Infrared spectroscopy as a tool to reconstruct past lake-ecosystem changes Method development and application in lake-sediment studies

Carsten Meyer-Jacob

Department of Ecology and Environmental Science Umeå 2015

This work is protected by the Swedish Copyright Legislation (Act 1960:729) © Carsten Meyer-Jacob 2015 ISBN: 978-91-7601-372-4 Cover: Torneträsk, photo by Carsten Meyer-Jacob Electronic version available at http://umu.diva-portal.org/ Printed by: Servicecenter KBC Umeå, Sweden 2015

Everything is flowing—going somewhere, animals and so-called lifeless rocks as well as water. Thus the snow flows fast or slow in grand beauty-making glaciers and avalanches; the air in majestic floods carrying minerals, plant leaves, seeds, spores, with streams of music and fragrance; water streams carrying rocks… While the stars go streaming through space pulsed on and on forever like blood globules in Nature's warm heart.

- John Muir

i

Table of Contents

Table of Contents i Abbreviations ii List of papers iii Author contributions iv Abstract v Introduction 1

Quantification of sediment components using IR spectroscopy 2 Indirect reconstruction of environmental variables using IR spectroscopy 3

Objectives 4 Methods 7

Infrared spectroscopy 7 Spectral analysis and model development 7 Independent measurement of biogenic silica 7 Lake-water TOC reconstruction 8 Other analytical methods 8

Major results and discussion 10 Advances in the development of IR spectroscopic methods in paleolimnology 10 IR spectroscopic methods applied in paleolimnological studies 11 3.6 million years of lake-ecosystem response to climate variability in the Siberian Arctic 11 Sensitivity of a large-lake system to Holocene climate changes in subarctic Sweden 12 Long-term perspective on recent changes in lake-water TOC 15

Conclusions 17 Future prospects 18 References 19 Acknowledgments 25

ii

Abbreviations ARbSi Accumulation rate of biogenic silica

ARTOC Accumulation rate of total organic carbon

bSi Biogenic silica

cal yr BP Calendar year before present (i.e., A.D. 1950)

FTIR spectroscopy Fourier transform infrared spectroscopy

IR Infrared

Ma/ka Million years/kilo years

OM Organic matter

PLS regression Partial least squares regression

R2cv Cross-validated coefficient of determination

RMSECV Root mean square error of cross-validation

RMSEP Root mean square error of prediction

TIC Total inorganic carbon

TOC Total organic carbon

VNIR spectroscopy Visible-near infrared spectroscopy

iii

List of papers

This doctoral thesis is a summary of the following papers, which are referred to in the text by their Roman numerals:

I. Meyer-Jacob, C., Vogel, H., Gebhardt, A. C., Wennrich, V., Melles, M., and Rosén, P.

2014. Biogeochemical variability during the past 3.6 million years recorded by FTIR spectroscopy in the sediment record of Lake El'gygytgyn, Far East Russian Arctic. Climate of the Past, 10(1):209-220.

II. Meyer-Jacob, C., Vogel, H., Boxberg, F., Rosén, P., Weber, M. E., and Bindler, R. 2014.

Independent measurement of biogenic silica in sediments by FTIR spectroscopy and PLS regression. Journal of Paleolimnology, 52(3):245-255.

III. Meyer-Jacob, C., Bindler, R., Bigler, C., Leng, M. J., and Vogel, H. Holocene ecosystem

ontogeny and element cycling in the subarctic catchment of lake Torneträsk, NW Sweden – Large lake sensitivity to climate change. Manuscript.

IV. Meyer-Jacob, C., Tolu, J., Bigler, C., Yang, H., and Bindler, R. 2015. Early land use and

centennial scale changes in lake-water organic carbon prior to contemporary monitoring. Proceedings of the National Academy of Sciences of the United States of America, 112(21):6579-6584.

Paper II is reproduced with kind permission from Springer Science and Business Media.

iv

Author contributions

Authors are referred to by their initials.

Authors: CMJ: Carsten Meyer-Jacob, HV: Hendrik Vogel, ACG: A. Catalina Gebhardt, VW: Volker Wennrich, MM: Martin Melles, PR: Peter Rosén, FB: Florian Boxberg, MEW: Michael E. Weber, RB: Richard Bindler, CB: Christian Bigler, MJL: Melanie J. Leng, JT: Julie Tolu, HY: Handong Yang Paper I CMJ, MM, and PR conceived the study; CMJ performed analyses of calibration set samples; CMJ performed data analysis; ACG and VW contributed to the calculation of accumulation rates, CMJ wrote the manuscript; HV, MM, and PR commented on the manuscript. Paper II CMJ, HV, and PR conceived the study; FB prepared samples and performed analyses with contributions from CMJ; MEW contributed data; CMJ performed data analysis; CMJ wrote the manuscript with contributions from HV and RB; MEW commented on the manuscript. Paper III CMJ and HV conceived the study; CMJ performed infrared spectroscopic and geochemical analyses; CB performed diatom analyses; MJL performed diatom isotope analyses; CMJ performed data analysis with input from RB and HV; CMJ wrote the manuscript with contributions from RB and HV; CB and MJL commented on the manuscript. Paper IV CMJ and RB conceived the study; CMJ performed infrared spectroscopic and geochemical analyses; JT performed Py-GC/MS analyses; CB performed diatom analyses; HY performed 210Pb-dating; CMJ performed data analyses with input from RB; CMJ wrote the manuscript with input from RB; JT, CB, and HY commented on the manuscript.

v

Abstract

Natural archives such as lake sediments allow us to assess contemporary ecosystem responses to climate and environmental changes in a long-term context beyond the few decades to at most few centuries covered by monitoring or historical data. To achieve a comprehensive view of the changes preserved in sediment records, multi-proxy studies – ideally in high resolution – are necessary. However, this combination of including a range of analyses and high resolution constrains the amount of material available for analyses and increases the analytical costs. Infrared spectroscopic methods are a cost-efficient alternative to conventional methods because they offer a) a simple sample pre-treatment, b) a rapid measurement time, c) the non- or minimal consumption of sample material, and d) the potential to extract quantitative and qualitative information about organic and inorganic sediment components from a single measurement.

The main objective of this doctoral thesis was twofold. The first part was to further explore the potential of Fourier transform infrared (FTIR) and visible-near infrared (VNIR) spectroscopy in paleolimnological studies as a) an alternative tool to conventional methods for quantifying biogenic silica (bSi) – a common proxy of paleoproductivity in lakes – in sediments and b) as a tool to infer past lake-water total organic carbon (TOC) levels from sediments. In a methodological study, I developed an independent application of FTIR spectroscopy and PLS modeling for determining bSi in sediments by using synthetic sediment mixtures with known bSi content. In contrast to previous models, this model is independent from conventional wet-chemical techniques, which had thus far been used as the calibration reference, and their inherent measurement uncertainties. The second part of the research was to apply these techniques as part of three multi-proxy studies aiming to a) improve our understanding of long-term element cycling in boreal and arctic landscapes in response to climatic and environmental changes, and b) to assess ongoing changes, particularly in lake-water TOC, on a centennial to millennial time scale.

In the first applied study, high-resolution FTIR measurements of the 318-m long sediment record of Lake El’gygytgyn provided a detailed insight into long-term climate variability in the Siberian Arctic over the past 3.6 million years. Highest bSi accumulation occurred during the warm middle Pliocene (3.6-3.3 Ma), followed by a gradual but variable decline, which reflects the first onset of glacial periods and then the finally full establishment of glacial–interglacial cycles during the Quaternary. The second applied study investigated the sediment record of Torneträsk in subarctic northern Sweden also in relation to climate change, but only over the recent post-glacial period (~10 ka). By comparing responses to past climatic and environmental forcings that were recorded in this large-lake system with those recorded in small lakes from its catchment, I determined the significance and magnitude of larger-scale changes across the study region. Three different types of response were identified over the Holocene: i) a gradual response to the early landscape development following deglaciation (~10000-5300 cal yr BP); ii) an abrupt but delayed response following climate cooling during the late Holocene, which occurred c. 1300 cal yr BP – about 1000-2000 years later than in smaller lakes from the area; and iii) an immediate response to the ongoing climate change during the past century. The rapid, recent response in a previously rather insensitive lake-ecosystem emphasizes the unprecedented scale of ongoing climate change in northern Fennoscandia. In the third applied study, VNIR-inferred lake-water TOC concentrations from lakes across central Sweden showed that the ongoing, observed increase in surface water TOC in this region was in fact preceded by a long-term decline beginning already AD 1450-1600. These dynamics coincided with early human land use activities in the form of widespread summer forest grazing and farming that ceased over the past century. The results of this study show the strong impact of past human activities on past as well as ongoing TOC levels in surface waters, which has thus far been underestimated. The research in this thesis demonstrates that infrared spectroscopic methods can be an essential component in high-resolution, multi-proxy studies of past environmental and climate changes.

Keywords: Fourier transform infrared spectroscopy, visible-near infrared spectroscopy, PLS regression, biogenic silica, climate change, carbon cycling, lake-water quality, geochemistry, paleolimnology, Holocene, Lake El’gygytgyn, Torneträsk

1

Introduction

Knowledge of past baseline conditions and their natural variability is key to understanding concurrent climatic and environmental changes and their implications for ecosystem management (Smol, 1992). To determine the risks of ongoing and projected climate change, it is essential to understand how terrestrial and aquatic ecosystems have evolved and responded to past climate change as well as the superimposed human impacts over tens to possibly thousands of years. However, instrumental records cover only a few decades to at most a few centuries and are thus too short to capture the scale of natural variability and to differentiate between human-induced and natural environmental changes. In addition, monitoring data are especially sparse for Arctic and subarctic landscapes that are currently undergoing the most pronounced changes due to ongoing climate warming (IPCC, 2007). Natural archives, such as ice cores, tree-rings, speleothems, peat records, or marine and lacustrine sediments, preserve information about past environmental changes and allow us to reconstruct and study past climatic and environmental dynamics beyond the perspective of instrumental data. Lake-sediment studies have contributed to improving our understanding of concurrent ecosystem responses to environmental problems, such as surface-water acidification (e.g., Battarbee, 1984; Renberg et al., 1993), eutrophication (e.g., Anderson et al., 1993; Bennion et al., 2004), soil erosion (e.g., Dearing and Jones, 2003, and references therein), pollution (e.g., Bindler et al., 2009; Boyle et al., 1998; Brännvall et al., 1999), and climate change (e.g., Smol and Douglas, 2007; Smol et al., 2005).

To obtain a comprehensive view of the recorded changes in sediment archives, multi-proxy studies including biological and geochemical proxies are essential (Engstrom and Wright, 1984), and to capture the long- and short-term variability, these studies are ideally performed at a high resolution. However, these two requirements – multi-proxy and high resolution – place strong demands on the limited amount of material that is available for a given sediment level for all of the desired analyses. Consequently, there is great value in high-quality analytical techniques that are either non-destructive or require small sample masses and which are rich in information. Examples of non-destructive techniques include x-ray fluorescence spectroscopy (XRF) based on either core scanning (Zolitschka et al., 2001) or quantitative analyses of discrete samples (Boyle, 2000; Rydberg, 2014) and visible- and near-infrared spectroscopy (VNIR). Examples of techniques requiring small masses include combined analyses of carbon, nitrogen and their isotopes and Fourier transform infrared spectroscopy (FTIR, mid-infrared).

Infrared (IR) spectroscopic methods such as FTIR and VNIR spectroscopy facilitate the characterization of molecular structures, and are widely used analytical techniques for process monitoring and product quality control in different industries such as agriculture, food, textiles, petrochemicals, and pharmaceuticals. In addition to applications in industrial processes, IR spectroscopy is a common analytical tool in chemistry, biology, and medicine to determine, for example, the composition of biological tissues (e.g., Foley et al., 1998; Movasaghi et al., 2008), proteins (e.g., Barth, 2007) or minerals (e.g., Farmer, 1974). However, applications of IR spectroscopy in paleoenvironmental and paleoclimatological studies are still scarce despite the advantages that these techniques offer, i.e., i) a simple sample pre-treatment, ii) a fast (< 1 min/sample) and straightforward measurement, iii) the non-consumption of sample material in the case of VNIR or the only small amount of material required for analysis (0.01 g) in the case of FTIR spectroscopy, and iv) the potential to extract a multitude of information from single measurements. Because of these characteristics, IR spectroscopy is particularly suitable for multi-proxy and high-resolution studies.

The application of IR spectroscopy to sediment samples is complicated by the complex nature of this sample matrix, which consists of a multitude of organic and inorganic compounds and whose IR spectral signatures may superpose each other (Fig. 1). Therefore, IR spectroscopy, when applied as an empirically based quantitative tool, depends on large sets of calibration samples and complex chemometrics (e.g., McClure, 1994; Siesler et al. 2002). However, the number of publications as well as the number of different applications using IR spectroscopy in paleoenvironmental studies has slowly increased since the 1990’s. For example, IR spectroscopy has been used to characterize organic substances in studies of environmental archives such as marine and lacustrine sediments (e.g., Belzile

2

et al., 1997; Braguglia et al., 1995; Calace et al., 1999; Calace et al., 2006; Mecozzi and Pietrantonio, 2006) as well as in peat (e.g., Chapman et al., 2001; Cocozza et al., 2003). The technique has been used to qualitatively assess the spatial variability in surface sediments (Korsman et al., 1999; Rydberg et al., 2012) but also more frequently to directly quantify specific sediment components or to indirectly infer environmental variables such as lake-water pH, lake-water total organic carbon (TOC) (discussed later in the thesis), and air temperature.

Figure 1. FTIR spectra of a sediment sample with ~10% biogenic silica (bSi) compared to the spectra of pure bSi with its characteristic IR absorption peaks around 1100, 945, 800, and 470 cm-1. FTIR spectra are stacked for better visualization.

Quantification of sediment components using IR spectroscopy

A number of studies quantified specific sediment components by establishing calibration functions between the IR spectral information of a set of samples and the corresponding conventionally measured concentrations of a component of interest. This function can subsequently be applied to other sediment samples to quantify the sediment component of interest exclusively based on the IR spectrum of the sample. One of the first quantifications had been demonstrated by Chester and Elderfield (1968) who estimated the biogenic silica (bSi) content in marine sediments by means of IR spectroscopy. Other studies successfully quantified different silicate minerals (e.g., Bertaux et al., 1998; Bertaux et al., 1996; Sifeddine et al., 1994; Wirrmann et al., 2001) and carbonates (e.g., Mecozzi et al., 2001) using mid-IR spectroscopy, and total concentrations of, e.g., organic and inorganic carbon, nitrogen, or phosphorus (e.g., Balsam and Deaton, 1996; Chang et al., 2005; Malley et al., 2000; Malley et al., 1999) and chlorophyll a and its derivatives (e.g., Das, 2007; Das et al., 2005; Michelutti, 2005; Michelutti et al., 2010; Rein and Sirocko, 2002; Wolfe et al., 2006) using VNIR spectroscopy.

More recently, the use of FTIR spectroscopy in conjunction with partial least square (PLS) regression has been shown to provide accurate measurements of biogeochemical sediment components such as TOC, total inorganic carbon (TIC), and bSi in sediments (Hahn et al., 2011; Rosén et al., 2010; Vogel et al., 2008). To date, this approach has been used in several studies to determine the past variability of biogeochemical proxies in lacustrine (e.g., Brigham-Grette et al., 2013; Brisset et al., 2013; Cunningham et al., 2013a; Cunningham et al., 2013b; Hahn et al., 2013; Vogel et al., 2013; Vogel et al., 2008; Vogel et al., 2010) as well as marine studies (Sprenk et al., 2013). PLS regression is particularly suitable for the analysis of IR spectra because it can be used to analyze data sets with strongly collinear, noisy, and numerous X-variables. This regression method models the relationship between two matrices, which in this case are the IR spectral information of a sediment sample and the corresponding concentration of the proxy of interest in the analyzed sample (Wold et al., 2001). The first applications of FTIRS and PLS regression were developed for specific sediment records using internal downcore calibration models built on a subset of samples from the actual study site (Vogel et al., 2008). As a next step, Rosén et al. (2011b) introduced the potential for a general calibration model that relies on globally distributed surface and downcore sediment samples and which demonstrated the possibility of a site-independent application of FTIR spectroscopy and PLS regression to quantify

3

sediment components as an alternative to conventional methods. This approach still relies, however, on calibrating the original samples to be used in the model against conventional methods.

Indirect reconstruction of environmental variables using IR spectroscopy

A common approach in paleoenvironmental studies is the use of indirect calibrations, so-called transfer functions, to infer environmental variables that are not directly measurable in the sediment. This approach has widely been used in the interpretation of changes in diatom communities, which are, for example, particularly sensitive to changes in lake-water pH (Battarbee et al., 2001). Conceptually similar transfer functions have also been established using IR spectral information to infer environmental conditions such as the lake-water chemistry or climatic conditions that prevailed during the sediment formation. These indirect calibrations are based on IR spectra of surface sediment samples from a set of lakes that covers a large environmental gradient and the corresponding variable of interest. The transfer function resulting from the calibration between these two datasets can then be applied to the IR spectrum of another sediment sample to reconstruct the proxy of interest only by means of its IR signature. There are three basic assumptions to this approach using IR spectroscopy: i) organic and inorganic compounds forming the sediment are strongly affected by and therefore reflect the environmental conditions present in a lake and its catchment during sediment formation, transport, and accumulation; ii) IR measurements are sensitive to these variations in the sediment composition; and iii) diagenetic processes do not significantly alter the IR spectral information, which allow the reconstruction of the proxy of interest. The first such indirect calibrations using VNIR spectroscopy were established by Korsman et al. (1992) and Nilsson et al. (1996), who developed PLS regression models to infer lake-water pH from the IR spectral signature of sediment samples. Other studies have since used, for example, temperature-related variations in the sediment composition to establish transfer functions between IR spectral information and air temperature (Rosén et al., 2000; Trachsel et al., 2010).

The first calibration model for lake-water TOC had been developed by Nilsson et al. (1996), which indicated the potential of the technique to reconstruct past TOC levels by VNIR spectroscopy from sediment archives despite a relatively small calibration set size of 25 samples. This approach was followed up by Rosén (2005) using surface samples from 100 lakes in northernmost Sweden and further extended by Cunningham et al. (2011), who extended the geographic scope of the calibration set by including an additional 40 samples from lakes in central and southern Sweden. These latter lakes are part of the Swedish national trend lake monitoring program and most have water-chemistry data available since the mid-1980’s (data host: Institutionen för vatten och miljö, SLU). Applications of the latter calibration model showed a good agreement between VNIR inferred and conventionally measured monitoring data for four lakes that are included in the Swedish national environmental monitoring program since 1987 (Cunningham et al., 2011).

The same methodology used for the Swedish lake-water TOC model was used to develop a comparable model for lake-water dissolved organic carbon (DOC) based on samples from Arctic Canada (Rouillard et al., 2011). This study demonstrated that the approach to use VNIR spectroscopy to infer past lake-water OC concentrations is not restricted to Scandinavian lakes and can also be used in regions with different environmental settings. To date, these VNIR-inferred models for lake-water TOC/DOC have been applied with promising results in other studies aiming at, for example, reconstructing past mire, permafrost, and treeline dynamics (Jones et al., 2011; Kokfelt et al., 2009; Rosén, 2005; Rydberg et al., 2010); assessing factors driving changes in lake-water chemistry (Bragee et al., 2015; Rosén et al., 2011a; Rosén and Hammarlund, 2007); and tracking ecosystem responses to climate variations (Reuss et al., 2010; Rosén et al., 2009; Rouillard et al., 2012). VNIR-inferred lake-water TOC concentrations were further used to improve the modeling and assessment of reference conditions for water management (Valinia et al., 2014).

4

Objectives

The main objectives of the research conducted in this doctoral thesis are twofold. First, the evaluation and further development of current applications of IR spectroscopy to a) quantify biogeochemical sediment components, with a focus on bSi (Papers I, II, III), and b) to infer past lake-water TOC levels (Paper IV). Second, the application of these methods in multi-proxy studies aiming at a) improving our understanding of natural long-term environmental variations in response to climate change in northern landscapes, and b) assessing recent environmental changes in a longer-term context (Papers I, III, IV).

Applications of FTIR spectroscopy and PLS regression to quantify biogeochemical sediment components, such as TOC, TIC, bSi, have thus far been applied to several sediment records that are up to 340 ka old and representing different environmental settings from the Mediterranean to the Arctic. However, it remains uncertain whether a single robust FTIR calibration model can be developed for sediment sequences that formed under substantially different climatic and environmental conditions beyond a few glacial-interglacial cycles, and that are potentially affected by more significant diagenetic changes (Paper I). The sediment sequence from Lake El’gygytgyn, Far East Russian Arctic, recorded information about the climatic and environmental evolution of the Arctic during the past 3.6 Ma – from the warm middle Pliocene over the onset of glacial-interglacial cycles until present (Brigham-Grette et al., 2013; Melles et al., 2012). The main aim of Paper I was to develop and apply FTIR-calibration models for the three sediment components in a very long sediment record (>300 m, 3.6 Ma), and thereafter to quantitatively examine changes in lake productivity inferred from bSi over this timeframe in light of established climatic changes. IR measurements of purified bSi from this record have shown that diagenetic processes altered the structure of bSi in sediments that formed during the past 250 ka (Chapligin et al., 2012). During this time period, Si-OH groups of bSi declined by 26% compared to Si-O-Si groups as consequence of a condensation reaction, following the continuous burial and maturation in the sediment. Such diagenetic changes may potentially complicate the accurate quantification of biogeochemical sediment components.

One limitation of current FTIR-PLS regression models for TOC, TIC, and bSi is that these models rely on the determination of calibration values based on conventional techniques, and thus inherently depend on the accuracy and uncertainties of these techniques. For example, bSi is conventionally measured by wet-chemical digestion techniques, which use an alkaline solution such as NaOH or Na2CO3 to sequentially extract the Si. Conley (1998) showed in an inter-laboratory comparison that results from these wet-chemical digestion techniques can vary significantly between laboratories. Although the precision of individual participating laboratories was quite high, the uncertainty for the bSi measurements between laboratories was 20-35% for samples with higher bSi (~40%) and 30-68% for samples with low bSi (<10%). These uncertainties between laboratories can arise from, e.g. variable execution, such as the concentration of the digestion solution or the slope determination, contaminated water systems, and lack of an absolute accuracy because there are no BSi standards. Furthermore, conventional leaching methods potentially overestimate bSi concentrations, especially in clay mineral rich samples, because of dissolution of minerogenic silica from alumino-silicates during the leaching process (Clymans et al., 2015; Kamatani and Oku, 2000; Ohlendorf and Sturm, 2008). In addition to the uncertainty introduced by basing a calibration approach on another method with its uncertainties, site-specific internal FTIR calibrations require IR spectroscopic and at least some conventional measurements of the proxy of interest. While this takes advantage of the faster analysis time of FTIR spectroscopy, it still requires duplicate work in performing a wet-chemical technique on a subset of samples (e.g. 420 samples in the Lake El’gygytgyn study). One specific objective was thus to develop a stand-alone FTIR calibration model for quantifying bSi in sediments that is independent of conventional measurements and their underlying measurement uncertainties (Paper II). This model would further have the potential to be universally applicable and therefore remove the need to develop time-consuming and expensive internal calibrations for each individual study site.

In Paper III, FTIR-inferred bSi concentrations were used as part of a multi-proxy study of the large lake Torneträsk, which is situated in northern subarctic Sweden. Ongoing climate change strongly affects this area and has thus far led to a mean annual temperature increase of 2.5°C (Callaghan et al., 2013). During the past decades, the Torneträsk catchment has been the focus of numerous studies

5

aiming to develop an improved process understanding in aquatic and terrestrial ecosystems (e.g., (Karlsson et al., 2010; Åkerman and Johansson, 2008), as well as the focus of several paleoenvironmental studies of small lakes that aimed to study the Holocene environmental and climatic development in the area (e.g., Barnekow, 1999; Bigler et al., 2002). Because of the multitude of studies conducted in the Torneträsk catchment and the relatively low human impact on the area, the lake represents an ideal site for investigating the regional environmental response to climate change in northernmost Sweden since deglaciation. In addition, these unique prerequisites allow comparing the sediment archive of a large-lake system, which integrates changes across a region, with the signals recorded in sediments of small lakes from its catchment, which are likely more sensitive to local catchment characteristics. Such a comparison between the sensitivity of large versus small lakes facilitates assessing the spatial and temporal variations in the responses to climate forcing across a larger region, which is key to understanding the magnitude and timing of concurrent ecosystem responses to climate change.

In Paper IV, VNIR-inferred lake-water TOC reconstructions were used to put recent changes in lake-water TOC into a longer-term perspective. Monitoring programs have recorded increasing TOC concentrations in lakes and rivers across parts of Europe and N-America (e.g., Driscoll et al., 2003; Monteith et al., 2007; Skjelkvale et al., 2005; Worrall et al., 2004) during the past few decades but the main drivers causing this phenomenon are still controversial. The few decades of overlap that now exist between monitoring and sediment records allow us to make quantitative comparisons between the recent trends measured in monitoring programs and the long-term patterns preserved in sediments. Several hypotheses have been proposed to explain the increase such as climate change (e.g., Ekström et al., 2011; Fenner and Freeman, 2011; Freeman et al., 2001; Hejzlar et al., 2003; Hongve et al., 2004; Tranvik and Jansson, 2002), declining acid deposition (e.g., De Wit et al., 2007; Evans et al., 2012; Evans et al., 2005; Monteith et al., 2007) or alterations of land use and land cover (e.g., Armstrong et al., 2010; Tuvendal and Elmqvist, 2011; Yallop et al., 2010).

All proposed hypotheses are based on the recent TOC trends observed in monitoring programs, which reflect the development in surface waters only over the past few decades. However, the lack of long-term monitoring data – over centuries and millennia – leaves us with an ambiguous understanding of the past trajectory of TOC concentrations. For example, many of the climate-related hypotheses imply some change in ecosystem and soil functioning, which in turn lead to new and possibly unprecedented TOC levels in surface waters (Fig. 2A). In contrast, if the decline in acid deposition were the main trigger for the recent TOC dynamics, then the observed increase would largely represent a recovery process that returns surface-water TOC to more natural, pre-industrial concentrations (Fig. 2B). The likely interaction of multiple stressors would have led to an even more complex past TOC trajectory (Fig. 2C). Thus, it is indispensable to understand the long-term trajectory of past TOC dynamics to identify the key drivers behind the recent TOC increase and to predict future changes in TOC levels.

One specific research aim in Paper IV was to assess whether past lake-water TOC dynamics across central boreal Sweden, an area that has also seen a widespread increase in TOC levels over the past decades, are lake-specific or if there are similarities across the landscape, which would suggest a common driver behind these dynamics. The second aim was to use VNIR-inferred lake-water TOC reconstructions together with other geochemical and biological proxies to study the long-term Holocene lake-water TOC development and to identify potential driving factors behind the reconstructed TOC dynamics.

6

Figure 2. Schematic diagram showing the response in TOC concentrations in lake ecosystems to (A) the onset of a new stress, (B) the decline of a previous stress, and (C) a combination of multiple stressors from the past to the future (modified from: Battarbee, 1999; Battarbee et al., 2005).

In summary, the specific objectives of each paper were as follows:

Paper I: Can robust FTIR-calibration models for TOC, TIC, and TIC be developed for the 3.6 Ma old sediment record of Lake El’gygytgyn? How did biogeochemical sediment components vary in this sediment record in response to the climatic and environmental evolution of the Arctic from the middle Pliocene to present?

Paper II: Is it possible to develop a stand-alone FTIR-calibration model for quantifying bSi in sediments that is independent of conventional measurements of bSi and their underlying measurement uncertainties?

Paper III: What is the regional environmental development in response to climate change in the Torneträsk area in northernmost subarctic Sweden? How large is the spatial and temporal variability of these responses across such a large region?

Paper IV: What is the long-term trajectory of VNIR-inferred lake-water TOC dynamics in central boreal Sweden? Are these dynamics lake-specific and driven by local factors or similar across the landscape as consequence of common regional factors? What are the driving factors behind the reconstructed TOC dynamics?

7

Methods

Infrared spectroscopy

IR spectroscopy allows the determination of specific sediment components because molecules absorb IR radiation at specific wavelengths (wavenumbers) depending on the structural and atomic composition of the molecules. In this thesis two different spectroscopic methods were used: FTIR spectroscopy – employing the mid-IR part of the electromagnetic spectrum from 2667 to 25000 nm (3750-400 cm-1) – and VNIR spectroscopy – employing the visible to near-IR part of the electromagnetic spectrum from 400 to 2500 nm (25000-4000 cm-1). Prior to IR analyses, sediment samples were freeze-dried, ground to a fine powder to reduce grain size effects, and stored in the same climate controlled room (25±0.2 °C) as the IR spectrometer at least 6 h before the measurement to minimize temperature effects. For FTIR analyses, samples were additionally mixed with oven- dried spectroscopic-grade potassium bromide (KBr) with a ratio of 11 mg sample to 500 mg KBr and subsequently homogenized using a mortar and pestle. A Bruker IFS 66v/S FTIR spectrometer equipped with a diffuse reflectance accessory (Paper I) and a Bruker Vertex 70 equipped with a HTS-XT accessory unit (Papers I-IV) were used for the FTIR measurement in diffuse reflectance mode. VNIR spectra were recorded in diffuse reflectance mode with a NIRSystems 6500 instrument (FOSS NIRSystems Inc.) equipped with a spinning cup module.

Spectral analysis and model development

Prior to model development, linear baseline correction was applied to normalize FTIR spectra and to remove baseline shifts and tilts by setting two points of the recorded spectrum to zero (3,750 and 2,210–2,200 cm-1). For both the recorded FTIR and VNIR spectra, multiplicative scatter-correction (MSC) was applied to separate chemical light absorption from physical light scatter (Geladi et al., 1985). This removes noise arising from varying effective path lengths and particle size and maximizes the remaining signal, which then reflects only compositional differences of the samples. PLS regression modeling was used to develop calibration models between the IR spectral information and the proxy of interest (Wold et al., 2001, and references therein). This regression extension of principle component analyses is particularly suitable for data sets with strongly collinear and numerous X-variables, such as the case for IR spectra. PLS regression allows the modeling of the relationship between two matrices, which in this case are the IR spectral information of a sediment sample (X-variable) and the corresponding concentration of the proxy of interest (Y-variable). Cross-validated coefficient of determination (R2cv) and root mean square error of cross-validation (RMSECV) that resulted from 7-fold cross-validation were used to evaluate the internal model performance. This form of model evaluation uses 6/7 of the available data in a calibration set to predict the proxy of interest of the remaining 1/7 of the data exclusively based on their IR signature.

Independent measurement of biogenic silica

In aquatic systems, bSi is formed by the sedimentation of siliceous microfossils of, e.g., sponges, silicoflagellates, chrysophytes, radiolarians, and phytoliths. However, in most lakes the total bSi pool is predominantly built up of the siliceous remains of diatoms, which are one of the main aquatic primary producers. The bSi concentration in sediments is therefore commonly interpreted as a proxy of diatom abundance, which reflects in most systems the diatom production (Conley and Schelske, 2001). bSi, a hydrated amorphous form of silica, has four distinct main IR absorption bands centered around 470 cm-1, 800 cm-1, 945 cm-1, and 1100 cm-1 (Gendron-Badou et al., 2003; Patwardhan et al., 2006; Schmidt et al., 2001) (Fig. 1), which allow its detection and quantification in sediments via FTIR spectroscopy. Synthetic sediment mixtures with known bSi content were produced to create a FTIR-based bSi quantification in sediments that is independent of conventional-techniques as a calibration reference. These synthetic samples consist of mixtures between purified diatom frustules that were extracted from sediment cores and major sediment components such as quartz and calcite as well as two natural lake sediments, which have a more complex composition. All together the prepared samples cover a bSi gradient from 0 to 100% in ~2% steps.

8

Lake-water TOC reconstruction

Lakes receive different inputs of autochthonous and allochthonous organic matter (OM) depending on climate and catchment characteristics such as vegetation cover and composition, hydrological flow patterns, or nutrient supply (Findlay and Sinsabaugh, 2003). Changes in the quantity and quality of lake-water carbon ultimately determine the composition of the material accumulating at the lake bottom and thus leave their fingerprint in the lake sediment. These changes are recorded in the sediment’s VNIR spectrum, which is particularly sensitive to changes in the OM composition. Past lake-water TOC concentrations were reconstructed by using a calibration model between VNIR spectra of lake surface sediments, the most recently accumulated material, and the corresponding TOC concentrations in the water column (Fig. 3). This model was originally developed by Rosén (2005) using samples from 100 lakes in northern Sweden and was extended by Cunningham et al. (2011) who included an additional 40 lakes from central and southern Sweden that are part of the Swedish national monitoring program. The here presented lake-water TOC reconstructions are based on an updated version of the extended model, which consists of surface sediments from 140 lakes distributed throughout Sweden including nemoral, boreal, and subarctic sites and covers a lake-water TOC gradient from 0.7 to 22 mg·L−1. After reanalysis of the training set samples, the updated model comprises 7 PLS regression components and exhibits a R2cv of 0.65 and a RMSECV of 3.0 mg·L−1 (Fig. 4).

Figure 3. Schematic diagram of important factors influencing lake-water quality and visualization of the calibration approach between the measured lake-water TOC and the VNIR spectral information of the surface sediment.

Other analytical methods

In addition to the information obtained from applications of IR spectroscopy, other geochemical, biological, and isotopic techniques have been used throughout this thesis to obtain a comprehensive view of the reconstructed climatic and environmental variations in the individual studies. Variations in major and trace element concentrations of the sediment were analyzed by wavelength dispersive X-ray fluorescence using a Bruker S8 Tiger spectrometer equipped with an Rh anticathode X-ray tube (Rydberg, 2014), which is located in our paleolimnology laboratories at Umeå University. Besides the element concentrations, the data matrix provided information about, for example, changes in weathering conditions such as through the ratio between elements (e.g., K/Al) or weathering indices that utilize several elements.

9

Figure 4. (left) VNIR-inferred vs. measured lake-water TOC. (right) Distribution of lakes included in the lake-water TOC calibration set across northern (grey) and southern Sweden (green).

Sediment TOC and total nitrogen were determined by using a Flash EA 2000 elemental analyzer (Thermo Fisher Scientific) at the Swedish University of Agricultural Science, Department of Ecology and Forest Management, in Umeå, Sweden. Variations in the ratio between these two components helped to identify changes in the OM source because aquatic-derived OM has generally a high nitrogen content with a low C/N ratio (<10), whereas terrestrial-derived OM has a lower nitrogen content and higher C/N ratios (>20) (Meyers and Ishiwatari, 1993). Molecular-level information about the OM composition in the sediment, specifically lignin concentrations – a proxy of higher plant OM, was also obtained by pyrolysis-GC/MS (Tolu et al., 2015).

Biological proxies included fossil assemblages of pollen, providing information about the vegetation history, and diatoms, which are sensitive to changes in the environmental conditions and thus allow the quantitative reconstruction of lake-water pH and qualify other changes in water quality. Charcoal counts, performed together with the pollen counts, provided information about changes in the past fire frequency.

Stable isotope proxies included bulk OM δ15N and δ13C as well as diatom δ18O and δ30Si. These isotopic ratios helped to reveal past changes in nutrient cycling as well as atmospheric circulation (e.g., Leng and Barker, 2006; Leng et al., 2009; Meyers and Ishiwatari, 1993).

10

Major results and discussion

Advances in the development of IR spectroscopic methods in paleolimnology

In Paper I, calibration models using FTIR spectroscopy and PLS regression were developed for quantifying TOC, TIC, and bSi in the long sediment record of Lake El’gygytgyn in the Far East Russian Arctic (>300 m, 3.6 Ma). These calibration models are based on site-specific calibrations between FITR spectral information and conventionally measured concentrations of TOC (n = 309; 0–2.9% TOC), TIC (n = 152; 0–0.4% TIC), and bSi (n = 420; gradient: 0.9–56.5%). The developed models yield good statistical performances with a R2CV between 0.86 and 0.91 and a RMSECV between 3.1 and 7.0%, respectively (Fig. 4). This shows that robust FTIR models for biogeochemical proxies can be established despite large variations in climatic and environmental conditions that prevailed during the formation of such a long sediment archive over 3.6 Ma. During this time period, the study area transitioned from a forested Arctic landscape in the warm middle Pliocene to the tundra-dominated landscape of today (Brigham-Grette et al., 2013; Melles et al., 2012), sedimentation rates varied over one order of magnitude (4-45 cm ka−1) (Nowaczyk et al., 2013), and long-term diagenetic processes altered the sediment by, for example, affecting the structural composition of bSi (Chapligin et al., 2012). The developed model for these sediment components were subsequently applied to 6771 samples from the entire sediment record of Lake El’gygytgyn to quantitatively examine changes in lake productivity in response to established climatic changes during the middle Pliocene to present.

As stated in the Introduction, conventional analyses come with an inherent uncertainty, which then underlies the FTIR calibration, and the development of an internal calibration – such as for El’gygytgyn – requires additional work. In Paper II, the use of synthetic sediment mixtures with known bSi content (0-100%) allowed the development of a FTIR-calibration model for bSi in sediments that is independent from conventional methods, in contrast to previously developed models. Synthetic mixtures between purified bSi and pure quartz and carbonate (calcite), two common sediment compounds, as well as two natural sediments yielded six individual series (n = 51) with a total of 306 samples, covering the range from 0 to 100% bSi. The PLS regression between the defined bSi content and the FTIR spectra of the samples exhibits a strong model performance with a R2cv of 0.97 and a RMSECV of 4.7%. This strong correlation between defined and FTIR-inferred bSi demonstrates the direct relationship between the bSi content of a sample and its FTIR signature. Subsequent application of this model to natural, lacustrine and marine, sediments shows a good correlation between FTIR-inferred and wet-chemically measured bSi concentrations (R2 = 0.87; RMSEP = 5.9%). Particularly good agreements between the two measuring techniques exist for samples with comparatively simple sediment compositions, in which variations in bSi dominate compositional changes. At present, the used synthetic mixtures do not include the larger variations in the quantity and quality of OM that occur in natural samples; this may explain the somewhat lower prediction accuracy for organic-rich samples.

Figure 4. Conventionally measured versus FTIR-inferred concentrations of (a.) biogenic silica (bSi), (b.) total organic carbon (TOC), and (c.) total inorganic carbon (TIC) based on the internal calibrations from Lake El’gygytgyn (modified from Paper I).

11

FTIR-inferred bSi concentrations for the same record based on the independent calibration and an internal calibration, which relies on reference values from conventional wet-chemical measurements, reflect the same trends, but differ with respect to absolute values (Fig. 5). The independent calibration yields lower concentrations for samples with low bSi content compared to conventional measurements and therefore also to the internal calibration model. Several studies have shown the potential bSi overestimation by wet-chemical digestion methods, particularly in samples with low bSi content, because of the dissolution of minerogenic silica from alumino-silicates during the sample digestion (Clymans et al., 2015; Kamatani and Oku, 2000; Ohlendorf and Sturm, 2008). By using FTIR spectroscopy and PLS regression modeling, it is possible to distinguish between biogenic (amorphous) Si and other crystalline minerals that contain Si. Furthermore, the lower bSi values given by the independent calibration are supported by total diatom counts that were available for one of the tested sediment sequences. In general, the general good agreement regarding the inferred trends in bSi highlights the robustness of bSi quantifications by FTIR spectroscopy.

The results of Paper II show that a FTIR calibration model for quantifying bSi in sediments can be developed by using synthetic sediments that is independent of conventional wet-chemical measurements and their underlying measurement uncertainties. The presented approach allows the establishment of FTIR-spectroscopy as a stand-alone technique for quantifying bSi in sediments as an alternative to conventional methods. This approach further eliminates the additional work that is required for each new project when developing an internal calibration based on conventional reference values. The independent calibration model has consequently been used to quantify bSi in the sediment records that were studied in Papers III and IV.

Figure 5. Downcore plot of FTIR-inferred (line plots) and wet-chemically measured (open circles) biogenic silica (bSi) concentrations for the sediment sequence from Lake El’gygytgyn. FTIR-inferred concentrations are based on the independent calibration (black line) and internal calibration based on wet-chemical measurements (grey line) (Paper I) (modified from Paper II).

IR spectroscopic methods applied in paleolimnological studies

3.6 million years of lake-ecosystem response to climate variability in the Siberian Arctic

Lake El’gygytgyn is a 170-m deep and 110 km2 lake in the Far East Russian Arctic. The sediment archive of Lake El’gygytgyn offers the unique possibility to study climatic and environmental changes in the Arctic over the past 3.6 Ma because this region has never been directly affected by Northern Hemisphere glaciation and has thus continuously accumulated sediment since its formation following a meteorite impact (Brigham-Grette et al., 2013; Melles et al., 2012). By applying the internal FTIR-calibration models for TOC, TIC, and bSi, described previously in the thesis, to 6771 samples throughout this 318-m long record, it was possible to obtain a detailed insight into the variability of these three sediment components from the middle/late Pliocene to the Quaternary (Paper I). In Lake El’gygytgyn, long-term variations in aquatic bioproduction are primarily driven by climate change, which ultimately controls the light availability, recycling of nutrients, and allochthonous nutrient

12

supply to the lake (e.g., Melles et al., 2006). FTIR-inferred concentrations show strong variations throughout the sediment sequence ranging from 0 to 2.2% for TOC, from 0 to 5.2% for TIC, and from 1 to 56% for bSi (Fig. 6). Carbonates, as indicated by the TIC content, occur primarily at the base of the sediment record and are potentially the result of initial hydrothermal activity induced by the meteorite impact.

Because sedimentation rates have changed over one order of magnitude (4-45 cm ka−1) during the last 3.6 Ma (Nowaczyk et al., 2013), accumulation rates were calculated to assess quantitative changes in the FTIR-inferred proxies. These revealed similarly large variations with ranges of 0-5.2 g∙m-2∙yr-1 for TOC and 1-70 g∙m-2∙yr-1 for bSi, respectively (Fig. 6). Highest ARbSi, indicating favorable conditions for diatom production, occurred during the period 3.6 to 3.3 Ma ago. This period overlaps with the early Pliocene warm period (~5-3 Ma ago) – a period that is characterized by ~2-3°C higher global surface temperatures compared to today (Brierley et al., 2009; Haywood and Valdes, 2004), reduced high-latitude ice sheet extents (Hill et al., 2010; McKay et al., 2012), as well as a global sea level 10-40 m above present (Raymo et al., 2011). This bioproduction maximum is followed by a stepwise drop in ARbSi during the period 3.3 to 3.15 Ma ago and a further gradual ARbSi decrease during the past 3.15 Ma. Compared to average ARbSi during the middle Pliocene, maximum ARbSi are 1.4 times lower between 3.15-1.5 Ma ago and 2.8 times lower between 1.5 Ma to 0.125 Ma ago. This long-term decline in ARbSi, which suggests a long-term deterioration of climatic conditions for diatom production, agrees with the first increased occurrence of glacial periods and finally full establishment of high amplitude glacial/interglacial cycles during the Quaternary (Brigham-Grette et al., 2013; Melles et al., 2012). Elevated ARTOC are associated with interglacial periods, when in-lake and terrestrial-derived OM inputs exceed decomposition, as well as with glacial periods, when OM is preserved due to oxygen-depleted bottom waters.

The agreement between ARbSi and the established general variations in climate emphasizes the climate sensitivity of diatom production in Lake El’gygytgyn, and thus the value of this record for assessing long-term climate variability in the Arctic. Interestingly, ARbSi were not significantly elevated during the so-called “super interglacials” marine isotope stages 31 and 11c despite particularly high pollen-inferred temperatures of up to 4-5°C warmer than present (Melles et al., 2012; Vogel et al., 2013). The here presented FTIR-inferred concentrations and accumulation rates of bSi contributed to the synopsis of the results of the El’gygytgyn drilling project, which provide a comprehensive view of the climatic and environmental development in the Siberian Artic over the past 3.6 Ma (Brigham-Grette et al., 2013; Melles et al., 2012).

Sensitivity of a large-lake system to Holocene climate changes in subarctic Sweden

Over the past century, ongoing climate change has thus far led to a mean annual temperature increase of 2.5°C in the Torneträsk area and a decrease of the mean lake ice-cover duration of Torneträsk by ~40 days (Callaghan et al., 2013). Several paleoenvironmental studies of small lakes in the area showed the sensitivity of these lakes to climatic and environmental changes during the Holocene. For example, diatom and chironomid communities responded to temperature variations (Bigler et al., 2006; Bigler et al., 2002) and higher elevation lakes recorded changes in glacier activities and catchment erosion (Berglund et al., 1996; Rubensdotter and Rosqvist, 2003; Snowball and Sandgren, 1996). In Paper III, the changes recorded in these small lakes, which are likely more sensitive to local catchment characteristics, were compared to those changes recorded in Torneträsk itself, which is a large lake (70-km long; 330 km2) that integrates changes across the entire region. This comparison allows an assessment of the spatial and temporal variations in the responses to climate forcing across the region and to determine when the magnitude of environmental change was sufficient to affect a large lake such as Torneträsk and by inference the entire region rather than only registering in more sensitive, local areas.

Over the past 10000 years, three periods with distinct changes in response to major regional changes in the landscape were recorded in the sediment record of Torneträsk (Paper III). Significant changes in the sediment composition of the lake followed a) the gradual early landscape development between c. 10000 and c. 5300 cal yr BP, b) increased soil erosion in the catchment from c. 1300 until c. 100 cal

13

Figure 6. (a.) LR04 global marine isotope stack for the past 3.6 Ma compared with (b.) FTIR-inferred biogenic silica (bSi) concentrations (line plot) and accumulation rates (ARbSi; area plot; note logarithmic scale), (c.) FTIR-inferred total organic carbon (TOC) concentrations (line plot) and accumulation rates (ARTOC; area plot), and (d.) FTIR-inferred total inorganic carbon (TIC) concentrations (line plot) and accumulation rates (ARTIC; area plot) in the Lake El´gygytgyn sediment record. Periods of high ARbSi are highlighted with orange bars (3.15 Ma to 0.125 Ma: ARbSi > 10 g∙m-2∙yr-1). Dashed lines indicate breaks in the scale with maximum values plotted above (modified from Paper I).

14

yr BP, and c) the ongoing climate change (Fig. 7). However, the timing and type of response to these different forcings differed fundamentally. During the early landscape development, changes in the sediment record reflect the successive development of vegetation and soils, which drove the element cycling in the landscape and in Torneträsk. For example, initially slow chemical weathering in the catchment limited the nutrient supply to the lake and suppressed notable diatom production until c. 7700 cal yr BP. These earlier dynamics likely superposed potential lake-ecosystem responses to climate variations that have been observed in small lakes situated in the Torneträsk catchment (e.g., Bigler et al., 2002; Hammarlund et al., 2002).

Compared to this gradual response to the successive early landscape development, the second major change at c. 1300 cal yr BP was recorded as an abrupt change in sediment composition and occurred following a period of c. 4000 years of relatively stable sediment geochemistry. At 1300 cal yr BP, several proxies show a change in the sediment composition toward material with a more terrestrial-derived and less-weathered composition, which would reflect an increased erosion of soils in the Torneträsk catchment. Paleoenvironmental studies of small lakes in the Torneträsk catchment show first indications for a gradual climate cooling in the area during the late Holocene that led to a treeline

Figure 7. Weathering degree (WD), FTIR-inferred biogenic silica (bSi), total organic carbon (TOC), total nitrogen (TN), C:N ratios, δ13C, and δ15N analyzed in the sediment profile Co1282 from Torneträsk for the past 10000 years, showing three periods of major change in the sediment composition associated with the early landscape development, increased soil erosion during the late Holocene, and ongoing climate change (from Paper III).

15

retreat and increased soil erosion at higher elevations (Berglund et al., 1996; Rubensdotter and Rosqvist, 2003; Snowball and Sandgren, 1996). However, these changes started already ~2000 years before this increase in catchment soil erosion was registered in Torneträsk. The abrupt and delayed or late response suggests that a certain threshold for soil erosion and transportation was crossed around 1300 cal yr BP. From that point onwards, the magnitude of change related to the gradual, ongoing climate cooling was sufficient to destabilize soils and increase soil erosion across the region and finally also affect the large-lake ecosystem rather than only specific areas of the watershed that are more sensitive to environmental changes.

The third recorded major ecosystem change is related to the ongoing climate change over the last century. The composition of the sediment that accumulated in Torneträsk during the past century returned thus far to a composition that is similar to that prior to 1300 cal yr BP, which suggests a stabilization of the catchment, such as with increased vegetation cover. In contrast to the gradual response to the early landscape development and to the abrupt but late response to gradual climate cooling during the late Holocene, the ongoing lake ecosystem response to recent climate change was immediate and thus far of equal magnitude. This recent, rapid change highlights the unprecedented scale of ongoing climate change in subarctic Fennoscandia during the past century.

Long-term perspective on recent changes in lake-water TOC

During the 1980’s–1990’s reconstructions of lake-water pH in Sweden using diatom assemblages had shown that underlying the modern acidification of surface waters in southern Sweden there were other, earlier human-related changes in pH (Renberg, 1990; Renberg and Hellberg, 1982; Renberg et al., 1993). Namely, after deglaciation >10000 cal yr BP lakes underwent a natural acidification and during much of the Holocene pH values in the SW Swedish lakes were below pH 6. However, beginning 2500-1500 cal yr BP, the pH in the studied lakes increased to values >6 in conjunction with evidence of human land use. As land-use practices changed in the late 19th century, the culturally elevated pH in these lakes began to decline – already in advance of the later effects of acid rain during the latter half of the 20th century. In similar fashion, application of the VNIR-inferred model for lake-water TOC shows a more complex history involving long-term changes beyond those observed in monitoring.

As in other areas across Europe and N America (e.g., Driscoll et al., 2003; Monteith et al., 2007; Skjelkvale et al., 2005; Worrall et al., 2004), increasing TOC concentrations have also been recorded in lakes in central boreal Sweden during the past three decades (Swedish national environmental monitoring). VNIR-inferred lake-water TOC concentrations for four lakes across this region (Paper IV) show that the recent increase recorded by monitoring data was preceded by a long-term decrease that started already between AD 1450 and AD 1600. At the four study sites, TOC concentrations decreased from values in the range of 12 to 18 mg L-1 to minimum values between 8 and 10 mg L-1 in the early to mid-20th century, and since then have returned subsequently to levels around 10-16 mg L-1 during recent times. The common past lake-water TOC trajectory at all studies sites, including an earlier study in one lake in SW Sweden (Rosén et al., 2011a) as well as a recent study of two lakes in S Sweden (Bragee et al., 2015), demonstrates the landscape-wide significance of this past decline and suggests a common driver behind these dynamics.

The multi-proxy analysis of the complete Holocene record of one of the studied lakes, Lång-Älgsjön, showed that relatively stable lake-water TOC levels prevailed for almost 9000 years prior to the decline at c. AD 1450. Although climate is one of the hypothesized drivers of the recent increase in TOC, significant changes in climate during this period had limited influence on the reconstructed TOC levels. During the past 550 years, geochemical and biological proxies reveal that lake-water TOC dynamics are strongly coupled to historical patterns in human landscape utilization in central boreal Sweden. For example, coinciding with the TOC decline, several proxies indicate increasing catchment disturbance such as an enhanced catchment erosion and input of terrestrial OM, as well as the increased occurrence of pollen associated with a more open forest landscape and cultural plants (Fig. 8). These dynamics could be linked to widespread but low-intensive summer forest grazing and small-scale farming that expanded in importance in central boreal Sweden in the 15th century. This

16

type of landscape utilization, which occurred at summer forest farms several km from the home villages, included the grazing of livestock in the forest, production and processing of milk, haymaking on mires for fodder, and occasionally the cultivation of cereals. To improve forest grazing, the canopy cover was opened and unwanted field-layer vegetation such as dwarf shrubs and mosses was grazed, removed or burned, leading to an altered, more-open forest composition (Segerström and Emanuelsson, 2002; Segerström et al., 1996). The alteration of the forest structure and continuous removal of biomass would reduce the export of labile carbon from the terrestrial to the aquatic system and thus lead to a TOC decline in surface waters.

Figure 8. Comparison of (a) the VNIR-inferred lake-water total organic carbon (TOC) development to geochemical and biological proxy data analyzed in the sediment record of Lång-Älgsjön for the past 10500 years: (b) sediment total organic carbon (TOC); (c) total lignin, indicating changes in the input of terrestrial-derived organic matter (OM); (d) FTIR-inferred biogenic silica (bSi); (e) K/Al ratio, indicating changes in the weathering degree of the sediment and consequently changes in catchment erosion; (f) diatom-inferred pH reconstruction; (g) grouped diatom assemblage composition; and pollen counts of (h) apophytes (pollen from native plants favored by human disturbance) and (i) anthropochores (pollen from cultivated plants) (modified from Paper IV).

Summer forest farming and grazing gradually became less important in central Sweden during the 19th century following the industrial revolution and vanished in the vicinity of Lång-Älgsjön during the 1940’s (Swedish National Heritage Board database). Shortly after the cessation of this type of landscape utilization lake-water TOC levels started to return towards pre-human impact conditions. The results of Paper IV do not allow isolating the contribution of other hypothesized drivers, such as climate change and the recovery from acidification, to ongoing lake-water TOC dynamics. However, the effect of these processes on lake-water TOC dynamics during the past 10500 years appears to be minor compared to the scale of change associated with the impact of extensive early human-landscape interactions in the study area. Paper IV demonstrates that early land use strongly affected past TOC dynamics and that its impact on ongoing changes in surface-water TOC has thus far been underestimated.

17

Conclusions

The results of this doctoral thesis demonstrate the value and further potential of IR spectroscopic methods in paleolimnological studies. Their advantages (a simple sample pre-treatment, the non- or minimal consumption of sample material, a rapid measurement time, and the multitude of information on organic and inorganic components) make these techniques cost-efficient alternatives to conventional methods to quantify specific sediment components, such as bSi, as well as to allow the reconstruction of past environmental conditions that are otherwise challenging to infer, such as past lake-water TOC levels.

The application of FTIR spectroscopy in the context of the El’gygytgyn drilling project facilitated a detailed quantification of bSi in >6000 samples, which would have been very time-consuming with conventional techniques (Paper I). In addition to bSi, the use of FTIR spectroscopy provided measurements of two more sediment components, TOC and TIC, through the same analyses. One limitation of using FTIR spectroscopy and PLS regression modeling to quantify bSi is the dependence of this approach on conventional methods and their inherent uncertainties as a calibration reference. This hitherto existing limitation was overcome by creating and then analyzing synthetic sediment mixtures with known bSi content (Paper II). Applications of the resulting calibration model to natural sediment samples yield robust bSi measurements. This further development represents an important step forward toward establishing FTIR spectroscopy as a stand-alone technique for measuring bSi in sediments.

Applied as part of multi-proxy studies (Paper I, III, IV), IR spectroscopic methods contributed to answering specific research questions in this thesis that aimed at improving our understanding of natural long-term environmental variations in response to climate change in northern landscapes as well as providing a long-term perspective on recent environmental changes, which include a range of human impacts (Paper IV). The main conclusions of these studies are as follows:

Paper I: FTIR-inferred proxies related to bioproduction in the sediment record of Lake El’gygytgyn provided a detailed insight into long-term climate variability in the Siberian Arctic over the past 3.6 million years. Highest in-lake production occurred during the warm middle Pliocene (3.6-3.3 Ma) and gradually declined, indicating a successive climate deterioration, during the late Pliocene and Quaternary. Variations in bioproduction agree with established climate patterns such as the first occurrence of glacial periods and the finally full establishment of glacial–interglacial cycles during the Quaternary.

Paper III: Comparing the response of a large-lake system to past climatic and environmental forcings with those recorded in the sediments of more climate sensitive small lakes from its catchment allowed determining the significance and magnitude of changes across a larger region. Over the Holocene three different types of responses were recorded in the sediment of the subarctic, 330-km2 lake Torneträsk in northern Sweden: i) a gradual response to the early landscape development following deglaciation, ii) an abrupt but delayed response following climate cooling during the late Holocene, and iii) an immediate response to ongoing climate change during the past century. This rapid, recent response recorded in the comparatively climate insensitive record from a large lake emphasizes the unprecedented scale of ongoing climate change in northern Fennoscandia.

Paper IV: Past lake-water TOC reconstructions by VNIR spectroscopy across central Sweden revealed that early human land use activities in the form of widespread summer forest grazing and farming strongly affected past lake-water TOC dynamics and still contribute to ongoing surface-water TOC changes. This thus far underestimated role of early human-landscape interactions should be considered when assessing ongoing TOC changes and formulating appropriate reference conditions for water management.

18

Future prospects

The ultimate goal is to establish the independent calibration model for quantifying bSi in sediments as a stand-alone technique that is freely available to anyone. However, before this can be achieved, current limitations and knowledge gaps regarding its universal applicability need to be resolved. For example, the calibration model exhibits a bias to somewhat lower values for organic-rich samples because of the fact that no such samples are currently included in the synthetic sediment mixtures that underlie the model. To overcome this current limitation and assure a reliable applicability to a wide range of sediment types, it would be important to add additional synthetic mixtures that reflect the full range of sediment matrices encountered in natural sediments. The obvious next step would therefore be to include synthetic mixtures between bSi and OM into in the independent calibration. Further remaining questions are for example: To what extent the method might be biased by other sediment compounds that contain amorphous forms of SiO2 such as volcanic glasses, duripans or hydrothermal concretions? In general, these components do not largely contribute to lacustrine sediments, but Si contributions from these components have shown to be significant when analyzing samples with low bSi content by conventional wet-chemical methods (Clymans et al., 2015). Similar to the independent calibration model for bSi, it could be possible to develop calibration models for other major sediment components such as carbonates, quartz, feldspar, and clay minerals.

VNIR-inferred lake-water TOC concentrations in Paper II and by Cunningham et al. (2011) are in good agreement with monitoring data that are available for some of the studied lakes since the late 1980’s. These results corroborate the credibility of the sediment-inferred lake-water TOC levels. Yet it remains unclear which specific organic sediment constituents that are recorded in the VNIR spectrum allow the transfer function for lake-water TOC. To answer this question would not only help to reduce the uncertainty of the method but also to better understand the processes that lead to variations in TOC levels in surface waters. One promising way to achieve this goal could be the use of pyrolysis-GC/MS. The technique provides molecular-level chemical information about the composition of OM in sediments, which can help to determine the OM origin and the processes contributing to its composition (Tolu et al., 2015). An analysis of the sediment samples included in the calibration set for lake-water TOC by pyrolysis GC/MS may allow determining the components driving the transfer function as well as to characterize variations in the OM quality in lakes across nemoral, boreal, and subarctic Sweden and possibly connect these to specific catchment characteristics.

The fact that a model comparable to the Swedish lake-water TOC could be developed for Arctic Canada (Rouillard et al., 2011) demonstrates that the approach is not restricted to Scandinavian lakes and thus has general applicability. If these two calibration sample sets could successfully be combined, it might be possible to create a universal model for northern lakes that could be applied to lakes that are outside of the geographical calibration range of these two data sets, and thus allow an application of the technique in other regions without the need to first generate a cost and labor intensive local calibration set. Such a unified model would open the way to study the effect of different types of early human activities on carbon cycling in other regions, particularly in those that are currently experiencing changes in surface-water TOC concentrations such as the United Kingdom or NE America.

19

References

Åkerman, H.J., Johansson, M., 2008. Thawing permafrost and thicker active layers in sub-arctic Sweden. Permafrost Periglac.

19, 279-292.

Anderson, N.J., Rippey, B., Gibson, C.E., 1993. A comparison of sedimentary and diatom-inferred phosphorus profiles:

implications for defining pre-disturbance nutrient conditions. Hydrobiologia 253, 357-366.

Armstrong, A., Holden, J., Kay, P., Francis, B., Foulger, M., Gledhill, S., McDonald, A.T., Walker, A., 2010. The impact of

peatland drain-blocking on dissolved organic carbon loss and discolouration of water; results from a national survey. J.

Hydrol. 381, 112-120.

Balsam, W.L., Deaton, B.C., 1996. Determining the composition of late Quaternary marine sediments from NUV, VIS, and NIR

diffuse reflectance spectra. Mar. Geol. 134, 31-55.

Barnekow, L., 1999. Holocene tree-line dynamics and inferred climatic changes in the Abisko area, northern Sweden, based on

macrofossil and pollen records. Holocene 9, 253-265.

Barth, A., 2007. Infrared spectroscopy of proteins. Biochim. Biophys. Acta-Bioenerg. 1767, 1073-1101.

Battarbee, R.W., 1984. Diatom analysis and the acidification of lakes. Philos. Trans. R. Soc. B-Biol. Sci. 305, 451-477.

Battarbee, R.W., 1999. The importance of palaeolimnology to lake restoration. Hydrobiologia 395, 149-159.

Battarbee, R.W., Anderson, N.J., Jeppesen, E., Leavitt, P.R., 2005. Combining palaeolimnological and limnological approaches

in assessing lake ecosystem response to nutrient reduction. Freshw. Biol. 50, 1772-1780.

Battarbee, R.W., Jones, V.J., Flower, R.J., Cameron, N.G., Bennion, H., Carvalho, L., Juggins, S., 2001. Diatoms. Springer,

Netherlands.

Belzile, N., Joly, H.A., Li, H., 1997. Characterization of humic substances extracted from Canadian lake sediments. Can. J.

Chemistry 75, 14-27.

Bennion, H., Fluin, J., Simpson, G.L., 2004. Assessing eutrophication and reference conditions for Scottish freshwater lochs

using subfossil diatoms. J. Appl. Ecol. 41, 124-138.

Berglund, B.E., Barnekow, L., Hammarlund, D., Sandgren, P., Snowball, I.F., 1996. Holocene forest dynamics and climate

changes in the Abisko area, northern Sweden: the Sonesson model of vegetation history reconsidered and confirmed. Ecol.

Bull., 15-30.

Bertaux, J., Fröhlich, F., Ildefonse, P., 1998. Multicomponent analysis of FTIR spectra; quantification of amorphous and

crystallized mineral phases in synthetic and natural sediments. J. Sediment. Res. 68, 440.

Bertaux, J., Ledru, M.P., Soubiès, F., Sondag, F., 1996. The use of quantitative mineralogy linked to palynological studies in

palaeoenvironmental reconstruction: the case study of the" Lagoa Campestre" lake, Salitre, Minas Gerais, Brazil. C.R. Acad.

Sci. II 323, 65-72.

Bigler, C., Larocque, I., Peglar, S.M., Birks, H.J.B., Hall, R.I., 2002. Quantitative multiproxy assessment of long-term patterns of

Holocene environmental change from a small lake near Abisko, northern Sweden. Holocene 12, 481-496.

Bindler, R., Renberg, I., Rydberg, J., Andrén, T., 2009. Widespread waterborne pollution in central Swedish lakes and the Baltic

Sea from pre-industrial mining and metallurgy. Environ. Pollut. 157, 2132-2141.

Boyle, J.F., 2000. Rapid elemental analysis of sediment samples by isotope source XRF. J. Paleolimn. 23, 213-221.

Boyle, J.F., Mackay, A.W., Rose, N.L., Appleby, P.G., 1998. Sediment heavy metal record in Lake Baikal: natural and

anthropogenic sources. J. Paleolimn. 20, 135-150.

Bragée, P., Mazier, F., Nielsen, A.B., Rosén, P., Fredh, D., Broström, A., Granéli, W., Hammarlund, D., 2015. Historical TOC

concentration minima during peak sulfur deposition in two Swedish lakes. Biogeosciences 12, 307-322.

Braguglia, C.M., Campanella, L., Petronio, B.M., Scerbo, R., 1995. Sedimentary humic acids in the continental margin of the

Ross Sea (Antarctica). Int. J. Environ. An. Ch. 60, 61-70.

Brierley, C.M., Fedorov, A.V., Liu, Z.H., Herbert, T.D., Lawrence, K.T., LaRiviere, J.P., 2009. Greatly expanded tropical warm

pool and weakened Hadley circulation in the early Pliocene. Science 323, 1714-1718.

Brigham-Grette, J., Melles, M., Minyuk, P., Andreev, A., Tarasov, P., DeConto, R., Koenig, S., Nowaczyk, N., Wennrich, V.,

Rosén, P., Haltia, E., Cook, T., Gebhardt, C., Meyer-Jacob, C., Snyder, J., Herzschuh, U., 2013. Pliocene warmth, polar

amplification, and stepped Pleistocene cooling recorded in NE Arctic Russia. Science 340, 1421-1427.

Brisset, E., Miramont, C., Guiter, F., Anthony, E.J., Tachikawa, K., Poulenard, J., Arnaud, F., Delhon, C., Meunier, J.D., Bard, E.,

Suméra, F., 2013. Non-reversible geosystem destabilisation at 4200 cal. BP: Sedimentological, geochemical and botanical

markers of soil erosion recorded in a Mediterranean alpine lake. Holocene 23, 1863-1874.

Brännvall, M.L., Bindler, R., Renberg, I., Emteryd, O., Bartnicki, J., Billström, K., 1999. The Medieval metal industry was the

cradle of modern large scale atmospheric lead pollution in northern Europe. Environ. Sci. Technol. 33, 4391-4395.

Calace, N., Capolei, M., Lucchese, M., Petronio, B., 1999. The structural composition of humic compounds as indicator of

organic carbon sources. Talanta 49, 277-284.

20

Calace, N., Cardellicchio, N., Petronio, B.M., Pietrantonio, M., Pietroletti, M., 2006. Sedimentary humic substances in the

northern Adriatic sea (Mediterranean sea). Mar. Environ. Res. 61, 40-58.

Callaghan, T.V., Jonasson, C., Thierfelder, T., Yang, Z.L., Hedenås, H., Johansson, M., Molau, U., Van Bogaert, R., Michelsen, A.,

Olofsson, J., Gwynn-Jones, D., Bokhorst, S., Phoenix, G., Bjerke, J.W., Tømmervik, H., Christensen, T.R., Hanna, E., Koller,

E.K., Sloan, V.L., 2013. Ecosystem change and stability over multiple decades in the Swedish subarctic: complex processes

and multiple drivers. Philos. Trans. R. Soc. B-Biol. Sci. 368.

Chang, C.-W., You, C.-F., Huang, C.-Y., Lee, T.-Q., 2005. Rapid determination of chemical and physical properties in marine

sediments using a near-infrared reflectance spectroscopic technique. Appl. Geochem. 20, 1637-1647.

Chapligin, B., Meyer, H., Swann, G.E.A., Meyer-Jacob, C., Hubberten, H.W., 2012. A 250 ka oxygen isotope record from diatoms

at Lake El'gygytgyn, far east Russian Arctic. Clim. Past. 8, 1621-1636.

Chapman, S., Campbell, C., Fraser, A., Puri, G., 2001. FTIR spectroscopy of peat in and bordering Scots pine woodland:

relationship with chemical and biological properties. Soil Biol. Biochem. 33, 1193-1200.

Chester, R., Elderfield, H., 1968. The infrared determination of opal in siliceous deep-sea sediments. Geochim. Cosmochim. Acta

32, 1128-1140.

Clymans, W., Barão, L., Van der Putten, N., Wastegård, S., Gísladóttir, G., Björck, S., Moine, B., Struyf, E., Conley, D.J., 2015.

The contribution of tephra constituents during biogenic silica determination: implications for soil and palaeoecological

studies. Biogeosciences 12, 3789-3804.

Cocozza, C., D'Orazio, V., Miano, T., Shotyk, W., 2003. Characterization of solid and aqueous phases of a peat bog profile using

molecular fluorescence spectroscopy, ESR and FT-IR, and comparison with physical properties. Org. Geochem. 34, 49-60.

Conley, D.J., 1998. An interlaboratory comparison for the measurement of biogenic silica in sediments. Mar. Chem. 63, 39-48.

Conley, D.J., Schelske, C.L., 2001. Biogenic silica, In: Smol, J.P., Birks, H.J.B., Last, W.M. (Eds.), Tracking Environmental

Change Using Lake Sediments, vol 3, Terrestrial Algal, and Siliceous Indicators. Kluwer Acad., Dordrecht, Netherlands, pp.

281-293.

Cunningham, L., Bishop, K., Mettavainio, E., Rosén, P., 2011. Paleoecological evidence of major declines in total organic carbon

concentrations since the nineteenth century in four nemoboreal lakes. J. Paleolimn. 45, 507-518.

Cunningham, L., Vogel, H., Nowaczyk, N., Wennrich, V., Juschus, O., Persson, P., Rosén, P., 2013a. Climatic variability during

the last interglacial inferred from geochemical proxies in the Lake El’gygytgyn sediment record. Palaeogeogr. Palaeoecol.

386, 408-414.

Cunningham, L., Vogel, H., Wennrich, V., Juschus, O., Nowaczyk, N., Rosén, P., 2013b. Amplified bioproductivity during

Transition IV (332 000–342 000 yr ago): evidence from the geochemical record of Lake El'gygytgyn. Clim. Past 9, 679-686.

Das, B., 2007. Reconstruction of historical productivity using visible-near-infrared (VNIR) reflectance properties from boreal

and saline lake sediments. Aquat. Ecol. 41, 209-220.

Das, B., Vinebrooke, R.D., Sanchez-Azofeifa, A., Rivard, B., Wolfe, A.P., 2005. Inferring sedimentary chlorophyll concentrations

with reflectance spectroscopy: a novel approach to reconstructing historical changes in the trophic status of mountain lakes.

Can. J. Fish. Aquat. Sci. 62, 1067-1078.

De Wit, H.A., Mulder, J., Hindar, A., Hole, L., 2007. Long-term increase in dissolved organic carbon in streamwaters in Norway

is response to reduced acid deposition. Environ. Sci. Technol. 41, 7706-7713.

Dearing, J.A., Jones, R.T., 2003. Coupling temporal and spatial dimensions of global sediment flux through lake and marine

sediment records. Glob. Planet. Change 39, 147-168.

Driscoll, C.T., Driscoll, K.M., Roy, K.M., Mitchell, M.J., 2003. Chemical response of lakes in the Adirondack Region of New York

to declines in acidic deposition. Environ. Sci. Technol. 37, 2036-2042.

Ekström, S.M., Kritzberg, E.S., Kleja, D.B., Larsson, N., Nilsson, P.A., Graneli, W., Bergkvist, B., 2011. Effect of Acid Deposition

on Quantity and Quality of Dissolved Organic Matter in Soil-Water. Environ. Sci. Technol. 45, 4733-4739.

Engstrom, D.R., Wright, H., 1984. Chemical stratigraphy of lake sediments as a record of environmental change. In: Haworth,

E.Y., Lund J.W. (Eds.), Lake sediments and environmental history: studies in palaeolimnology and palaeoecology in honour

of Winifred Tutin. Leicester University Press, pp. 11–67.

Evans, C.D., Jones, T.G., Burden, A., Ostle, N., Zielinski, P., Cooper, M.D.A., Peacock, M., Clark, J.M., Oulehle, F., Cooper, D.,

Freeman, C., 2012. Acidity controls on dissolved organic carbon mobility in organic soils. Glob. Change Biol. 18, 3317-3331.

Evans, C.D., Monteith, D.T., Cooper, D.M., 2005. Long-term increases in surface water dissolved organic carbon: Observations,

possible causes and environmental impacts. Environ. Pollut. 137, 55-71.

Farmer, V.C., 1974. The infrared spectra of minerals, Mineralogical Society Monograph 4. Adlard & Son, Dorking, Surrey

Fenner, N., Freeman, C., 2011. Drought-induced carbon loss in peatlands. Nat. Geosci. 4, 895-900.

Findlay, S., Sinsabaugh, R.L., 2003. Aquatic ecosystems: Interactivity of dissolved organic matter. Elsevier Academic,

Burlington, MA.

21

Foley, W.J., McIlwee, A., Lawler, I., Aragones, L., Woolnough, A.P., Berding, N., 1998. Ecological applications of near infrared

reflectance spectroscopy a tool for rapid, cost-effective prediction of the composition of plant and animal tissues and aspects

of animal performance. Oecologia 116, 293-305.

Freeman, C., Evans, C., Monteith, D., Reynolds, B., Fenner, N., 2001. Export of organic carbon from peat soils. Nature 412, 785.

Geladi, P., Macdougall, D., Martens, H., 1985. Linearization and scatter-correction for near-infrared reflectance spectra of meat.

Appl. Spectrosc. 39, 491-500.

Gendron-Badou, A., Coradin, T., Maquet, J., Fröhlich, F., Livage, J., 2003. Spectroscopic characterization of biogenic silica. J.

Non-Cryst. Solids 316, 331-337.

Hahn, A., Kliem, P., Ohlendorf, C., Zolitschka, B., Rosén, P., Team, T.P.S., 2013. Climate induced changes as registered in

inorganic and organic sediment components from Laguna Potrok Aike (Argentina) during the past 51 ka. Quat. Sci. Rev. 71,

154-166.

Hahn, A., Rosén, P., Kliem, P., Ohlendorf, C., Zolitschka, B., 2011. Comparative study of infrared techniques for fast

biogeochemical sediment analyses. Geochem. Geophy. Geosy. 12.

Hammarlund, D., Barnekow, L., Birks, H.J.B., Buchardt, B., Edwards, T.W.D., 2002. Holocene changes in atmospheric

circulation recorded in the oxygen-isotope stratigraphy of lacustrine carbonates from northern Sweden. Holocene 12, 339-

351.

Haywood, A.M., Valdes, P.J., 2004. Modelling Pliocene warmth: contribution of atmosphere, oceans and cryosphere. Earth

Planet. Sci. Lett. 218, 363-377.

Hejzlar, J., Dubrovsky, M., Buchtele, J., Ruzicka, M., 2003. The apparent and potential effects of climate change on the inferred

concentration of dissolved organic matter in a temperate stream (the Malse River, South Bohemia). Sci. Total Environ. 310,

143-152.

Hill, D.J., Dolan, A.M., Haywood, A.M., Hunter, S.J., Stoll, D.K., 2010. Sensitivity of the Greenland Ice Sheet to Pliocene sea

surface temperatures. Stratigraphy 7, 111-121.

Hongve, D., Riise, G., Kristiansen, J.F., 2004. Increased colour and organic acid concentrations in Norwegian forest lakes and

drinking water - a result of increased precipitation? Aquat. Sci. 66, 231-238.

IPCC: Climate change 2007: The physical science basis, in: Contribution of Working Group I to the Fourth Assessment Report of

the Intergovernmental Panel on Climate Change, edited by: Solomon, S., Qin, D., Manning, M., Chen, Z., Marquis, M.,

Averyt, K. B., Tignor, M., and Miller, H. L., Cambridge University Press, Cambridge and New York, 2007

Jones, V.J., Solovieva, N., Self, A.E., McGowan, S., Rosén, P., Salonen, J.S., Seppä, H., Valiranta, M., Parrott, E., Brooks, S.J.,

2011. The influence of Holocene tree-line advance and retreat on an arctic lake ecosystem: a multi-proxy study from

Kharinei Lake, North Eastern European Russia. J. Paleolimn. 46, 123-137.

Kamatani, A., Oku, O., 2000. Measuring biogenic silica in marine sediments. Mar. Chem. 68, 219-229.

Karlsson, J., Christensen, T.R., Crill, P., Förster, J., Hammarlund, D., Jackowicz-Korczynski, M., Kokfelt, U., Roehm, C., Rosén,

P., 2010. Quantifying the relative importance of lake emissions in the carbon budget of a subarctic catchment. J. Geophys.

Res. 115.

Kokfelt, U., Rosén, P., Schoning, K., Christensen, T.R., Förster, J., Karlsson, J., Reuss, N., Rundgren, M., Callaghan, T.V.,

Jonasson, C., Hammarlund, D., 2009. Ecosystem responses to increased precipitation and permafrost decay in subarctic

Sweden inferred from peat and lake sediments. Glob. Change Biol. 15, 1652-1663.

Korsman, T., Nilsson, M., Ohman, J., Renberg, I., 1992. Near-infred reflectance spectroscopy of sediments - A potential method

to infer the past pH of lakes. Environ. Sci. Technol. 26, 2122-2126.

Korsman, T., Nilsson, M.B., Landgren, K., Renberg, I., 1999. Spatial variability in surface sediment composition characterised by

near-infrared (NIR) reflectance spectroscopy. J. Paleolimn. 21, 61-71.

Leng, M.J., Barker, P.A., 2006. A review of the oxygen isotope composition of lacustrine diatom silica for palaeoclimate

reconstruction. Earth-Sci. Rev. 75, 5-27.

Leng, M.J., Swann, G.E.A., Hodson, M.J., Tyler, J.J., Patwardhan, S.V., Sloane, H.J., 2009. The potential use of silicon isotope

composition of biogenic silica as a proxy for environmental change. Silicon 1, 65-77.

Malley, D., Lockhart, L., Wilkinson, P., Hauser, B., 2000. Determination of carbon, carbonate, nitrogen, and phosphorus in

freshwater sediments by near-infrared reflectance spectroscopy: rapid analysis and a check on conventional analytical

methods. J. Paleolimn. 24, 415-425.

Malley, D., Rönicke, H., Findlay, D., Surname, F., 1999. Feasibility of using near-infrared reflectance spectroscopy for the

analysis of C, N, P, and diatoms in lake sediments. J. Paleolimn. 21, 295-306.

McClure, W.F., 1994. Near-infrared spectroscopy - The giant is running strong. Anal. Chem. 66, A43-A53.

McKay, R., Naish, T., Carter, L., Riesselman, C., Dunbar, R., Sjunneskog, C., Winter, D., Sangiorgi, F., Warren, C., Pagani, M.,

Schouten, S., Willmott, V., Levy, R., DeConto, R., Powell, R.D., 2012. Antarctic and Southern Ocean influences on Late

Pliocene global cooling. Proc. Natl. Acad. Sci. U S A 109, 6423-6428.

22

Mecozzi, M., Pietrantonio, E., 2006. Carbohydrates proteins and lipids in fulvic and humic acids of sediments and its

relationships with mucilaginous aggregates in the Italian seas. Mar. Chem. 101, 27-39.

Mecozzi, M., Pietrantonio, E., Amici, M., Romanelli, G., 2001. Determination of carbonate in marine solid samples by FTIR-ATR

spectroscopy. The Analyst 126, 144-146.

Melles, M., Brigham-Grette, J., Glushkova, O.Y., Minyuk, P.S., Nowaczyk, N.R., Hubberten, H.-W., 2006. Sedimentary

geochemistry of core PG1351 from Lake El’gygytgyn—a sensitive record of climate variability in the East Siberian Arctic

during the past three glacial–interglacial cycles. J. Paleolimn. 37, 89-104.

Melles, M., Brigham-Grette, J., Minyuk, P.S., Nowaczyk, N.R., Wennrich, V., DeConto, R.M., Anderson, P.M., Andreev, A.A.,

Coletti, A., Cook, T.L., Haltia-Hovi, E., Kukkonen, M., Lozhkin, A.V., Rosén, P., Tarasov, P., Vogel, H., Wagner, B., 2012. 2.8

million years of Arctic climate change from Lake El'gygytgyn, NE Russia. Science 337, 315-320.

Meyers, P.A., Ishiwatari, R., 1993. Lacustrine organic geochemistry--an overview of indicators of organic matter sources and

diagenesis in lake sediments. Org. Geochem. 20, 867-900.

Michelutti, N., 2005. Recent primary production increases in arctic lakes. Geophys. Res. Lett. 32.

Michelutti, N., Blais, J.M., Cumming, B.F., Paterson, A.M., Rühland, K., Wolfe, A.P., Smol, J.P., 2010. Do spectrally inferred

determinations of chlorophyll a reflect trends in lake trophic status? J. Paleolimn. 43, 205-217.

Monteith, D.T., Stoddard, J.L., Evans, C.D., de Wit, H.A., Forsius, M., Høgåsen, T., Wilander, A., Skjelkvåle, B.L., Jeffries, D.S.,

Vuorenmaa, J., Keller, B., Kopácek, J., Vesely, J., 2007. Dissolved organic carbon trends resulting from changes in

atmospheric deposition chemistry. Nature 450, 537-U539.

Movasaghi, Z., Rehman, S., Rehman, I.U., 2008. Fourier transform infrared (FTIR) spectroscopy of biological tissues. Appl.

Spectrosc. Rev. 43, 134-179.

Nilsson, M.B., Dåbakk, E., Korsman, T., Renberg, I., 1996. Quantifying relationships between near-infrared reflectance spectra

of lake sediments and water chemistry. Environ. Sci. Technol. 30, 2586-2590.

Nowaczyk, N.R., Haltia, E.M., Ulbricht, D., Wennrich, V., Sauerbrey, M.A., Rosén, P., Vogel, H., Francke, A., Meyer-Jacob, C.,

Andreev, A.A., Lozhkin, A.V., 2013. Chronology of Lake El'gygytgyn sediments – a combined magnetostratigraphic,

palaeoclimatic and orbital tuning study based on multi-parameter analyses. Clim. Past 9, 2413-2432.

Ohlendorf, C., Sturm, M., 2008. A modified method for biogenic silica determination. J. Paleolimn. 39, 137-142.

Patwardhan, S.V., Maheshwari, R., Mukherjee, N., Kiick, K.L., Clarson, S.J., 2006. Conformation and assembly of polypeptide

scaffolds in templating the synthesis of silica: an example of a polylysine macromolecular “switch”. Biomacromolecules 7,

491-497.

Rein, B., Sirocko, F., 2002. In-situ reflectance spectroscopy - analysing techniques for high-resolution pigment logging in

sediment cores. Int. J. Earth-Sci. 91, 950-954.

Renberg, I., 1990. A 12600 year perspective of the acidification of Lilla Oresjon, southwest Sweden. Philos. Trans. R. Soc. Lond.

Ser. B-Biol. Sci. 327, 357-361.

Renberg, I., Hellberg, T., 1982. The pH history of lakes in southwestern Sweden, as calculated from the subfossil diatom flora of

the sediments. Ambio 11, 30-33.

Renberg, I., Korsman, T., Birks, H.J.B., 1993. Prehistoric increases in the pH of acid-sensitive Swedish lakes caused by land-use

changes. Nature 362, 824-827.

Reuss, N.S., Hammarlund, D., Rundgren, M., Segerström, U., Eriksson, L., Rosén, P., 2010. Lake Ecosystem Responses to

Holocene Climate Change at the Subarctic Tree-Line in Northern Sweden. Ecosystems 13, 393-409.

Rosén, P., 2005. Total organic carbon (TOC) of lake water during the Holocene inferred from lake sediments and near-infrared

spectroscopy (NIRS) in eight lakes from northern Sweden. Biogeochemistry 76, 503-516.

Rosén, P., Bindler, R., Korsman, T., Mighall, T., Bishop, K., 2011a. The complementary power of pH and lake-water organic

carbon reconstructions for discerning the influences on surface waters across decadal to millennial time scales.

Biogeosciences 8, 2717-2727.

Rosén, P., Cunningham, L., Vonk, J., Karlsson, J., 2009. Effects of climate on organic carbon and the ratio of planktonic to

benthic primary producers in a subarctic lake during the past 45 years. Limnol. Oceanogr. 54, 1723-1732.

Rosén, P., Dåbakk, E., Renberg, I., Nilsson, M., Hall, R., 2000. Near-infrared spectrometry (NIRS): a new tool for inferring past

climatic changes from lake sediments. The Holocene 10, 161.

Rosén, P., Hammarlund, D., 2007. Effects of climate, fire and vegetation development on Holocene changes in total organic

carbon concentration in three boreal forest lakes in northern Sweden. Biogeosciences 4, 975-984.

Rosén, P., Vogel, H., Cunningham, L., Hahn, A., Hausmann, S., Pienitz, R., Zolitschka, B., Wagner, B., Persson, P., 2011b.

Universally applicable model for the quantitative determination of lake sediment composition using Fourier transform

infrared spectroscopy. Environ Sci Technol 45, 8858-8865.

Rosén, P., Vogel, H., Cunningham, L., Reuss, N., Conley, D.J., Persson, P., 2010. Fourier transform infrared spectroscopy, a new

method for rapid determination of total organic and inorganic carbon and biogenic silica concentration in lake sediments. J.

Paleolimn. 43, 247-259.

23

Rouillard, A., Michelutti, N., Rosén, P., Douglas, M.S.V., Smol, J.P., 2012. Using paleolimnology to track Holocene climate

fluctuations and aquatic ontogeny in poorly buffered High Arctic lakes. Palaeogeogr. Palaeoecol. 321, 1-15.

Rouillard, A., Rosén, P., Douglas, M.S.V., Pienitz, R., Smol, J.P., 2011. A model for inferring dissolved organic carbon (DOC) in

lakewater from visible-near-infrared spectroscopy (VNIRS) measures in lake sediment. J. Paleolimn. 46, 187-202.

Rubensdotter, L., Rosqvist, G., 2003. The effect of geomorphological setting on Holocene lake sediment variability, northern

Swedish Lapland. J. Quat. Sci. 18, 757-767.

Rydberg, J., 2014. Wavelength dispersive X-ray fluorescence spectroscopy as a fast, non-destructive and cost-effective analytical

method for determining the geochemical composition of small loose-powder sediment samples. J. Paleolimn. 52, 265-276.

Rydberg, J., Klaminder, J., Rosén, P., Bindler, R., 2010. Climate driven release of carbon and mercury from permafrost mires

increases mercury loading to sub-arctic lakes. Sci Total Environ 408, 4778-4783.

Rydberg, J., Rosén, P., Lambertsson, L., De Vleeschouwer, F., Tomasdotter, S., Bindler, R., 2012. Assessment of the spatial

distributions of total- and methyl-mercury and their relationship to sediment geochemistry from a whole-lake perspective. J.

Geophys. Res.-Biogeosci. 117.

Schmidt, M., Botz, R., Rickert, D., Bohrmann, G., Hall, S.R., Mann, S., 2001. Oxygen isotopes of marine diatoms and relations to

opal-A maturation. Geochim. Cosmochim. Acta 65, 201-211.

Segerström, U., Emanuelsson, M., 2002. Extensive forest grazing and hay-making on mires - vegetation changes in south-

central Sweden due to land use since Medieval times. Veg. Hist. Archaeobot. 11, 181-190.

Segerström, U., Hornberg, G., Bradshaw, R., 1996. The 9000-year history of vegetation development and disturbance patterns

of a swamp-forest in Dalarna, northern Sweden. Holocene 6, 37-48.

Siesler H.W., Ozaki Y., Kawata S. and Heise H.M. 2002. Near-infrared spectroscopy: principles, instruments, applications.

Wiley, Weinheim.

Sifeddine, A., Fröhlich, F., Fournier, M., Martin, L., Servant, M., Soubiès, F., Turcq, B., Suguio, K., Volkmer-Ribeiro, C., 1994. La

sédimentation lacustre indicateur de changements des paléoenvironnements au cours des 30 000 dernières années (Carajas,

Amazonie, Brésil). C.R. Acad. Sci. Paris 318, 1645-1652.

Skjelkvåle, B.L., Stoddard, J.L., Jeffries, D.S., Tørseth, K., Høgåsen, T., Bowman, J., Mannio, J., Monteith, D.T., Mosello, R.,

Rogora, M., Rzychon, D., Vesely, J., Wieting, J., Wilander, A., Worsztynowicz, A., 2005. Regional scale evidence for

improvements in surface water chemistry 1990-2001. Environ. Pollut. 137, 165-176.

Smol, J.P., 1992. Paleolimnology: an important tool for effective ecosystem management. J. Aquat. Ecosys. Health 1, 49-58.

Smol, J.P., Douglas, M.S.V., 2007. Crossing the final ecological threshold in high Arctic ponds. Proc. Natl. Acad. Sci. U S A 104,

12395-12397.

Smol, J.P., Wolfe, A.P., Birks, H.J., Douglas, M.S., Jones, V.J., Korhola, A., Pienitz, R., Rühland, K., Sorvari, S., Antoniades, D.,

Brooks, S.J., Fallu, M.A., Hughes, M., Keatley, B.E., Laing, T.E., Michelutti, N., Nazarova, L., Nyman, M., Paterson, A.M.,

Perren, B., Quinlan, R., Rautio, M., Saulnier-Talbot, E., Siitonen, S., Solovieva, N., Weckström, J., 2005. Climate-driven

regime shifts in the biological communities of arctic lakes. Proc. Natl. Acad. Sci. U S A 102, 4397-4402.

Snowball, I., Sandgren, P., 1996. Lake sediment studies of Holocene glacial activity in the Karsa valley, northern Sweden:

Contrasts in interpretation. Holocene 6, 367-372.

Sprenk, D., Weber, M.E., Kuhn, G., Rosén, P., Frank, M., Molina-Kescher, M., Liebetrau, V., Röhling, H.-G., 2013. Southern

Ocean bioproductivity during the last glacial cycle – new detection method and decadal-scale insight from the Scotia Sea.

Geological Society, London, Special Publications 381, 245-261.

Tolu, J., Gerber, L., Boily, J.F., Bindler, R., 2015. High-throughput characterization of sediment organic matter by pyrolysis-gas

chromatography/mass spectrometry and multivariate curve resolution: A promising analytical tool in (paleo)limnology.

Anal. Chim. Acta 880, 93-102.

Trachsel, M., Grosjean, M., Schnyder, D., Kamenik, C., Rein, B., 2010. Scanning reflectance spectroscopy (380–730 nm): a novel

method for quantitative high-resolution climate reconstructions from minerogenic lake sediments. J. Paleolimn. 44, 979-

994.

Tranvik, L.J., Jansson, M., 2002. Climate change - Terrestrial export of organic carbon. Nature 415, 861-862.

Tuvendal, M., Elmqvist, T., 2011. Ecosystem services linking social and ecological systems: River brownification and the

response of downstream stakeholders. Ecol. Soc. 16, 21.

Valinia, S., Futter, M.N., Cosby, B.J., Rosén, P., Fölster, J., 2014. Simple models to estimate historical and recent changes of

total organic carbon concentrations in lakes. Environ. Sci. Technol. 49, 386-394.

Wirrmann, D., Bertaux, J., Kossoni, A., 2001. Late Holocene paleoclimatic changes in Western Central Africa inferred from

mineral abundance in dated sediments from Lake Ossa (Southwest Cameroon). Quat. Res. 56, 275-287.

Vogel, H., Meyer-Jacob, C., Melles, M., Brigham-Grette, J., Andreev, A.A., Wennrich, V., Tarasov, P.E., Rosén, P., 2013. Detailed

insight into Arctic climatic variability during MIS 11c at Lake El'gygytgyn, NE Russia. Clim. Past 9, 1467-1479.

Vogel, H., Rosén, P., Wagner, B., Melles, M., Persson, P., 2008. Fourier transform infrared spectroscopy, a new cost-effective

tool for quantitative analysis of biogeochemical properties in long sediment records. J. Paleolimn. 40, 689-702.

24

Vogel, H., Wagner, B., Zanchetta, G., Sulpizio, R., Rosén, P., 2010. A paleoclimate record with tephrochronological age control

for the last glacial-interglacial cycle from Lake Ohrid, Albania and Macedonia. J. Paleolimn. 44, 295-310.

Wold, S., Sjöström, M., Eriksson, L., 2001. PLS-regression: a basic tool of chemometrics. Chemometr Intell Lab Syst 58, 109-

130.

Wolfe, A.P., Vinebrooke, R.D., Michelutti, N., Rivard, B., Das, B., 2006. Experimental calibration of lake-sediment spectral

reflectance to chlorophyll a concentrations: methodology and paleolimnological validation. J. Paleolimn. 36, 91-100.

Worrall, F., Harriman, R., Evans, C.D., Watts, C.D., Adamson, J., Neal, C., Tipping, E., Burt, T., Grieve, I., Monteith, D., Naden,

P.S., Nisbet, T., Reynolds, B., Stevens, P., 2004. Trends in dissolved organic carbon in UK rivers and lakes. Biogeochemistry

70, 369-402.

Yallop, A.R., Clutterbuck, B., Thacker, J., 2010. Increases in humic dissolved organic carbon export from upland peat

catchments: the role of temperature, declining sulphur deposition and changes in land management. Clim. Res. 45, 43-56.

Zolitschka, B., Mingram, J., Van Der Gaast, S., Jansen, J.F., Naumann, R., 2001. Sediment logging techniques, In: Last, W.M.,

Smol, J.P., (Eds.), Tracking Environmental Change Using Lake Sediments, vol 1, Basin Analysis, Coring, and Chronological

Techniques. Kluwer Acad., Dordrecht, Netherlands, pp. 137-153.

25

Acknowledgments

There are many people to whom I owe gratitude for contributing to this research as well as for making the past four years such a valuable and positive experience. First of all, I want to thank my supervisor Rich. Thank you for the great support and guidance, for being there when needed, for always taking the time to discuss science but also other things in life. Special thanks to my co-supervisor, Hendrik. Thank you for your enthusiasm, motivation, and support, but most of all for encouraging me six years ago to go to Abisko for an internship with Peter; this was the start of a great adventure. I would like to thank Peter for sharing his passion and enthusiasm for science and nature, for giving me the opportunity to come to Umeå. I thank all my co-authors for their comments, ideas, and contributions to the manuscripts that are included in this thesis. Thank you, Catalina, Volker, Martin, Florian, Michael, Christian, Melanie, Julie, Handong, Peter, Hendrik, and Rich.

I want to thank Jon, Dominque, and Sophia for all the good times, for helping whenever help was needed, and most of all for putting up with my quirks. I’m glad we spent this time together.

I am grateful to current and former members of the “paleo” group in Umeå, namely Julie, Johan, Carolina, Christian, Jonatan, Erik, Åsa, Manuela, Sofia, Marina, Ingemar, and Erik, for help, fruitful discussions, and creating a good spirit at work. Thank you to all Wednesday innebandy players at EMG for the good times on the court; to Ingrid, Jolina, Mattias, and all other administrators at EMG; and to Carin for being such a good-spirited help in the lab.

And I want to thank my family. Mama and Papa, thank you for supporting me in all different ways, for believing in me; without you none of this would have been possible!