up so called water-damaged arable lands round the lake and the two outlets - Altersundet and the Northern outlet (sv: Norra avloppet). In the resolution of the Persöfjärden sjösänkningsföretag av åren 1932, [The water lowering project of Persöfjärden of the years of 1932] (Norrbygdens vattendomstol, B &

U.D.14/1932 ) the work described includes the digging of two channels, and alteration of the existing outlet Altersundet.

The effect on water chemistry of AS soils in Persöfjärden has been investigated by Byrsten and Sandberg (2005) and Erixon (2009). In their thesis, Byrsten and Sandberg (2005) describes how highly increased concentrations of Fe, Al, Mn and Co where found in the lake when investigating water chemistry of the year 2004 (Erixon 2005 in Byrsten and Sandberg 2005) as compared to the water chemistry measured in the year 2000 by the Swedish Agricultural University’s Riksinventeringen 2000, and how when the water chemistry of Persöfjärden was investigated for abnormalities as compared to “unaffected” waters of northern Sweden, Zn and Ni where marked highly abnormal during all the months investigated (June through September, 2004), and Cd, Cr and Cu where marked highly abnormal during parts of the months investigated.

Byrsten and Sandberg (2004) concludes that the dry year of 2004, with following low subsoil water, enabled oxidation of AS soils in the catchment, and were a probable cause to the change in water chemistry.

2.2 Sampling

The samples consist of a deep core, in Persöfjärden, and a transect of surface cores, in

Persöfjärden through adjacent bay. Samples from 7 locals were extracted (table 1); 1 local in the lake and 6 six in the adjacent marine bay (figure 1). The sediment extraction was done both through HTH-corer (Renberg and Hansson 2008), local 2 through 8, and Russian peat corer, location 2. The HTH-core and Russian peat cores at location two were collected with the purpose to create a deep core representing the sedimentation at this, the deeper part of the lake.

Table 1. Description of sampling, where RPC is the Russian peat corer.

2.2.1 Sampling equipment

Two cores were collected at sight 2 using a Russian peat corer. Before sampling the lake, depth was measured by using a sinker with the purpose to find the deep end of the lake, and thus the sedimentation zone. The cores were kept intact and enfolded in plastic foil upon storage in a fridge before handling in lab.

One core was collected at each of the seven locals, one in Persöfjärden and the remaining six in the marine bay, using the HTH-corer. The HTH-core gathered in local 2 were gathered in proximity to the two RPC cores earlier described, in an attempt to create a kind of reference of connected deep core throughout all three depths. Upon extraction of the core, the samples were divided into sections of 1 cm (local 2) or 5 cm (local 3 - 6) by using the HTH standard equipment (as described by Renberg and Hansson 2008). The samples were collected into plastic bags and stored in a fridge before handling in lab.

2.3 Analysis

All laboratory analysis was conducted at Umeå University, EMG laboratories. The analyses consisted of analysis through loss on ignition (LOI) and X-ray fluorescence (XRF) analysis of selected depths of each core. The aim of this thesis was assessed to benefit from a rather high resolution in the making of the geochemical analysis. The selections of depths are further described in following subchapters. The results of the laboratory analyses were used for elemental investigation and for correction of depths of cores.

2.3.1 LOI and XRF analysis

LOI is a calculating organic content in soils. By burning samples in crucibles, in a muffle furnace at 550C, 3 h, the organic matter is expected to burn away; leaving the possibility to calculate the concentration of the lost organic content in the material. All collected depths were analysed for LOI in cores HTH 2 through HTH 8. In the cores Younger RPC and Older RPC, every other centimetre was collected for the LOI analysis. These results were later also used to match the overlay of the RPC-cores and create a corrected depth. From the results of the LOI, a corrected depth was created for the RPC’s, where the measured depth of core Older RPC is corrected and lifted 6 cm.

The XRF measures concentrations of a selection of elements: Na, Mg, Al, Si, P, S, Cl, K, Ca, Ti, V, Mn, Fe, Ni, Cu, Zn, As, Br, Rb, Sr, Zr, Ba, W and Pb. The analysis was made with a Bruker S8- Tiger WD-XRF, as described in Rydberg (2014).

Approximately 200 mg dried and homogenized sample of every sixth centimetre, from depth 0,5 (HTH 2 core) to depth 118 (measured depth, RPC), and every tenth centimetre, from depth 118 to 158 cm (measured depth, RPC) were analysed. Certified reference materials (CRMs)

QC73048, QC70314 and QC73310 were used for the quality control. Because of its a high number of values measured under lower limit of detection, wolfram (W) in location 2 was excluded from further analysis.

2.4 Statistics and data interpretation

To test if the studied heavy metals (Ni, Zn, Cu, Fe) in the lake Persöfjärden sediment originated from sulfide weathering (research question 1), I tested the correlation of the concentrations through Pearson’s correlation analysis. By visually inspect the correlation clusters in a visualized matrix, I assessed the extent of metals clustered with elements of a likely sulfide or silicate origin. Elements grouping with the relatively stable elements Ti and Zr are here assumed to derive from low intensity-weathering (Koinig et al. 2003, Boës et al. 2010), elements such as Ti and Zr is thus used a proxy for low intensity-weathering. Similarly, S was used as a proxy for sulfides, as it is a crucial element in sulfides and has been shown to be fairly immobile after deposition in at least some sediments (Bindler et al 2008). Heavy metals with an insignificant or even negative correlation to Ti does potentially have trends driven by the high intensity

weathering caused by AS soils (Sohlenis and Öborn 2004) and are considered relevant for

further analysis. For research question #2, I used the vertical distribution of the metals judged to have a sulfidic origin, to assess if there were indications of past events of increased metal inputs.

Events are defined as peaks readable in at least two proxies. The significance of spatial gradient in the marine system (research question 3) was assessed using a pearson's regression where the distance from the shoreline was used as the independent variable.

When presenting correlations, R-value above 0,7 is considered a strong correlation, and R-values between 0,3 - 0,5 is considered a weak correlation, treated as a tendency. p-values are presented in tables in the appendix and in the text as: *** p<0,01, **p<0,05, *p<0,1.

3. Results

All XRF results are presented in tables in the appendix (appendix 1 and 2).

3.1 Sediment geochemistry

The vertical distribution of elements (W excluded) in the Lake Persöfjärden is shown in figure 3.

Similarities in trends are found for metals Cu and F, and metals Ni and Zn. The trends of Cu and Fe both show elevated concentrations at 80 and 40cm. The trends of Ni and Zn both show fluctuating elevated concentrations between 40 and 20cm depth. The LOI-trend seems to keep to a generally stable level from 180 cm to about 50cm, where an increase in concentration appears.

Figure 3. Temporal changes of elements and LOI through the corrected depth of local 2, Persöfjärden (excluding W).

When testing for correlation through Spearman rank correlation (figure 4), the heavy metals of interest (Ni, Cu, and Zn) all showed a negative correlation to Ti, the assumed proxy of low intensity weathering. Here Ni showed a stronger, and Zn a weaker, positive correlation to S, while Cu shows no such correlation. Based on the individual correlation scores Ni, Zn, Cu and Fe all correlates to some extent with LOI.

Figure 4. Spearman rank correlation analysis of elemental concentration found from XRF-analysis of the deep core of sediments of Persöfjärden (cores HTH 2, Younger RPC and Older RPC). Each element is correlated to every other and presented by R-value and a representative colour for the R-value.

3.2 Historical events in sulfide weathering and denudation

Four general peaks (readable in at least two proxies) are found in the plotted values; event 1 at 2o cm (Ni, Zn, Ni:LOI, Ni:LOI), event 2 at 40 cm (Cu, Fe), event 3 at approximately 90 cm (Ni:LOI, Zn:LOI, Cu, Fe) and event 4 at approximately 110 cm (Ni:LOI, Zn:LOi) (figure 5). All trends seem more stable in general below 120 cm, indicating an increase in disturbance in the soils above that level. When plotting Ni- and Zn values normalized to LOI (figure 5), the trends become more fluctuating than those of elemental concentration (figure 3), but the observed peak in concentration at 20 cm is still readable in the normalized trends.

Figure 5. Temporal trends of Ni, Zn, Cu, Fe, Ti, LOI, and ratios Ni:LOI, Zn:LOI, Cu:LOI and Fe:LOI, with observed events 1-4 marked out.

Event 1 (figure 5) consists of a decrease in concentrations of Ni and Zn after an increase from 40 cm depth. The decrease in Ni:LOI and Ti:LOI shows a similar sudden decrease at 20 cm, but the decrease is preceded of an earlier increase, starting at approximately 70 cm depth. Event 2 (figure 5) consist of a peak in concentration of Cu and Fe at 40 cm depth. This increase is however not present in the plotted normalized values; Cu:LOI and Fe:LOI, which instead seems slowly decreasing at this depth. Event 3 (figure 5) consists of a peak in Cu and Fe, and lower peaks in Cu:LOI, Fe:LOI, Ni:LOI and Zn:LOI at approximately 90 cm depth. Event 4 (figure 5) consist of an increase in only ratios Ni:LOI and Zn:LOI. The increase starts at 120 cm depth, peaks at 110 cm depth and decreases to a new stable level at 100 cm depth. The Increase is not visible in the values of concentration for Ni and Zn. Event 1, 3, and 4 are interpreted as results of sulfide weathering and denudation because of the consistent peaks through concentrations and normalized values. However, event 2 is rather interpreted to be the result of a higher LOI, and consequently an increase in inbound Fe and Cu.

3.3 Spatial gradient in sulfide weathering product

Fe is excluded from this analysis, since previous studies in the system and similar systems finds Fe to be sedimented already in the freshwater bodies, earlier in the system (Lindström 2017, Erixon 2009). When plotting the uppermost samples to a x-axis representing distance, the concentrations for Ni and Zn seems roughly divided into two means - one for the three locals closer to the shore, and another, lower for the three locals further away from the shore (figure 6).

The concentrations for Cu seem instead more levelled. A sloping trend for Ni and Zn is

confirmed by the R-values of the Pearson’s correlation test (figure 7) for both Ni (-0,76***) and Zn (-0,72***), while Cu shows no significant R-value.

Figure 6: Concentrations of Ti, LOI, Ni, and Zn of the uppermost 10 cm of sediment samples from core HTH 2 through 8. Y-axis consist of distance (m) from outlet.

When making a Pearson’s regression analysis of Ni and Zn to the independent variable Distance (figure 7), the R²-value, being the coefficient of determination, gives a degree of explanation of 58%*** (Ni), respectively 52%*** (Zn), for the impact of distance on the linear trends.

Figure 7. Pearson’s correlation and regression of values from the upper 10 cm of samples in Brändöfjärden, where the metals are the dependent variables (y-axis) and distance from inlet is the independent variable (x-axis).

4. Discussion

4.1 Origin of Ni, Zn, Cu and Fe in Persöfjärden

From the observed low correlation between the studied metals (Ni, Zn, Cu and Fe) and proxies for low intensity-weathering (i.e Ti), I argue that these metals are mainly from sulfide

weathering or atmospheric deposition of metal containing aerosols. In the latter case, metals are expected to have been deposited along with organic matter on the soils surface (Nygård et al.

2012) and subsequently enter freshwater ecosystems with organic solutes (Driscoll et al. 1988).

Since all four studied metals (Ni, Zn, Cu and Fe) correlates with LOI to some extent, surface soils are indeed a possible source of these metals. However, when the vertical distribution of

concentrations of Ni, Zn, Cu and Fe are normalized to LOI (figure 5), peaks in metals are still readable, most visible for Ni and Zn. This is interpreted as an indication of metals partially originating from sulfide weathering rather than atmospheric deposition through organic solutes.

In terms of correlation to lead (Pb) only Fe shows a significant, but weak, correlation. A correlation to Pb could imply a same source origin, likely the smelter in Skellefteå (Rönnskärsverken) located approximately 110km south of the studied catchment.

While the origin of Ni, Zn, and Cu seems fairly straightforward, the origin of Fe seems more uncertain. Besides having an origin from organic surface soils (Driscoll et al 1988), Fe is also known for being mobile after deposition in sediments (Davison et al. 1982). In redox-beneficial environments in sediments, Fe becomes mobile, and a travel up through the sediment column is enabled (Davison et al. 1982). Establishing the actual sedimentation depth of certain irons is hence hard. Fe is also commonly mentioned as one of the lesser mobile metals in AS soils, and so it leaches to a lesser extent. When Toivonen and Österholm (2011) instead finds AS soil-leachate enriched with Fe, they discuss two possible reasons to the finding; the pH in the soil reaches levels below 3,5 where Fe is more soluble, or less oxygen is accessible for exchange in the soil pores, inhibiting immobilization of Fe. Wennström (2017) too discusses increased leakage of Fe, but in coastal Norrbotten waters, and suggest it to be a potential result of local geochemistry, of the soils holding the soluble divalent iron. When Erixon (2009) studied the water chemistry of Persöfjärden, they found that Fe comes in high concentrations during the seasonal variation- periods where sulfide denudation is less palpable in the water chemistry. Erixon notices that the concentration of Fe during these inter episodes of sulfide denudation reaches a much higher level than recorded on the Finnish side. Erixon notes that the water chemistry lacks Fe at the outlet of Persöfjärden and discusses that this would probably be connected to the concurrent lack of organic material, as the two binds complexes and sediments earlier in the waters.

By excluding silicates, I derive that the metals (Ni, Zn, Cu and Fe) rather originates from sulfide- than silicate weathering but because of the correlation of the metals to LOI, I cannot dismiss a potential source in organic solutes too. I base this reasoning on two additional lines of evidences:

i) at least Zn and Ni were positive correlated to sulfur which is heavily enriched in AS soils (Nordmyr et al. 2006); and ii) previous studies have indicated that Ni, Zn, Cu and Fe is exported to a larger extent from AS soils than from non-sulfide soils (Sohlenius and Öborn 2004,

Sohlenius 2015, Lax 2005, Toivonen and Österholm 2011).

4.2 Historical land-use and its effect on element leakage

The sulfide weathering and denudation which is causing fluctuations in the sediment chemistry in Persöfjärden is not constant in a long term (centennial-scale) perspective. Thisbecomes evident from the variations in Ni, Zn, Cu and Fe, which as previously argued have a predominant origin form AS soils, with depth in my sediment cores (figure 5). But is this due to events linked to historical land-use? To be able to link the observed vertical variations in metal concentrations to historical events, the sediment age needs to be inferred using known sedimentation rates in the area. Lindström (2017) studied lakes near Luleå, in a catchment somewhat similar to ours.

They estimated the sedimentation rate, based on the changes in sediment proportions and known year of lake isolation from the sea, to 0,2 cm/year (Lindström 2017). Persöfjärden was isolated from the sea around 1750 (Byrsten and Sandberg 2004). Applying the calculated sedimentation rate of Lindström (2014) on Persöfjärden would give 52cm depth of lake sediment, before reaching the marine sediments beneath. This is of course a very rough estimation, since many local factors affects sedimentation rates in waters, but it does coincide with the measured start of increase in LOI at about 60 cm depth, in the Persöfjärden sediment.

If we were to assume that 60 cm depth represents the isolation of Persöfjärden from the sea in 1750, the sediment depth representing the 1930 lake lowering project would amount to

approximately 20 cm depth. The calculated depth of the lake lowering project then coincides with the earlier presented event 1 (figure 5), with its peak in Ni and Zn concentrations, and peak in normalized values Ni:LOI and Zn:LOI. Because of the visibility of the peaks throughout the normalized values, the peak in concentration of Ni and Zn in event 1, is interpreted to be the result of an increase in sulfide weathering and/or denudation. With the same assumption of sedimentation rate, the beginning of the 19th century and its intensified ditching of forests and arable lands in Sweden, amounts to about 40 cm depth, coinciding with the beginning of the increased concentrations of Ni and Zn, leading up to the concentration peak described as event 1 (figure 5). The Ni and Zn-values normalized to LOI (Ni:LOI, Zn:LOI) are at this depth already at stable high levels since about 90cm depth, which could be caused by the low LOI values of this level, and thus does not necessarily contradict the increase in concentration. Event 2 (figure 5) might also be induced by anthropogenic land use, but the insignificant increase in normalised values of metals (Cu and Fe) to LOI, compared to the rather obvious increase in LOI and

concentrations of Cu and Fe, suggest that this might be a result of additional erosion and particle transport (Trettin et al. 1996 in Lindström 2017) rather than an increased sulfide weathering.

With an assumed lake isolation at 60 cm sediment depth, the proposed events 3 & 4 (figure 5) would represent fluctuations in trends in marine sediments, and events induced before an intensified land-use in the area. I interpret these peaks in concentrations and ratios as effects of non-identified, non-anthropogenic, major disturbances.

4.3 The spatial gradient of Ni and Zn

Both Ni and Zn decreased in concentration with increasing distance from the shoreline i.e. the presumed AS source, in Brändöfjärden (Fig. 9). In short, more than 52 % of the variation in these trends in concentration is explained by the distance from the inlet of the marine bay,

Brändöfjärden. Indeed, my results indicate that metal transport from the sulfide soils have been substantial enough to be visible in the marine sediments and makes a good complement to the relatively few previous studies I have found discussing the subject (e.g Nordmyr et al. 2008).

However, two weaknesses are easily identified; i) the origin of Ni and Zn are only partly presumed to originate from AS soils; and ii) the data set was limited, leaving room for a discussion about representability. Further studies would benefit from a bigger data set, or combined investigation areas, for a more representative representation.

5. Conclusions

I found that: i) Nickel (Ni), Zink (Zn), Copper (Cu) and Iron (Fe) in the sediments had a partial likely origin from AS soils; ii) vertical variations in Ni, Zn, Cu and Fe concentrations suggested events of increased metal transport in the past; and iii) there is a tendency that Zn and Ni assumed from AS soils have caused elevated metal concentration in marine sediment deposited in Persöfjärden. Inferred sediment rates suggest that elevated metal concentrations occurred in sediment deposited during the 18-19^th and 20^th century. I argue that these periods correspond to periods of increased drainage of the AS soils during past agricultural activities in the catchment.

6. Acknowledgments

I want to thank KolArcitc for funding this project. I also want to thank supervisors Gustav Sohlenius, Geological Survey of Sweden, and Jonatan Klaminder, Umeå University, for commitment, guidance, and advice during the course of work. Lastly, I want to thank June Johansson and Jimmy Markström for sound and kind support, and Jimmy again for fun and rewarding weeks in field and laboratory.

7. References

Becher, M, Sohlenius, G and Öhrling, C. 2019. Sur Sulfatjord - egenskaper och utbredning.

Rapport/SGU: 2019:13. Uppsala: SGU.

Bindler, R, Wik-Persson, M and Renberg, I. 2008. Landscape-scale patterns of sediment sulfur accumulation in Swedish lakes. Journal of Paleolimnology 31:61-70. doi: 10.1007/s10933-007- 9095-0

Boës X, Rydberg, J, Martinez-Cortizas, A, Bindler, R, Renberg I. 2010. Evaluation of

conservative lithogenic elements (Ti, Zr, Al and Rb) to study anthropogenic element enrichments in lake sediments. Journal of Paleolimnology 46:75-87. doi: 10.1007/s10933-011-9515-z

Boman, A, Fröjdö, S, Backlund K and Åström M. 2010. Impact of isostatic land uplift and artificial drainage on oxidation of brackish-water sediments rich in metastable iron sulfide.

Geochimica et Cosmochimica Acta 74(4):1268-1281. doi:10.1016/j.gca.2009.11.026 Byrsten, S., and Sandberg, A. 2005. Metaller och fisk i Persöfjärden : en undersökning av metallhalter i fisk och ytvatten. BSc-Diss, Luleå University of Technology

http://urn.kb.se/resolve?urn=urn:nbn:se:ltu:diva-48175

Davison, W, Woof, C and Rigg, E. 1982. The dynamics of iron and manganese in a seasonally anoxic lake; direct measurement of fluxes using sediment traps. Limnology and Oceanography 27(6):987-1003. doi:10.4319/lo.1982.27.6.0987

Dent, D and Pons, L. 1995. A world perspective on acid sulphate soils. Geoderma 67:263-276.

doi:10.1016/0016-7061(95)00013-E

Driscoll, CT, Fuller, RD and Simone, DM. 1988. Longitudinal Variations in Trace Metal Concentrations in a Northern Forested Ecosystem. Journal of Environmental Quality 17(1):101- 107. doi: 10.2134/jeq1988.00472425001700010015x

Erixon, P. 2009. Klimatstyrda sulfidoxidationer som orsak till surhet och höga metallhalter i vattendrag i norra Sverige (Science report / Luleå University of Technology). Luleå. Retrieved from http://urn.kb.se/resolve?urn=urn:nbn:se:ltu:diva-22383

Fitzpatrick, R.W., Mosley, L.M., Raven, M.D., Shand, P. 2017. Schwertmannite formation and properties in acidic drain environments following exposure and oxidation of acid sulfate soils in irrigation areas during extreme drought. Geoderma 308:235-251.

doi:10.1016/j.geoderma.2017.08.012

Ivarsson H and Jansson M. 1995. Sources of acidity in running waters in central northern Sweden. Water, Air, and Soil Pollution 84:233-251. doi:10.1007/BF00475342

Job, T, Penny, D and Hua, Q. 2018. Metal enrichment in estuarine sediments proximal to acid sulfate soils as a novel palaeodrought proxy. Science of The Total Environment. 612:247-256.

doi:10.1016/j.scitotenv.2017.08.157

Koinig, KA, Shotyk, W, Lotter, AF, Ohlendorf, C and Sturm, M. 2003. 9000 years of geochemical evolution of lithogenic major and trace elements in the sediment of an alpine lake–the role of climate, vegetation, and land-use history. Journal of Paleolimnology 30:307–320. doi:

10.1023/A:1026080712312

Lax, K. 2005. Stream plant chemistry as indicator of acid sulphate soils in Sweden. Agricultural and Food Science 14(1):83-97. doi:10.2137/1459606054224165

Lindström C. 2017. Acid sulfate soils and metal accumulation in sediments in the Rosån catchment, northern Sweden. Ma-Diss. Uppsala University

http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-325213

Nordmyr, L, Boman, A, Åström M and Österholm P. 2006. Estimation of leakage of chemical elements from boreal acid sulphate soils. Boreal environment research 11:261-273

Nordmyr, L, Åström M and Peltola P. 2008 Metal pollution of estuarine sediments caused by leaching of acid sulphate soils. Estuarine, Coastal and Shelf Science 76(1):141-152.

doi:10.1016/j.ecss.2007.07.002

Norrbygdens vattendomstol, B. & U.D14/1932. Persöfjärden sjösänkningsföretag av åren 1932.

Decided 1932-12-12.

Nygård, T, Steinnes, E and Royset, O. 2012. Distribution of 32 Elements in Organic Surface Soils: Contributions from Atmospheric Transport of Pollutants and Natural Sources. Water, Air, and Soil Pollution 223:699-713. doi:10.1007/s11270-011-0895-5

Nystrand, M and Österholm, P. 2013. Metal species in a Boreal river system affected by acid sulfate soils. Applied Geochemistry 31:133-141. doi:10.1016/j.apgeochem.2012.12.015

Nystrand, M, Österholm, P, Changxun, Y and Åström, M. 2016. Distribution and speciation of metals, phosphorus, sulfate and organic material in brackish estuary water affected by acid sulfate soils. Applied Geochemistry 66:264-274. doi:10.1016/j.apgeochem.2016.01.003

Proske, U, Heijnis, H and Gadd P. 2014. Using X-ray fluorescence core scanning to assess acid sulfate soils. Soil Research 52(8):760-768. doi:10.1071/SR14103

Renberg, I and Hansson, H. 2008. The HTH sediment corer. Journal of Paleolimnology.

40:655–659.

Saarinen, T and Kløve, B. 2012. Past and future seasonal variation in pH and metal

concentrations in runoff from river basins on acid sulphate soils in Western Finland. Journal of Environmental Science and Health 47:1614-1625. doi:10.1080/10934529.2012.680783

SGU, Geological Survey of Sweden. 2020. Kartvisaren sura sulfatjordar

https://apps.sgu.se/kartvisare/kartvisare-sur-sulfatjord.html (Retrived 2020-03-31) Sohlenius, G, Aroka, N, Wåhlén, H, Uhlbäck, J. and Persson, L. 2015. Sulfidjordar och sura sulfatjordar i Västerbotten och Norrbotten. SGU-rapport 2015:26. Sveriges geologiska undersökning (SGU)

Sohlenius G, Bastani, M, Persson, L, Lax, K. 2009. On the recognition of areas with problematic sulphidic sediments using multi-disciplinary data. Environmental Geology 56:973-984.

doi:10.1007/s00254-008-1199-y

Sohlenius, G. & Öborn, I., 2004. Geochemistry and partitioning of trace elements in acid sulphate soils in Sweden and Finland before and after sulphide oxidation. Geoderma 122(2):167–175. doi:10.1016/j.geoderma.2004.01.006

SMHI. 2016. Vattenwebben. http://vattenweb.smhi.se/modelarea/ (Retrieved 2020-03-31)

Sukitprapanon, T, Suddhiprakarn, A, Kheoruenromne, and Gilkes R.J. 2018. Partitioning and potential mobilization of aluminum, arsenic, iron, and heavy metals in tropical active and post- active acid sulfate soils: Influence of long-term paddy rice cultivation. Chemosphere 197:691- 702. doi:10.1016/j.chemosphere.2018.01.099

Toivonen, J and Österholm, P 2011. Characterization of acid sulfate soils and assessing their impact on humic boreal lake. Journal of Geochemical Exploration 110(2):107-117. doi:

10.1016/j.gexplo.2011.04.003

Virtanen, S., Puustinen, M., Yli-Halla, M.. 2017. Oxidation of iron sulfides in subsoils of cultivated boreal acid sulfate soil fields – based on soil redox potential and pH measurements.

Geoderma 308:252-259. doi:10.1016/j.geoderma.2017.05.020

Wesström, I., Hargeby, A. and Tonderski, K. 2017. Miljökonsekvenser av markavvattning och dikesrensning - En kunskapssammanställning. Naturvårdsverket. Rapport 6777.

Åström, M and Björklund, A. 1997. Geochemistry and acidity of sulphide-bearing postglacial sediments of western Finland. Environmental Geochemistry and Health 19:155-164.

doi:10.1023/A:1018462824486

Appendix 1.

Appendix 1. XRF results for elements in sediments of Persöfjärden (local 2). Values in red undergoes LLD as described in Rydberg (2014).

Appendix 2

Appendix 2. XRF results for elements in sediments of local 3,4,5,6,7 and 8.

Appendix 3

Appendix 3. p-values of spearman rank correlation for elements and LOI in sediments of Persöfjärden (local 2).