ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2016

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1420

Spatiotemporal streamflowvariability in a boreal landscape

Importance of landscape composition for catchmenthydrological functioning

REINERT HUSEBY KARLSEN

ISSN 1651-6214ISBN 978-91-554-9680-7urn:nbn:se:uu:diva-302400

Dissertation presented at Uppsala University to be publicly examined in Hambergsalen,Villavägen 16, Uppsala, Friday, 21 October 2016 at 10:00 for the degree of Doctor ofPhilosophy. The examination will be conducted in English. Faculty examiner: Dr Sarah E.Godsey (Idaho State University ).

AbstractKarlsen, R. H. 2016. Spatiotemporal streamflow variability in a boreal landscape. Importanceof landscape composition for catchment hydrological functioning. (Avrinningens rumsliga ochtidsmässiga variation i ett borealt landskap. Landskapets betydelse för avrinningsområdetshydrologiska funktion). Digital Comprehensive Summaries of Uppsala Dissertations fromthe Faculty of Science and Technology 1420. 64 pp. Uppsala: Acta Universitatis Upsaliensis.ISBN 978-91-554-9680-7.

The understanding of how different parts of a landscape contribute to streamflow by storingand releasing water has long been a central issue in hydrology. Knowledge about what controlsstreamflow dynamics across landscapes can further our understanding of how catchments storeand release water, facilitate predictions for ungauged catchments, and improve the managementof water quality and resources. This thesis makes use of data from the Krycklan catchment innorthern Sweden. Streamflow data from 14 catchments (0.12 - 68 km2) with variable landscapecharacteristics such as topography, vegetation, wetland cover, glacial till soils and deepersediment soils were used to investigate spatial patterns and controls on runoff.

The differences in specific discharge (discharge per unit catchment area) between nearbycatchments were large at the annual scale, and have the same magnitude as predicted effects ofa century of climate change or the observed effects of major forestry operations. This variabilityis important to consider when studying the effects of climate change and land use changeson streamflow, as well as for our understanding of geochemical mass balances. Streamflowfrom different catchments was strongly related to landscape characteristics. The distribution ofwetland areas had a particularly strong influence, with an annual specific discharge 40-80%higher than catchments with high tree volume on till soils. During drier periods, catchments withdeeper sediment soils at the lower elevations of Krycklan had a higher base flow compared toboth forested till and wetland catchments. This pattern was reversed at high flows. The storagesreleasing water to streams in downstream sediment areas were able to maintain base flow forlonger periods and were less influenced by evapotranspiration compared to the more superficialtill and wetland systems.

The results of this thesis have led to a better understanding of the landscape wide patternsof streamflow during different seasons and time scales. The strong associations to landscapecharacteristics and variable spatial patterns with season and antecedent conditions form the basisfor a conceptual understanding of the processes and spatial patterns that shape the heterogeneityof streamflow responses in boreal catchments.

Keywords: streamflow, catchment hydrology, boreal, water balance, spatiotemporalvariability, landscape analysis, climate change, recession curve

Reinert Huseby Karlsen, Department of Earth Sciences, LUVAL, Villav. 16, UppsalaUniversity, SE-75236 Uppsala, Sweden.

© Reinert Huseby Karlsen 2016

ISSN 1651-6214ISBN 978-91-554-9680-7urn:nbn:se:uu:diva-302400 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-302400)

Akademisk avhandling som för avläggande av filosofie doktorsexamen i hydrologi vid Upp-sala universitet kommer att offentligen försvaras i Hambergsalen, Villavägen 16, Uppsala, fredagen 21 oktober 2016, klockan 10:00. Fakultetsopponent: Sarah E. Godsey (Idaho State University). Disputationen sker på engelska. Referat Karlsen, R. H. 2016. Avrinningens rumsliga och tidsmässiga variation i ett borealt landskap. Landskapets betydelse för avrinningsområdets hydrologiska funktion. Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 1420. Upp-sala: Acta Universitatis Upsaliensis. ISBN 978-91-554-9680-7. Hur olika delar av landskapet påverkar vattenbalansen och bidrar till avrinning har länge varit en central fråga inom hydrologin. Kunskap om vad som styr avrinningsdynamiken i ett land-skap kan öka vår förståelse av hur olika delar av landskapet bidrar till avrinning, hur avrin-ningsområden lagrar vatten och bildar avrinning, underlätta prognoser för avrinningsområden utan vattenföringsmätningar och förbättra hanteringen av vattenkvaliteten och vattenresurser. Denna avhandling använder data från Krycklans avrinningsområde i norra Sverige. Vattenfö-ringsdata från 14 delavrinningsområden (0.12 - 68 km2) med olika landskapskarakteristik såsom topografi, vegetation och jordarter, användes för att undersöka rumsliga mönster hos avrinningen över olika tidsperioder samt hur landskapet påverkar variabiliteten.

Skillnaderna i specifik avrinning (avrinning per areaenhet) mellan närliggande avrinnings-områden var stor för årliga värden, och är i samma storleksordning som effekterna av stora skogsavverkningar samt av förutspådda effekter av det kommande seklets förväntade klimat-förändringar. Denna variation är viktig att ta hänsyn till när man studerar hur klimatföränd-ringar och ändrad markanvändning påverkar avrinningen, liksom för vår förståelse av geoke-miska massbalanser. Avrinning från olika områden var starkt relaterad till deras landskaps-egenskaper. Förekomsten av våtmarker hade ett särskilt starkt inflytande. Områden med en stor andel våtmarker hade 40-80% högre årlig specifik avrinning än områden med hög trädvo-lym på moränjordar. Under torrare perioder hade områden med djupare sedimentjordar hög avrinning jämfört med både områden med skog på morän och med våtmarker. Under höga flöden var detta mönster omvänt. De vattenlager som bidrar till avrinning i sedimentområden kan upprätthålla basflöde under längre tidsperioder och påverkas mindre av evapotranspirat-ionen än de ytligare flödessystemen i morän och våtmarker.

Avhandlingen har givit en bättre förståelse av avrinningens rumsliga variation under olika årstider och i olika tidsskalor. Det starka sambandet mellan landskapskarakteristik och avrin-ningens varierande mönster under olika årstider och lagringsförhållanden utgör en grund för en begreppsmässig förståelse av de processer och rumsliga mönster som skapar heterogenitet-en i flödesrespons i boreala områden. Nyckelord: avrinning, vattenföring, hydrologi, boreal, vattenbalans, spatiotemporal varabilitet, landskapsanalys, klimatförändringar, recession kurva Reinert Huseby Karlsen, Institutionen för geovetenskaper, LUVAL, Villav. 16, Uppsala Uni-versitet, 752 36 Uppsala. © Reinert Huseby Karlsen 2016 ISSN 1651-6214 ISBN 978-91-554-9680-7

Til min familie

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Karlsen, R.H., Seibert, J., Grabs, T., Laudon, H., Blomkvist,

P., Bishop, K. (2016) The assumption of uniform specific dis-charge: unsafe at any time? Hydrological Processes, Accepted, doi:10.1002/hyp.10877

II Karlsen, R.H., Grabs, T., Bishop, K., Buffam, I., Laudon, H., Seibert, J. (2016) Landscape controls on spatiotemporal dis-charge variability in a boreal catchment. Water Resources Re-search, Accepted, doi:10.1002/2016WR019186.

III Teutschbein, C., Grabs, T., Karlsen, R.H., Laudon, H., Bishop, K. (2015) Hydrological response to changing climate condi-tions: Spatial streamflow variability in the boreal region. Water Resources Research, 51(12): 9425–9446, doi:10.1002/2015WR017337.

IV Karlsen, R.H., Bishop, K., Grabs, T., Ottosson-Löfvenius, M., Laudon, H., Seibert, J. (2016) The role of catchment physiog-raphy, storage and evapotranspiration on variability in stream-flow recessions. Manuscript in preparation.

Reprints of paper I, II and III were made with permission from Wiley. For papers I, II and IV I was responsible for most data preparation, all anal-ysis and had the main responsibility for the writing. For paper III I was re-sponsible for the preparation of the hydrologic data, runoff modelling and contributed to the writing.

Contents

Introduction ................................................................................................... 11 Spatial and temporal variability ............................................................... 12 

Aims and objectives ...................................................................................... 15 

Study area ..................................................................................................... 16 The Krycklan catchment .......................................................................... 16 Hydrology of the Krycklan catchment ..................................................... 19 

Data and methods .......................................................................................... 22 Streamflow data ........................................................................................ 22 Climate data.............................................................................................. 23 Catchment landscape characteristics ........................................................ 24 Methods .................................................................................................... 24 

Temporal aggregation .......................................................................... 24 Analysis of spatial variability .............................................................. 26 Specific discharge uncertainty ............................................................. 26 Hydrological modelling ....................................................................... 26 Streamflow recession analysis ............................................................. 27 

Results ........................................................................................................... 30 Spatiotemporal variability of specific discharge ...................................... 30 Landscape controls on specific discharge variability ............................... 31 Spatial variation of climate predictions .................................................... 34 Spatiotemporal variation in streamflow recessions .................................. 36 

Discussion ..................................................................................................... 39 Patterns and controls on spatial variation in specific discharge ............... 39 

Landscape patterns of specific discharge ............................................. 40 Landscape variability in predicted climate change effects .................. 42 

Variability in streamflow recession across Krycklan ............................... 43 Hydrological functioning of the Krycklan landscape ............................... 44 

Conclusions ................................................................................................... 47 Future research ......................................................................................... 48 

Acknowledgements ....................................................................................... 49 

Sammanfattning på svenska (Summary in Swedish) .................................... 51 

References ..................................................................................................... 54 

11

Introduction

The understanding of how different parts of a landscape contribute to streamflow by storing and releasing water has long been a central issue in hydrology. Although it is widely recognized that landscape characteristics, such as topography, soils and vegetation, influence the partitioning of water into storage, evapotranspiration and streamflow, detailed knowledge about how the variability of the landscape influences the runoff regime and the processes responsible remains limited. Knowledge about what controls streamflow dynamics across landscapes can further our understanding of how catchments store and release water, facilitate predictions for ungauged catchments, and improve the management of water resources, nutrients and pollutants (Doyle et al., 2005; Pinay et al., 2015; Wagener et al., 2007).

Hydrological systems exhibit a high degree of complexity and heteroge-neity in processes and responses, both in space and time. A search for simple organizing principles that could classify catchments and their hydrological functioning has been suggested as a way to bring order to the vast heteroge-neity in the observational record (Dooge, 1986; McDonnell et al., 2007; McDonnell and Woods, 2004; Sivapalan, 2005). A way towards this likely involves analyzing landscape heterogeneity in hydrological response to find concepts which can describe the variability and the patterns we observe (Hrachowitz et al., 2013). By comparing and classifying catchments, one can seek connections between hydrological functioning in different landscapes and hydro-climatic conditions, with the ultimate goal of understanding how catchment structure and climate interact to define catchment function in the transformation of precipitation into streamflow (Wagener et al., 2007).

Knowledge of such connections could eventually lead us to better under-stand the functional traits of catchments and form a basis for better model predictions (Blöschl et al., 2013; McDonnell et al., 2007; Sivapalan et al., 2003). Wagener et al. (2007) have suggested a framework that considers catchment hydrological functioning, such as the partitioning, storage and release of water in catchments, across spatial and temporal scales. Stream-flow, for which observations are widely available, integrates the processes occurring within a catchment, and its characteristics can therefore be a prom-ising basis for such analysis (e.g. Kirchner, 2009).

12

The boreal zone, containing 33% of the world’s forested area, holds about one third of the global soil carbon stock (Dixon et al., 1994; Humphreys et al., 2006; Scharlemann et al., 2014). The boreal zone plays an important part in the global carbon balance, and the boreal surface waters play a key role for the total carbon export from the boreal landscape (Tranvik et al., 2009; Wallin et al., 2013). Land use changes through forestry operations and ex-tensive drainage are common in many parts of the boreal landscape, espe-cially in Sweden. Additionally, predicted climate change is also expected to exert a large influence in northern latitudes. Thus, improved knowledge of the hydrology across the boreal landscape is crucial for the management of water resources and water quality. However, the hydrological role of differ-ent landscape elements in boreal catchments and their contributions to streamflow, as well as their interaction, remains poorly understood.

Spatial and temporal variability Hydrological systems display significant variability in space and time. Con-trolling factors such as weather and climate, geology, soils, vegetation, to-pography and human influences are often associated with this variability (Woods, 2005). Studying the controls on spatiotemporal variability of hydro-logical response and processes is not only important for understanding how the various elements of the landscape function independently of each other, but also how they interact. This can be especially important in the case of the boreal landscape which consists of a mosaic of landscape elements including forests, wetlands and lakes, since the interactions between these landscape elements in creating the runoff response in streams remains largely un-known.

Exploring this variability in a landscape perspective may be valuable for understanding the partitioning of water into storage, evapotranspiration and runoff across spatial and temporal scales. Great efforts in hydrological re-search have been directed towards understanding the consequences of alter-ing the landscape (e.g. forestry operations) or of a changing climate. Howev-er, one could argue that to properly predict how perturbations might influ-ence our future we first need to understand the magnitudes and dynamics of the present variability. Understanding the variability is also important for the development of models, especially with regards to the conceptualization of processes, regionalization and predictions in ungauged basins (Sivapalan, 2005).

Investigations of how landscape structure influences variability in stream runoff has long been a central effort in hydrology (e.g. Hewlett and Hibbert, 1967; Hoover and Hursh, 1943). Landscape characteristics have been found

13

to control not only streamflow magnitudes (Kuraś et al., 2008; Payn et al., 2009; van der Velde et al., 2013), but also transit times (McGuire et al., 2005; Soulsby et al., 2006; Tetzlaff et al., 2009), water storage (Sayama et al., 2011), streamflow recession (Bogaart et al., 2015; Tague and Grant, 2004) and runoff response (Dunne, 1983; Nippgen et al., 2011).

Considerable spatial variability in area specific discharge has been ob-served within a relatively small spatial scale (< 100 km2) in boreal catch-ments, both at the daily scale using snapshots (Lyon et al., 2012; Temnerud et al., 2007) and annual scale (Buttle and Eimers, 2009; Nicolson, 1988; Prepas et al., 2006). In some cases, a relation between spatial patterns of streamflow and landscape characteristics has been found, such as specific discharge being positively related to wetland areas (Lyon et al., 2012; Prepas et al., 2006).

It can be argued, however, that it is important to evaluate variability across the complete range of flow rates and under different antecedent or seasonal conditions. For example, ecological and biological effects of streamflow variability can manifest themselves at various time scales (Doyle et al., 2005; McClain et al., 2003; Tetzlaff et al., 2005). Different hydrologi-cal processes can also dominate, or become apparent, at different spatial and temporal scales (Blöschl and Sivapalan, 1995). Furthermore, the temporal and spatial scale of analysis has a large impact on which aspects of the hy-drological variability we observe (Woods, 2005). Atkinson et al. (2003), for instance, found that different complexities in models was needed to predict spatial patterns of streamflow under different conditions, and that a simple lumped model was adequate during wet periods but a more complex and distributed model was needed during dry periods. Similarly, Eder et al. (2003) found that increasing model complexity was needed to capture the variability at shorter (daily) temporal scales than at longer (seasonal and annual) temporal scales. They note that although the physical processes driv-ing the hydrological system might not change with timescale, the dominance of difference processes can depend on the timescale of analysis and variables such as antecedent wetness conditions.

Such dynamic (i.e. temporally variable) controls on spatial patterns have been observed for streamflow (Kirnbauer et al., 2005; Kuraś et al., 2008; Lyon et al., 2012; Payn et al., 2012), recession characteristics (Krakauer and Temimi, 2011), hydrological connectivity (Jencso and McGlynn, 2011), soil moisture (Grayson et al., 1997) and biogeochemistry (Ågren et al., 2014; Buffam et al., 2007). For example, Payn et al. (2012) found that topography had a decreasing influence on spatial patterns of base flow as streamflow gradually decreased, while subsurface characteristics likely gained influence. Hence, knowledge on dynamic spatial patterns and the associated controls

14

can give further information on the hydrological functioning of different parts of a landscape under different wetness conditions.

However, many of these studies on spatial variability of streamflow char-acteristics and association with landscape and climate have focused on coarse temporal aggregation such as annual flows, spatial snapshot cam-paigns or a limited period of the year (e.g. spring or base flow only), or on large scales where climate variability has a significant effect. There is thus a lack of characterization of variability of streamflow and its relation to land-scape characteristics across temporal scales and within a relatively small landscape subject to the same climate. Therefore, this thesis aims to charac-terize spatial and temporal variability in a boreal landscape to help close some of the current knowledge gaps.

15

Aims and objectives

The primary aim of this thesis was to exploit the spatial variability in stream-flow across a boreal landscape to improve the understanding of what con-trols that variability, and gain insight into the hydrological functioning of the landscape.

The specific objectives were I Quantify the spatial variability of specific discharge across time-

scales in a boreal landscape (paper I) II Identify the major factors controlling the spatiotemporal variability

in specific discharge and storage-discharge behavior (papers II, IV) III Predict the spatial variability in the runoff regime under the influence

of climate change (paper III)

16

Study area

The Krycklan catchment This thesis is based on data collected between 2008 and 2013 within the Krycklan Catchment Study (KCS, Laudon et al., 2013) in northern Sweden (64°25' N, 19°46' E), situated about 50 km west of the Baltic Sea coast (Fig-ure 1). The KCS consists of a 68 km2 catchment, with streamflow monitored for the main outlet and 13 subcatchments.

Figure 1. The Krycklan Catchment Study location (left), and detailed map (right) of the catchment outlets, distribution of soils, wetlands and lakes.

The climate is characterized by long winters and short summers. Mean an-nual temperature (1981-2010) is 1.8°C, with mean monthly temperatures in January of -9.5°C and +14.7°C in July. The mean annual precipitation is 614 mm, with about one third falling as snow and the 30-year average snow wa-ter equivalent is 180 mm (Laudon and Ottosson Löfvenius, 2016). The snow cover typically forms in November, with snow melt beginning in April. Po-

17

tential evapotranspiration (PET) shows a strong seasonal pattern, and usually peaks in June-July with a mean annual PET of 480 mm. This climatic pattern is reflected in the monthly streamflow distribution over the year, with spring snow melt producing the highest streamflow rates in April-May, which gen-erally decline during the early summer months, then rise again in the autumn before declining at the onset of sub-zero temperatures (Figure 2). About 40 to 50% of the total annual streamflow occurs during April and May (1981-2013).

Figure 2. Mean monthly a) temperature, b) precipitation, c) potential evapotranspira-tion (PET) and d) specific discharge at C7 for the study period 2008-2013. Gray shaded area shows the long term (1981-2013) range in mean ± standard deviation.

The bedrock in KCS is fairly uniform, with 94% Svecofennian metasedi-ments/metagraywacke, smaller amounts of acid and intermediate (4%) and basic (3%) metavolcanic rock. The terrain is gently undulating, with eleva-tions ranging from 127 to 372 m a.s.l. The upper parts of the catchment are dominated by Quaternary deposits of glacial till, while the material on the lower parts of the catchments is mainly postglacial sediment deposits of silt

18

and sand (hereafter referred to as sediment). Forests cover most of the land surface of the catchment (87%), with peat forming wetlands (9%), mainly at the upper altitudes, lakes (1%), rock outcrops (1%) and arable land (2%) covering the remaining area. Stands of Scots pine (Pinus sylvestris) make up 63% of the total forest volume and are commonly found on the drier upslope areas. Norway spruce (Picea abies, 26%) and birch (Betula spp., 10%) are found mainly in the lower lying, wetter areas along stream channels. The mean stand age is 62 years.

Table 1. A selection of the catchment characteristics considered in this thesis. See Table 3 for details.

Area Elevation Forest Mire Lake Till and

thin soils Sediment

soils Tree

volume

Catchment (ha) (m a.s.l.) (%) (%) (%) (%) (%) (m3 ha-1)

C1 48 279 98 2 0 100 0 187

C2 12 273 100 0 0 100 0 212

C4 18 287 56 44 0 49 0 83

C5 65 292 54 40 6 46 0 64

C6 110 283 71 25 4 65 0 117

C7 47 275 82 18 0 81 0 167

C9 288 251 84 14 2 76 4 150

C10 336 296 74 26 0 71 1 93

C12 544 277 83 17 0 75 6 129

C13 700 251 88 10 1 70 16 145

C14 1410 228 90 5 1 53 38 106

C15 1913 277 82 15 2 73 10 85

C16 6790 239 87 9 1 58 30 106

C20 145 214 88 10 0 65 21 59

Figure 3. Catchment maps showing a) the subcatchments, b) elevation and c) tree volume.

For this thesis the entire catchment (C16) and 13 subcatchments of Krycklan have been considered (Table 1, Figure 3). These catchments vary in size (12

19

to 6790 ha), portion of catchment covered by wetlands (0 to 44 %), till and thin soils (46 to 100%), sediment soils (0 to 38%) and forest tree volume (59 to 212 m3 ha-1).

Hydrology of the Krycklan catchment Forest hydrology has a long history in Sweden. Hydrological studies on the effects of forestry operations date back to the early 20th century (Grip, 2015). In the Krycklan area, forest hydrological research began about 100 years ago, and the Svartberget research catchment (C7 in this thesis) was established in 1979 with a focus on forest hydrology and geochemistry (Grip and Bishop, 1990; Laudon et al., 2013). The research later expanded to the 68km2 KCS in 2002 with increased focus on process-based research at the landscape scale.

Knowledge about different runoff generation mechanisms is important when studying the spatial and temporal patterns of stream runoff. These mechanisms and their interactions structure the patterns in stream runoff that we observe. I will therefore briefly summarize some previous hydrological research in Krycklan and similar environments concerning the dominant landscape types found in Krycklan, namely glacial till, wetlands and sedi-ment soils.

Much of the hydrological research in Krycklan, and elsewhere in Sweden, has focused on the role of glacial till soils, which dominate in areal extent. Ground-breaking work by Rodhe (1989, 1987, 1981), using oxygen-18 iso-tope as a tracer in Krycklan and elsewhere in Sweden, showed that storm runoff resulting from both snowmelt and rainfall consisted mainly of pre-event groundwater. Lundin (1982) demonstrated that saturated hydraulic conductivity in till soils decreased quickly with depth from the surface, and that raising the groundwater table into the more transmissive, near surface layers caused a large increase in stream discharge (Rodhe, 1989). Bishop (1991) explored the implications of these findings for stream chemistry in the Svartberget catchment, and proposed the term “transmissivity feedback” for this runoff generation mechanism. It has since been identified as the dominant runoff mechanism in a number of forested till soils (Bishop et al., 2011; Bishop, 1991; Grip and Rodhe, 1985; Laudon et al., 2004; Nyberg, 1995; Rodhe, 1989) and the explanation for how pre-event groundwater can be rapidly mobilized during runoff events (Bishop et al., 2004). Despite deep soil frost in the till soils during winter, this has not been found to significant-ly influence streamflow response in Swedish till soils (Nyberg et al., 2001).

When it comes to the hydrology of wetlands, Rodhe (1987) performed hydrograph separation on a wetland influenced catchment in Krycklan dur-

20

ing snowmelt (C4 in this thesis). He found a higher event water contribution in the wetland influenced catchment compared to the till areas. Similarly, Laudon et al. (2007), also using oxygen-18 hydrograph separation, found much higher event water contribution from wetlands, and that percentage of wetland cover could largely explain the differences in event water contribu-tion between 15 subcatchments in Krycklan during snowmelt. This was also supported by Lyon et al. (2010), who observed shorter mean transit times for wetland dominated catchments during snowmelt in Krycklan. Similar find-ings have been found elsewhere in the boreal region (Gibson et al., 1993). Peralta-Tapia et al. (2015b) investigated isotopic signatures in central Kryck-lan and found similar responses during spring snow melt and rain events for a wetland. The mechanisms responsible for the larger event water contribu-tions from wetlands were concluded to be rapid routing of overland flow on the frozen mire surface, saturation overland flow and preferential flow paths within the mire.

There has been less focus on the sediment soils in the lower areas of the Krycklan catchment. Generally, these types of sediment deposits represent larger reservoirs, with deeper groundwater tables showing less fluctuation over the year (Knutsson and Fagerlind, 1977). For other study areas, it has been suggested that sand and gravel deposits in the riparian zones could function as a hydrologic buffer maintaining higher base flows (Gaál et al., 2012; Shaman et al., 2004). The sediment deposits in Krycklan are associat-ed with deeper soils and higher permeability compared to the more shallow and compacted glacial till in the upland areas (Bishop, 1991; Mellander et al., 2004).

On the scale of the whole Krycklan catchment, the portion of wetland ar-eas has been found to be a key descriptor of hydrological and biogeochemi-cal response, with a larger proportion of event water (Laudon et al., 2007) and low mean transit times (Lyon et al., 2010) during snowmelt, higher spe-cific discharge during drier periods (Lyon et al., 2012), spatial patterns of dissolved organic carbon (Ågren et al., 2008; Buffam et al., 2007) and metal transport (Lidman et al., 2014). Additionally, contributions of deeper ground water during winter base flow show strong scale-dependence, with increas-ing groundwater contributions at larger scales (Peralta-Tapia et al., 2015a). These deeper inputs were also found to have a large impact on DOC and base-cation chemistry during low flow (Tiwari et al., 2014).

Generally, the past hydrological studies in Krycklan have focused on a few upland catchments (C2, C4, C7). Exceptions considering the full scale of the landscape are mainly based on isotope studies, without consideration of spatially distributed hydrometric data, and over limited time periods such as spring snow melt or base flow (Laudon et al., 2007; Lyon et al., 2010; Peralta-Tapia et al., 2015a). Distributed streamflow measurements across

21

Krycklan from three snapshot campaigns were used by Lyon et al. (2012), who found high spatial variability and changing spatial patterns between campaigns. Still, there remains a lack of knowledge about how the landscape as a whole responds hydrologically, across the range of shorter to longer timescales. Spatial variability across time scales and the role of the lower lying areas of the Krycklan catchment have not been the subject of earlier hydrological studies on streamflow.

22

Data and methods

Streamflow data Stream discharge was monitored at 14 partly nested catchments, including the main outlet between 2008 and 2013 (Table 2). The smaller catchments were monitored using V-notch weirs, flumes or well defined cross sections with road culverts. The three largest catchments (C14, C15, C16) were gauged at natural control sections. Hourly water level data were recorded automatically by using pressure transducers (Expert 3400, MJK A/S, Den-mark) connected to data loggers (CR1000, Campbell Scientific Inc., USA), capacitance rods (WR-HT 1000, TruTrack Inc., New Zealand), radar (RLS, OTT, Germany) or floats (Recorder Model X, OTT, Germany) for the dif-ferent catchments. Two or more loggers were deployed at each gauging sta-tion in case of failure of the primary device. This redundancy was also used for quality control of water height time series. For four catchments with heated housings, observations were possible year round for parts of the study period (Table 2). The remaining catchments were monitored over the ice-free season. The automatic water levels were corrected to account for logger offset using manual readings of reference water height. These were per-formed at regular intervals (at least biweekly during the growing season, more frequently during spring flood, and monthly during the winter season) in addition to during manual flow gauging. The water height time series of each logger was quality controlled manually (~1.8 million hourly data points) and filtered/corrected for influence of ice, drift/shifts and unrealistic water heights. Streamflow was gauged for defining rating curves using salt dilution (NaCl), velocity-area (current-meter, Acoustic Doppler Current Profiler) and time-volume measurements. Measurements at a certain occa-sion were often made at least in duplicates to increase confidence in the manual gauging, and these were averaged to create a single stage-discharge pair. A total of 325 stage-discharge pairs were used for the rating curves for the 14 catchments. Extrapolation beyond the highest streamflow gauging was on average only required for 0.4% of the hourly time series between October 2008 and September 2013 for all catchments.

Specific discharge, defined as stream discharge per unit catchment area, was calculated for each catchment. The catchment areas were calculated

23

using the D8 algorithm (O’Callaghan and Mark, 1984) on a 5m resolution digital elevation model (DEM), derived from airborne Light Detection And Ranging (LiDAR) measurements. Field mapping of catchment boundaries was used to modify the DEM derived catchment areas where needed, and questionable sections were further assessed with a 0.5 m resolution LiDAR DEM (Laudon et al., 2013, 2011).

Daily specific discharge time series were gap-filled using the HBV model (Bergström, 1976; Seibert and Vis, 2012) for periods with missing automatic water level data. The method of Jónsdóttir et al. (2008) was used, which adjusts the modelled data to achieve a smooth transition to the measured series preceding and following the data gap.

Table 2. Catchment gauging setup, summary of gap-filling and estimated uncertain-ty in streamflow measurements.

Area Infilled

Q Infilled

days Ɛ5-year Water level loggers Catchment Gauge type (ha) (%) (%) (%)

C1 90° V-notch weir 48 9 33 6 PT, TT C2 90° V-notch weira 12 18 37 8 PT, TTC4 90° V-notch weira 18 11 32 7 PT, TTC5 120° V-notch weir,

H-flumeb65 8 36 6 PT, TT

C6 Culvert 110 13 33 5 PT, TTC7 90° V-notch weirc 47 4 13 3 PT, TT, Float C9 Culvert 288 22 45 7 PT,TTC10 Culvert 336 24 46 12 PT,TTC12 Venturi flume 544 37 50 9 TTC13 Trapezoidal flume 700 24 52 10 PT,TTC14 Natural section 1410 33 52 11 TTC15 Natural section 1913 18 46 6 PT,TTC16 Natural section (bridge) 6790 18 43 9 TT,RLS C20 Culvert 145 27 45 11 TTa: heated weir since 2011; b: heated flume since 2012; c: heated weir since 1981 Ɛ5-year: estimated measurement uncertainty of 5 year aggregated discharge; PT: Pressure trans-ducer (MJK 3400, with Campbell Scientific CR1000); TT: TruTrack capacitance rods (WT-HR 1000); Float: OTT model X float strip chart recorder; RLS: OTT RLS Radar.

Climate data Precipitation, temperature, radiation and potential evapotranspiration data were collected at the Svartberget climate station, located in central Krycklan (64°14' N, 19°46' E, 225 m a.s.l, Figure 1) as part of the reference climate monitoring program at Vindeln experimental forests. Measurements fol-lowed standard WMO routines.

24

Both daily and hourly precipitation was used. Daily precipitation data were manually measured using a standard Swedish Meteorological and Hy-drological Institute rain gauge throughout the year. During the growing sea-son precipitation was also recorded at 10 minute intervals using a tipping bucket (ARG 100, Campbell Scientific, USA), and resampled to hourly val-ues for this study. Mean daily and hourly temperature was recorded 1.7 m above ground in a ventilated radiation shield. Global radiation was measured using a CM11 solarimeter (Kipp&Zonen, The Netherlands), and hourly av-erages were used. Daily potential evapotranspiration was estimated accord-ing to the Penman method using air temperature, humidity, net radiation and wind speed. Estimates were adjusted to the site using simultaneous meas-urements from a weighing lysimeter over the period 1981-84 (M. Ottosson-Löfvenius, pers. comm.).

Catchment landscape characteristics A number of physical landscape characteristics can potentially influence the hydrology of a catchment. Catchment characteristics covering topography, soils, land use and vegetation were selected to describe the variability within the Krycklan catchment (Table 3).

Methods

Temporal aggregation For paper I and II specific discharge variability and association to landscape characteristics were investigated at timescales ranging from daily to the en-tire five year period. The temporal resampling was done by aggregating spe-cific discharge over fixed periods, i.e. day, week, month, season and year. Seasons were divided into spring (AM), summer (JJA), autumn (SO) and winter (NDJFM) based on the procedure of the Swedish Meteorological and Hydrological Institute (Vedin, 1995), also used in paper IV. Annual specific discharge was aggregated over the hydrological year from October 1st to September 30th. Winter season discharge, which was gap-filled for most catchments, was excluded from analysis on shorter temporal scales than annual. For paper III the seasons were divided slightly differently because of expected changes in future temperature altering, for example, the onset of spring.

25

Table 3. Description of the catchment characteristics used in this thesis.

Catchment characteristic DescriptionArea Catchment area. Delineated with D8 algorithm using 5m LiDAR

DEM, with supporting field observations and 0.5m LiDAR DEM (Laudon et al., 2013, 2011).

MSCA Median subcatchment area following McGlynn et al. (2003). A meas-ure of drainage network organization.

Slope Catchment slope from 5m LiDAR DEM, calculated similar to Seibert and McGlynn (2007).

Tang. curvature Tangential curvature calculated perpendicular to slope gradient, i.e. a measure of flow divergence and confluence (Conrad et al., 2015).

Elevation Catchment mean elevation from 5m LiDAR DEM

EAS Catchment mean elevation above stream from 5m LiDAR DEM, cal-culated similar to Seibert and McGlynn (2007).

LFS Length of flow paths to stream network.

GFS Gradient along flow path to stream.

LFS/GFS Ratio median flow path length to median flow path gradient. See McGuire et al. (2005).

TWI Topographical wetness index, calculated similar to Grabs et al. (2009).

Forest Catchment forested area cover. See Laudon et al. (2013).

Wetland Catchment wetland cover, strongly correlated with peat soil cover which is left out here. See Laudon et al. (2013).

Lake Catchment lake cover. See Laudon et al. (2013).

Wet area Sum of wetland and lake area.

Till and thin soils Cover of till and thin soils, from quaternary deposits map of the Geo-logical Survey of Sweden (SGU). See Laudon et al. (2013).

Sediment soils Cover of silt, sand and glaciofluvial sediments, from quaternary depos-its map (SGU) See Laudon et al. (2013).

Tree volume Tree volume determined from LiDAR measurements and forest inven-tory. See Laudon et al. (2013).

Soil depth Mean catchment soil depth, calculated from the SGU soil depth model map (Daniels and Thunholm, 2014). Soil depth is here equivalent to depth to bedrock.

PETTurc and insolation

Spatially variable potential evapotranspiration (PET) based on the temperature and radiation driven Turc method, and potential insola-tion. See Lyon et al. (2012).

Paper III had some differences in the classification, excluded in papers II and IV due to low occurrence (e.g. rock outcrops) or strong correlation with other metric (e.g. aspect and insola-tion).

26

Analysis of spatial variability In paper I the spatial variability in specific discharge was quantified using different metrics, selected to facilitate comparison with other studies. These metrics were the coefficient of variation (CV), ratio of interquartile range and the median (CIQR), as well as the relative deviation of specific discharge from a subcatchment compared to the main outlet (C16). These metrics were then summarized using the total range and the median value.

To investigate the relationship between landscape characteristics and dis-charge characteristics we used both multivariate and univariate approaches. In paper II we used principal component analysis (PCA) on the landscape characteristics, and related the principal component scores of the catchments to specific discharge. Landscape characteristics were also used to group catchments to different classes using k-means clustering. In paper IV partial least squares (PLS) is used to relate spatial and temporal variability in reces-sion characteristics to landscape characteristics, and metrics of evapotranspi-ration and storage. Univariate analysis was made using non-parametric Spearman rank correlations (Spearman, 1904) between landscape character-istics and specific discharge (paper II), and streamflow indices (paper III).

Specific discharge uncertainty Uncertainty in specific discharge was quantified for each catchment in paper I. We estimated uncertainty in the measured discharge series based on ratingcurve regressions and in the gap-filled series based on modelled ranges (10th

and 90th percentile) of flow. Additionally, catchment area uncertainty wasincluded (5%) in the calculation of specific discharge. A Monte-Carlo exper-iment with 106 time series for each catchment was applied to evaluate theeffect of specific discharge uncertainty on measures of spatial variability.

Hydrological modelling The bucket-type precipitation-runoff model HBV was used to model stream-flow for the 14 subcatchments under current (1981-2010) and future (2061-2090) climate conditions in paper III.

For each of the 14 subcatchments, the HBV model was first calibrated on observed daily streamflow between 2008 and 2013, using precipitation, tem-perature and potential evaporation as forcing data. To consider parameter uncertainty, a total of 100 calibrations resulting in 100 optimal parameter sets were performed with a global genetic algorithm followed by a local Powell optimization (Seibert, 2000). Due to the short period of observed data for most sites, validation of the models was only made on C7 where longer term daily discharge is available. Based on calibration under the period 2008

27

to 2013, it was validated against streamflow data and signatures from year 1991 to 2000.

Bias corrected precipitation and temperature from 15 regional climate models (RCM) were used to simulate streamflow under current and future climate conditions. Potential evaporation, required as forcing data, was cal-culated using the temperature-based Hamon method and adjusted to match potential evaporation in Krycklan described above for the current climate. The results of RCM-driven and observed climate-driven HBV simulations were compared using streamflow signatures for all catchments under current climate. The 15 different RCMs combined with 100 optimal parameter sets resulted in 1,500 time series of 29 years (discarding the first year as model warmup) for both current and future climate for each catchment (87,000 annual streamflow series in total). Uncertainty in climate predictions was considered through the use of 15 different RCMs.

A total of 26 streamflow signatures describing water balance, annual and seasonal streamflow hydrograph and flow durations were calculated for cur-rent and future climate simulations. Relative changes in these signatures between current and future scenarios were used to relate changes to land-scape characteristics.

Streamflow recession analysis Two approaches were used for the recession analysis in paper IV. The first approach was a temporally lumped analysis where hours without influence of evapotranspiration, precipitation and snowmelt were included. The second approach considered individual recession events. Both analyses used the method of Brutsaert and Nieber (1977), where streamflow recession is ana-lyzed by relating change in discharge, dQ/dt, to concurrent rate of discharge, Q by

dQQ

dt (1)

where α and β are determined as intercept and slope using linear regression on data pairs of log(-dQ/dt) vs. log(Q).

In the lumped approach data points of dQ/dt were selected using hourly time series of discharge for each catchment for periods when precipitation, snow melt and evapotranspiration (ET) were assumed to be negligible based on the approach of Kirchner (2009) and Teuling et al. (2010). Assuming that change in storage (S) during these periods without precipitation and ET is only affected by discharge Q (dS/dt = -Q), this is then describing the storage

28

discharge relationship. The lumped approach thus gives a single dQ/dt-Q relationship for each catchment.

The individual recession analysis was performed using the same method, but for each recession event separately (Shaw and Riha, 2012). For this analysis periods with evapotranspiration were included, and only periods of precipitation and snow melt were excluded. Data were aggregated to daily time scales, due to large diurnal fluctuations during summer low flow and increased accuracy when considering individual events (Chen and Krajewski, 2016).

For comparing recession characteristics across catchments and between events we used β from equation 1, describing the degree of non-linearity of the recession, and recession timescale, τ, calculated as

1

1( )Q

Q (2)

where τ is dependent on discharge Q, except for the case of β = 1 (Q is a linear function of S, i.e. a linear reservoir) (Krakauer and Temimi, 2011). In addition, for the case of the analysis of individual recession events, we quan-tified possible shifts in recession rates between events (i.e. a translation of the dQ/dt-Q relationship). Typically, these shifts are quantified by fixing β between different recession events and analyzing variability in α (e.g. Bart and Hope, 2014; Shaw and Riha, 2012). However, this has some drawbacks as not being able to consider variation in β as well as α still being influenced by the choice of fixed β in the regression. If β is sufficiently stable between events compared to changes in the intercept α, this approach is considered applicable. But, due to larger variation in β in our data set, we applied a dif-ferent approach by using the single lumped dQ/dt-Q relationship as a refer-ence, and compared the individual events with respect to how much they deviate from this relationship. The concept of using a no-ET recession curve as a reference is widely applied, for example to estimate actual ET from streamflow records (e.g. Szilagyi et al., 2007). We quantified the shift in recession events as the ratio between the sum of discharge occurring over each observed recession event segment (QR-seg) and the sum of discharge occurring over the same period from the single lumped recession curve using the observed initial discharge (QR-lumped).

Multivariate partial least squares regression (PLS, Eriksson et al., 2006) was then applied to relate the recession characteristics β, τ and QR-seg/QR-lumped to catchment characteristics (spatial analysis), and metrics of antecedent storage and evapotranspiration (temporal analysis). For evapotranspiration we use the mean concurrent potential evapotranspiration (PETs), and for antecedent storage we use discharge at the onset of the recession (Qp0), cu-

29

mulative discharge prior to the event (between 10 and 2 days before event, Qp8) and modelled HBV storage at the onset of the recession event (SHBV).

30

Results

This thesis examines the spatial patterns of streamflow characteristics in the Krycklan catchment and their relation to landscape properties. In paper I the spatial variability in specific discharge is analyzed across temporal scales. Paper II explores the relationship between the spatial patterns of specific discharge and catchment characteristics. Paper III demonstrates the influ-ence of variable specific discharge in climate change predictions. Paper IV focuses on the spatiotemporal patterns of streamflow recession characteris-tics. The main results of these papers are summarized below.

Spatiotemporal variability of specific discharge The spatial variability in specific discharge in Krycklan was quantified at different temporal scales from daily to multi-annual in paper I. Compared to the main outlet (C16), the specific discharge of the subcatchments ranged between 61% and 150% on the annual time scale, with a median absolute deviation of 19% from C16. Comparable spatial variability was observed at the seasonal scales, with differences relative to C16 ranging between 72 and 175% for spring, 34 and 130% for summer as well as 46 and 175% during autumn. Moving to shorter temporal scales (monthly, weekly and daily), the spatial variability in specific discharge increased (Table 4).

Table 4. Spatial specific discharge variability for different aggregation periods. Note that annual metrics include gap-filled winter flow, while other aggregation periods do not.

Aggregation period

Range relative to

outlet (C16) Median absolute devia-tion from outlet (C16)

median CV

median CIQR

median Q mm day-1

Day 0-414 % 33 % 35 % 43 % 0.72 Week 0-248 % 30 % 31 % 36 % 0.81 Month 11-205 % 24 % 24 % 25 % 0.92 Spring 72-175 % 20 % 17 % 17 % 2.22 Summer 34-130 % 14 % 16 % 18 % 0.67 Autumn 46-175 % 24 % 22 % 25 % 0.93 Annual 61-150 % 19 % 18 % 19 % 0.88

31

The spatial patterns between years were consistent, with catchments having similar specific discharge relative to each other. For the annual specific dis-charge, spatial Spearman rank correlations varied between 0.80 and 0.96 for all combinations. This spatial consistency also existed for seasonal periods, for example the median Spearman correlation was 0.82 between different spring periods. Summer and autumn showed less correlation between differ-ent years, and also with other seasons. For example summer and spring were less correlated (Spearman rank correlation 0.02 to 0.75, median 0.40). In other words, the ranking between catchments changed between seasons, and it was not the same catchments having high and low flow during spring and summer.

Furthermore, the spatial variability tended to be larger during low flow periods compared to high flow periods (Figure 4). During weeks of high flow, the weekly spatial variability approached that of the annual variability. During drier weeks, the spatial variability was consistently larger. This pat-tern remained even when considering the higher measurement uncertainty during low flow conditions.

Figure 4. Spatial variability (CV) in weekly specific discharge plotted against mean specific discharge at C7 (as a measure of general catchment wetness). Error bars show range in CV when considering uncertainty in rating curve definition, gap-filling and catchment area. Shaded area show the range of CV for annual specific discharge. Vertical bars show median (dashed) and mean (solid) specific discharge at C7.

Landscape controls on specific discharge variability The results from paper I summarized above showed that the spatial variabil-ity in specific discharge in Krycklan persists over longer periods and that the

32

spatial patterns were variable over time. The relation between spatial varia-bility in specific discharge and catchment characteristics across various time periods was investigated in paper II.

Both the multivariate PCA and univariate analysis showed a strong asso-ciation between specific discharge and landscape characteristics. These spa-tial relationships changed between seasons and hydrological conditions. At the annual timescale the specific discharge was positively related to portion of wet areas (wetland and lake) covering the catchments and negatively to forest cover and tree volume (Figure 5). Catchments with larger portions covered by wet areas had roughly 40-80% higher annual specific discharge compared to catchments with the highest tree volumes on till soils. A similar spatial pattern was seen during spring snowmelt and wetter autumn seasons. The pattern of increasing specific discharge in catchments with greater pro-portions of wet areas was more pronounced during wet periods compared to drier periods. Catchment tree volume was negatively related to specific dis-charge during all five summer seasons, and drier autumn seasons.

Figure 5. Spearman rank correlations between specific discharge and individual catchment characteristics for seasons and hydrological years. White cells are not significant (p > 0.05).

Forested areas on both till and sediment soils had more similar annual spe-cific discharge (sediment ranged from +8 to +40% compared to till), but with larger differences during drier periods. Patterns at shorter temporal scales were more variable. For example during low flow periods, catchments with deeper sediment soils maintained a higher base flow compared to the till areas. This pattern was reversed for periods with high flow (Figure 6).

33

Figure 6. a) Top: Daily specific discharge at C7 and daily precipitation. Bottom: Time series of Spearman rank correlation between catchment specific discharge and sediment soil cover for different temporal aggregation scales. Only data between March 2011 and October 2013 are shown, while the full data set was used in the analysis. b) Weekly Spearman rank correlation (rs) for seasons spring to autumn plotted against weekly specific discharge at C7. Grey shaded area denotes non-significant correlations (p > 0.05).

The spatial differences and the changing spatial patterns with wetness states between the main landscape types in Krycklan, namely forested till, sedi-ment and wet areas, can be summarized by their flow duration curves (FDC, Figure 7). Both wet areas and forested till had high maximum flows com-pared to the sediment areas. As streamflow approaches lower values, the wet areas were able to maintain a higher flow, while the densely forested till catchments dry out much faster. Sediment areas have a much flatter FDC with a dampened response at high flows and higher base flows compared to the rest of the landscape.

34

Figure 7. Flow duration curves for catchments belonging to main landscape types (forests on till, sediment and wetland areas). Catchments with largely mixed charac-teristics (C9, C12, C13, C15) are not shown.

Spatial variation of climate predictions In paper III hydrological response was simulated for the 14 subcatchments using the HBV model under current (1982-2010) and future (2062-2090) climate conditions using an ensemble of 15 regional climate models. The focus of this paper was to estimate the variability among catchments with regards to the change in response between current and future climate. Simu-lated annual streamflow at the catchment outlet (C16) increased about 20% as a result of the projected increase in precipitation (15%) and temperature (+3.7°C). The seasonal changes vary, with little change projected over the summer (5% less streamflow), and larger differences during winter (+125%) and spring flood (-23%) for C16.

These changes resulted in a general shift in the streamflow regime and flow duration curves, where highest flows were reduced and base flow in-creases (Figure 8). This was found for all subcatchments. Despite the overall similarities in regime shift, there was a high variability in predicted change in the hydrological response among catchments. For example, did the medi-an projected increase in annual streamflow vary between 12 and 25% for the subcatchments.

35

Figure 8. (a) Simulated streamflow and (b) flow duration curves for catchment C16 under current and future climate scenarios. Solid lines are median of the simulations and the shaded areas mark the interquartile range (IQR, between 25th and 75th per-centile).

Several aspects of the predicted change in streamflow indices were related to the catchment characteristics, especially the contrast between forested areas and wetland areas. Forest and wetland areas are strongly correlated as non-forest areas are largely covered by wetlands in Krycklan, and can be inter-preted simultaneously. For example, catchments with more forest (and less wetland) were predicted to have a larger percentage increase in future annual streamflow and summer base flow index, but a smaller decrease in spring flood magnitude (Figure 9).

Figure 9. Relationships between catchment wetland cover and median projected change between current and future climate simulations for a) summer base flow index (BFI), b) spring peak flow and c) mean annual flow. Wetland cover is inverse-ly correlated to forested areas.

36

Spatiotemporal variation in streamflow recessions In paper IV we found event recession characteristics to be variable between different catchments in Krycklan and temporally variable between different events. Strong spatial relationships between catchment characteristics and recession behavior were found in the PLS analysis of individual recession events for recession non-linearity (β) and recession timescales (τ). Catch-ment topography (terrain slope and elevation above stream), proportion of deeper sediment soils and catchment area were found to be important factors for describing variability among catchments. Steeper slopes, higher elevation above stream and deeper sediment soils, found mostly at lower elevations, were related to longer recession timescales meaning that these areas can sustain their base flow for longer periods. Catchment area gained more im-portance as flow gradually decreased, and especially the largest catchments (C14, C15, C16) had long recession timescales. However, also the smaller C20, with sediment deposits near the stream channel, had long recession times and high base flow during low flow, indicating that soils play a key role on storage and release of water. The spatial patterns for the temporally lumped approach showed weaker relation to landscape characteristics than the individual recession events.

The degree of temporal variability in recession events was variable among catchments (Figure 10). We quantified the temporal variability using varia-tion in β and shifts in dQ/dt-Q (QR-seg/QR-lumped). Some catchments showed little temporal variability in dQ/dt-Q and β (e.g. wetland and lake areas C4, C5), while others showed large variations (e.g. small till soil catchments C1, C2). Larger catchments with deeper sediment soils along the stream channels (e.g. C16, C20) showed the highest variability in β, but less variation and shifts in recession rates dQ/dt-Q.

37

Figure 10. Plots of -dQ/dt-Q for catchments C2, C5 and C16. Plots show the single lumped relationship (red dashed line), all pairs of daily values (grey points) and the individual recession events used in the analysis. C2 shows high variability in reces-sion rates, while C5 and C16 show less variation in recession rates among different events.

The controls on temporal variability between recession events were evaluat-ed using different metrics for antecedent storage and evapotranspiration (ET) in PLS models for each catchment. In general, ET showed a strong influence on recession β, with lower β related to higher ET. This implies that stream-flow declines faster when ET is higher during the recession. However, this influence was variable among catchments. Larger catchment area was relat-ed to a lower influence of ET on the streamflow recession (Figure 11, left).

Shifts in streamflow recession, as seen in translations in dQ/dt-Q plots, were found to mainly be related to antecedent storage conditions rather than concurrent ET. A lower initial storage was related to faster decline in stream-flow. Antecedent storage was more important in catchments with a higher elevation above the stream network (EAS), and thus less important with the

38

low EAS found in catchments having lake influences and low gradient wet-lands (Figure 11, right).

Figure 11. The left-hand plot shows the PLS loadings of concurrent PET for ex-plaining temporal variability in β (non-linearity) for all 14 catchments. A higher absolute loading signifies higher importance in explaining the variability. The larger catchments show less influence of concurrent PET, with lower loadings in the PLS models. The right-hand plot shows loadings for explaining shifts in recession dQ/dt-Q between events (QR-seg/QR-lumped). Antecedent storage (discharge at onset of reces-sion, Qp0, shown here) was important for explaining these shifts between events, with higher loadings for catchments having higher elevation above stream. Open circles signify low loadings that were not considered important in the PLS models (i.e. Qp0 was not important for explaining temporal variability in recession shifts for two of the catchments).

39

Discussion

Patterns and controls on spatial variation in specific discharge In paper I we found a large spatial variability in specific discharge between nearby catchments in this boreal landscape. Not only was the presence of high short term spatial variability confirmed (cf. Lyon et al., 2012), but vari-ability on annual time scales was on the same order of magnitude as has been observed for the effect of clear-cutting large parts of forested catch-ments, or the predicted effects of a century of climate change in the region. For example, climate change has been predicted to increase streamflow by 10-30% (Andréasson et al., 2004) and clear-cutting experiments have shown increases of 20% to 35% in annual streamflow in boreal forests (Ide et al., 2013; Sørensen et al., 2009). In comparison, annual specific discharge varied between 61 and 150% compared to the catchment outlet, with a median ab-solute deviation of 19%. The annual variability is similar to what Nicolson (1988) found at Turkey Lakes in Canada, higher than observed at Hubbard Brook, USA, and lower than at Coweeta, USA, or Gomadansan, Japan (Yanai et al., 2015).

Not only did the degree of variability change between seasons and tem-poral scale of analysis, but also the spatial patterns of specific discharge changed. It was often not the same catchments having high flow during spring as having high flow during summer. Spatial rank correlations showed that the similarity in spatial patterns was stronger between periods of similar flow magnitudes, i.e. two high flow periods had stronger correlation than high and low flow periods. This indicates a changing spatial pattern with wetness conditions, also observed elsewhere for hydrological and biogeo-chemical responses and processes (Buffam et al., 2007; Grayson et al., 1997; Jencso and McGlynn, 2011; Payn et al., 2012). Spatial variability was rela-tively larger during periods of lower flow than during periods of higher flow. For instance, the weekly spatial variability during high flow approached the annual spatial variability. We hypothesize that the larger spatial differences during drier periods, observed across time scales, can be related to spatial differences in evapotranspiration, snow accumulation, and storage-release of water which is enhanced as the landscape dries out.

40

These patterns of variability disprove the assumption of uniform specific discharge that was previously made for this landscape. We also suggest that this variability is a valuable source of information about the hydrological functioning of these catchments, particularly with regards to the role of catchment storage in shaping the flowing regime.

Landscape patterns of specific discharge In paper II we found a strong association between spatial patterns of specific discharge and landscape characteristics. Variable spatial patterns were found depending on the temporal scale of analysis, wetness conditions and season. Analyzing spatial patterns across temporal scale revealed spatial patterns and insights into landscape patterns that were not visible when considering only annual or seasonal periods.

For annual and longer time periods there was a strong positive relation-ship between wet areas and specific discharge in Krycklan. Previous work in Krycklan has shown that these areas exert a large influence on hydrological and biogeochemical processes. In the boreal region in general, both similar (Prepas et al., 2006) and contrasting (Buttle and Eimers, 2009) results on the role of wetlands for annual specific discharge have been reported. These contrasting influences of wetlands are likely linked to their topographical position and connectivity, where headwater wetlands increase runoff and valley bottom wetlands along the stream may reduce runoff during drier periods (Hayashi et al., 2004; Quinton et al., 2003).

The average annual ET, estimated from the water balance, had a range of roughly 250-300 mm year-1 for wetland areas and 375-450 mm year-1 for forested catchments. This corresponded to about 55% and 80% of potential evapotranspiration, respectively. These annual differences in specific dis-charge between wet area and forested catchments can largely be related to the expected ET observed from latent flux or water balance estimation else-where in Sweden and the boreal zone in general (Grelle et al., 1999; Kasurinen et al., 2014; Peichl et al., 2013; Rosén, 1984; van der Velde et al., 2013).

The wet areas also showed a consistently positive relationship to specific discharge during snowmelt in spring. We relate this pattern to two factors. One is the difference in snow accumulation between open wet areas and forest. The second has to do with runoff generation mechanisms. Snow ac-cumulation has been observed to be about 30% higher in areas open com-pared to forests nearby Krycklan (Schelker et al., 2013; Sørensen et al., 2009). Snow interception in boreal forests can reach up to 30% of gross pre-cipitation (Lundberg and Koivusalo, 2003) and be 30% higher compared to open areas (Pomeroy and Schmidt, 1993). However, assuming a 30% higher

41

snow accumulation in wet areas compared to forested areas could only ac-count for 50-88% of the difference in spring specific discharge observed between catchments. We attribute the remaining differences to the shallow groundwater tables in wetlands compared to forested hillslopes (Kellner, 2001; Nilsson et al., 2008; Seibert et al., 2003), which are providing a lower storage deficit, combined with overland flow from snow melting on frozen peatland surfaces (Laudon et al., 2007) This could result in higher runoff coefficients for wet areas.

The change in correlation between landscape characteristics and specific discharge, both at seasonal timescales and shorter timescales, indicates that the contribution of streamflow from different parts of the landscape is wet-ness state dependent. Change in landscape controls on specific discharge between periods has also been identified by others (Kirnbauer et al., 2005; Kuraś et al., 2008) and been related to increasing importance of subsurface characteristics as flow gradually decreases during base flow (Payn et al., 2012).

For instance, during drier conditions we found patterns that were related to both vegetation and soil deposits. For the lowest flows, catchments with deeper sediment soils maintained a high specific discharge, while the more densely vegetated till areas had the lowest specific discharge. The sediment areas, located in the lower parts of the catchment, could function as a hydro-logic buffer where a relatively large reservoir maintains a stable release of base flow through longer and deeper flow paths during extended periods without precipitation or snowmelt (Knutsson and Fagerlind, 1977; Peralta-Tapia et al., 2015a; Tiwari et al., 2014). Some of these patterns related to low flow are not detectable during coarser aggregation periods due to the relatively small contribution from low flow to the total specific discharge, compared to the high flow events. This highlights the importance of consid-ering different temporal scales of analysis for a richer understanding of the spatial patterns of hydrological response.

The degree of spatial variability between the different catchment types was often larger during dry periods than wet periods (paper I). Assuming that annual ET is generally energy-limited rather than water-limited in Swe-den (van der Velde et al., 2013), increases in annual precipitation would have a larger impact on streamflow than ET. At the nearby Degerö mire, Peichl et al. (2013) found this effect on inter-annual variability between wet and dry years. There, ET remained stable between years (~200-300 mm year-

1), but streamflow showed large variability (~225-750 mm year-1). This would lead to smaller relative differences in annual specific discharge be-tween catchments with increasing precipitation, and vice versa for less pre-cipitation, assuming ET remains energy limited.

42

However, during more extreme dry periods ET might be constrained due to water availability. This can vary across the landscape depending on soil characteristics, vegetation and landscape position (Barr et al., 2009; Betts et al., 2001). This could lead to changing spatial patterns in specific discharge, for example as seen during dry periods between forested till and sediment areas. Therefore, we believe that spatial variability and patterns of specific discharge are not only controlled by landscape characteristics in Krycklan, but also influenced to a large degree by temporal variability in weather and climate. Small-scale, short-term variation in rainfall is one example of this. The lack of detailed information on this creates a source of noise in the short term analyses in this thesis.

Landscape variability in predicted climate change effects The results of paper III demonstrate that catchment responses to the same change in climate can be highly variable, and highlights the importance of considering present day spatial variability and its relation to landscape char-acteristics when thinking about the future (papers I, II). The current hydrological behavior was captured well by the HBV model using observed climate data from Krycklan, and a relatively good perfor-mance was achieved for the validation using forcing data from regional cli-mate models under current conditions. The spatial variability from the mod-elled current conditions (-24% to +30% relative to catchment outlet) was comparable to the observed range in variability over five years from paper I (-26% to +35%).

The predicted change in streamflow response under future climate condi-tions suggests more uniform flow conditions over the year, with a decrease in spring flood magnitude and an increase in base flow. Due to the warmer temperatures the ratio of rain to snow is predicted to increase, causing small-er spring snowmelt events. These general results are comparable to previous studies on the hydrological effects of climate change in Sweden (e.g. Andréasson et al., 2004; Bergström et al., 2001).

However, the various subcatchments exhibited different degrees of change between current and future conditions. Catchment responses to the same change in climate can be highly variable. This highlights the im-portance of considering present day spatial variability and its relation to landscape characteristics (papers I, II). Catchments with differences in the cover of wetland and forest showed different seasonal changes. These two landscape characteristics are strongly correlated, as non-forested areas are generally covered by wetlands at Krycklan and across much of the Sveco-fennian landscape. Forested areas had a larger increase in annual streamflow and summer base flow, while wetland areas show larger decreases in spring

43

streamflow compared to forested areas. The wetland areas have much higher spring specific discharge under the current climate, which can be linked to higher snow accumulation (paper II), and the warmer future climate with reduced snow packs can have a larger effect in the open wet areas. The high-er increase in future summer base flow for forested areas could be related to higher groundwater storage due to increased recharge and a higher rain to snow ratio during winter periods (de Wit et al., 2007). This effect is larger in forested areas compared to wetlands, where storage is already close to satu-ration. These findings suggest that the change in rain to snow ratio is im-portant for the spatially variable change in hydrologic response to future climate. Changes in the role of frozen soil and saturation excess conditions between wetland and forest will also be involved, complicating the situation.

Variability in streamflow recession across Krycklan For concurrent recession events analyzed in paper IV the catchment charac-teristics related to recession non-linearity (β) and recession timescale (τ) were terrain slope, deeper sediment deposits, catchment area and elevation. These characteristics co-vary to some degree, where catchments with more sediment soils generally have steeper slopes, deeper soils, larger catchment area and are located at lower elevations. Shorter recession timescales and lower β were related to smaller catchments in the upland areas, and vice versa for the lower lying sediment areas. The results are in agreement with several other studies where subsurface characteristics and terrain influence the recession characteristics (McMillan et al., 2014; Tague and Grant, 2004; Tallaksen, 1995). The spatial variability in recession characteristics was lower during spring following snowmelt than during summer or autumn. This indicates that the landscape behaves more uniformly during wet condi-tions compared to relatively dry conditions. This was also the case for the variability in specific discharge (paper I).

The lake outlet (C5) behaved consistently close to a linear reservoir (β ~ 1). This is in agreement with open water hydraulics and is typical for catch-ments with larger lake influences (Bogaart et al., 2015). Other catchments showed higher temporal variability in recession characteristics. We hypothe-size that this is related to the degree of heterogeneity in subsurface storage characteristics. Smaller catchments with wetland and lake influence showed the lowest temporal variability, till dominated catchments moderate variabil-ity and larger catchments with a mix of sediment, till and wetland showed larger variability for β. This is similar to other studies which found that larg-er areas with more heterogeneity were often associated with higher non-

44

linearity and variability (Clark et al., 2009; Harman et al., 2009; McMillan et al., 2014).

The larger influence of ET on recession rates seen for smaller catchments suggests that ET depletes the same store of water that controls streamflow recession. This implies that the system is relatively shallow, since the rooting zone is largely confined to the upper decimeters of soil (Bishop and Dambrine, 1995; Mellander et al., 2004). These catchments are dominated by till and peat soils, where groundwater levels are close to the surface in the near stream areas. Larger catchments, where deeper sediment soils are found along the stream network, showed less influence from ET, which is con-sistent with deeper flow paths (Peralta-Tapia et al., 2015a) which are less directly influenced by ET processes feeding the streams draining these catchments.

During recessions with lower antecedent storage there was a general pat-tern of faster recession rates compared to higher antecedent storage. These results are consistent with a number of recent studies showing that recession dynamics are related to catchment storage state (Bart and Hope, 2014; Patnaik et al., 2015; Shaw, 2015). This suggests that a single-reservoir stor-age-discharge relationship may not be applicable for some of the catchments in Krycklan, and that storage-discharge relationships vary depending on storage state. A general pattern of previous applications of single-reservoir storage discharge relationships suggests that these are more applicable in humid catchments and during wetter periods (e.g. Brauer et al., 2013). Nev-ertheless, Krycklan subcatchments with lower elevation above the stream network and with lake influence tended to show less dynamic behavior in recession rates and less influence of antecedent storage conditions. In these catchments an assumption of a single storage-discharge relationship may be more appropriate.

The spatial patterns of streamflow recessions were relatively stable be-tween events and wetness states, compared to the more variable pattern ob-served for specific discharge (paper II). There were, however, some tenden-cies of catchment area gaining importance during drier periods, and catch-ment terrain being relatively more important during wetter periods.

Hydrological functioning of the Krycklan landscape In this thesis I have demonstrated that a strong relation between streamflow and catchment characteristics exist in the Krycklan catchment. Through analysis of nearby catchments subject to a similar climate, it was possible to identify patterns related to landscape characteristics (even though weather, in the form of random spatial variability in rainfall at shorter timescales, is a

45

source of noise in the analysis of short term spatial patterns). This analysis made it possible to begin resolving the relation of landscape to specific run-off generation processes and the role of storage in how catchments translate precipitation into runoff.

In general, there are distinct patterns of behavior between different parts of the landscape. The main landscape elements of till, wetland, lakes and sediment show differences in partitioning of water and release to the stream network. Wetlands are important for both biogeochemistry and hydrology in Krycklan. Hydrologically, this thesis found that wetland area strongly influ-ences the water balance and specific discharge with high annual specific discharge, as well as during shorter wet periods. However, the wetland area does not appear to be a major factor for streamflow recession characteristics across Krycklan, but lakes exert a strong control on streamflow recession which is distinct from other landscape elements.

The more densely forested glacial till catchments have low annual specif-ic discharges, a more flashy hydrologic response and drain faster compared to the forested sediment and wetlands. The recession characteristics are sen-sitive to both storage and ET, resulting in fast recession rates and low specif-ic discharge during periods with high ET and lower storage.

The hydrological role of the lower sediment areas has previously been largely unexplored in Krycklan. Here we find that these areas strongly influ-ence the hydrological response across wetness conditions, and especially their importance for maintaining low flows through slow release of ground-water. The relatively low influence of ET on streamflow recession combined with deeper flow paths may explain the higher specific discharge observed in these catchments during dry summer conditions compared to other landscape elements. Deep percolation reduces the high flow peaks, and water is slowly released via longer flow paths maintaining a high low flow.

The temporal scale of analysis was shown to be important for a more complete understanding of the spatial patterns. Different information on the hydrological response emerge at different timescales. By analyzing a range from short to long timescales, a better understanding of low flow and reces-sion behavior was found that revealed changing patterns with season and wetness conditions. In general, the subcatchments behaved less differently during wetter conditions, with less spatial variability in discharge and more similar recession behavior. However, the associations between runoff behav-ior and landscape characteristics remain strong across wetness states. For spatial discharge variability over longer timescales and during wet periods, this is likely related to spatial differences in ET having a relatively larger effect during drier periods. When ET is limited mainly by available energy, an increase in precipitation will mainly result in higher specific discharge and, relatively speaking, similar specific discharge among catchments. Dur-

46

ing relatively dry periods, ET plays a relatively larger role in the water bal-ance and differences among catchments increase. Spatial variability in the storage characteristics and runoff generation likely play a key role for shorter timescales, and especially for low flow periods, reflected in the increased differences in both discharge and recession behavior during low flow.

The spatial variability between the nearby catchments remains large even at longer timescales and spatial patterns vary between seasons. Based on these results it is important to consider spatial variability in streamflow for studies on disturbance effects (e.g. land use and climate) and biogeochemical processes and exports. In this thesis, the importance was demonstrated for climate predictions, where the predicted change in streamflow was highly variable among the different catchments, and often related to the landscape characteristics. Thus, the choice of reference catchments for studies on dis-turbance can have an influence on the predicted effects.

Previous studies on landscape wide biogeochemical processes and exports in the Krycklan catchments have assumed a spatially uniform specific dis-charge. The results of this thesis imply that long-term mass balances need to be reevaluated (paper I). Future studies will also be able to better quantify the effect of the spatiotemporal variation in specific discharge on concentra-tion patterns such as hysteresis. This can increase our understanding of the relevant processes as well as linkages to landscape and flow paths.

Together, the results presented in this thesis show that patterns of stream-flow have high variability within the landscape. This variability shows strong association to landscape characteristics that may provide a basis for a simpler conceptual understanding of the heterogeneity in streamflow re-sponses of boreal catchments.

47

Conclusions

The main conclusions drawn from this thesis are: Specific discharge variability remained large even at longer time peri-

ods, and was on the same order as observed for major forest disturbance or predicted from a century of climate change. This has implications for not only hydrological studies, but also our understanding of biogeo-chemical processes and exports from across the landscape.

The spatial variability in the present day specific discharge also influ-ences the predicted climate change effects on streamflow.

Spatial variability in streamflow behavior is strongly related to catch-ment characteristics in this boreal landscape. The spatial patterns are temporally variable, and different patterns emerge between seasons and temporal scales of analysis. For annual timescales, the patterns are pro-portional to expected differences in evapotranspiration between wetland and forested areas.

Spring season differences can be explained largely by differences in snow accumulation between forested and open areas, with different run-off generation mechanisms for wetland and forests likely playing an im-portant role as well.

During summer low flow periods specific discharge is generally lower in areas with higher tree volumes and shallow soils that tend to dry out faster, while areas with deeper sediment soils and larger catchment area maintain a relatively higher base flow. Recession analysis showed that these areas have slower streamflow recession compared to upland till and wetlands, especially during drier periods.

Both evapotranspiration and antecedent storage influenced the rate of streamflow recessions at Krycklan. A larger influence of evapotranspira-tion was found for the smaller catchments and for catchments with lower elevation above the stream network (where groundwater can be assumed to be superficial). Lower antecedent storage generally leads to faster re-cession rates compared to high antecedent storage in this landscape, also indicating that a single-storage discharge relationship may not be appli-cable.

48

Future research There is still much to be learned about the distribution of water storage and release in the boreal landscape. As a next step, I would suggest to conceptu-alize the findings of this thesis in hydrological models that better represent the hydrological functioning of the different landscape elements and their positioning in the landscape. This will be useful for exploring effects of dis-turbance, processes identification through hypothesis testing, extrapolation to ungauged basins and landscape-wide geochemical mass balances. In addi-tion, much remains to be explored about winter base flow. Further infor-mation on patterns during winter could prove important for improved under-standing of storage-discharge relationships and the implications of climate change that is predicted to increase winter flows.

The results can also be used to assist in choosing appropriate model struc-tures, assist in planning of field monitoring campaigns to focus on areas of high importance for capturing variability or where the understanding of the hydrologic system remains poor. An important factor for further work in the Krycklan basin is to combine the rich data sets of streamflow, water chemis-try and natural tracers to improve our conceptual understanding of the sys-tem across spatial and temporal scales.

Ideally this thesis will help encourage other study sites to invest in ex-panding the spatial and temporal resolution of high quality flow measure-ments. These are a source of information about processes on their own, as in this study. However, even greater insights can be obtained by combining high-resolution flow data with other lines of investigation, such as hydrolog-ical tracers and subsurface hydrometric observations. This can help hydrolo-gy to answer its grand challenge of providing reliable information on water flow paths and transit times to geochemists and biologists (Bishop and Seibert, 2015).

49

Acknowledgements

I have been extremely lucky, with 3 fantastic supervisors! First of all I need to thank Jan, my main supervisor and cross-country nemesis. Despite being most of the time far away in Zürich, you’ve always found time to talk to me about my research (and sometimes snow conditions). Your advice has al-ways been helpful and you’ve definitely helped me see the bigger picture when I get lost in meaningless details. Thank you for the invitations to Zü-rich, which has been scientifically interesting and somewhat less pleasing skiing-wise for me. I’m still not sure if you invite me just to prove you are the faster/luckier skier again and again. The inevitable passing of time and aging is (hopefully) in my favor, and one day I will/might beat you in the track Jan!

Kevin, you are an endless source of ideas, optimism and enthusiasm. Coming to you with one idea, I usually left with a handful more and the feel-ing that anything is possible (but perhaps not always realistic…). I have learned so much from you and I’m beyond grateful for all the support you’ve given over the years, and for always being available. And you do make a killer soup!

Thomas, I don’t know how you know so much about everything, but somehow you do! You’re a walking encyclopedia of all things hydrology (and other stuff I know nothing about), and you’ve always given me great ideas and invaluable help! You’re really a true role model for me as re-searcher, and human being in general.

I think none of the work presented in this, and countless other theses, would have been possible without the King of North, Hjalmar! It is amazing what you have achieved with Krycklan, and I can hardly believe my eyes when I see all the new things happening up there every time I visit. It has truly been kanon to work with you and in Krycklan, and thanks for all the amazing advice and input with the writing of my papers.

Big thanks also to Claudia, for including me in your project. It’s been amazing to see how you make sense of big-data! Also thanks to my co-authors Ishi for the fantastic inputs and knowledge you have about Krycklan, and Mikaell for great help not only as a co-author but also for assistance with data and field work in Vindeln.

50

The data collected in this thesis wouldn’t exist had it not been for the amazing field crew in Krycklan! Without the work and help of Peder I would still be up in the forest collecting data. Thanks a million for showing me around Krycklan when I first arrived, for all the help you’ve provided me and for all the work you’ve laid down. Huge thanks also to Ida and Viktor for amazing field support! I also need to thank the Vindeln fieldstation crew, and especially Charlotta, Thomas and Pernilla. It has always been a great pleasure to come up to Umeå and Vindeln, and I’ve got a lot to thank Andrés for making it even better with dinners, climbing and tango (still not sure about that last one though).

Also huge thanks for helping me with field work (either through actually doing field work or having a great time drinking beer and taking in Vindeln): Stefan, José, Nino, Christian, Antoine, Aurélie, Fanny, Sylvain, Jeremy, Samuel, Juliane, Ana, Tum.

To everyone at Geo, thank you! Especially my office mates Bea and Tito, I don’t know how you’ve learned to live with my mess. Either you’ve just accepted it or I file it under Costa-Rican politeness. Also thanks to the “sen-iors” for making Geo a great place to work, study and teach; especially Al-lan, Roger, Giuliano, Sven, Tomas. Thanks, also non-Geoians, for all the amazing times over lunch, dinners, cycling, skiing, hiking, running, after-work, climbing, midsummer parties etc. Jean-Marc (frisbee kept me (kind of) sane!), Colin, Marc, Agnes, Bea, Martin, Adam, Lebing, Diana, Peter, Ida, Kaycee, Eduardo, Magnus, Anne, Hanne, Josefin, Nino, Anna, Audrey, Jo-han, Liang, Dorothée, Jon, Saba, Carmen, Albin, Zhibing, Viveca, Lichuan, Johanna, Korbi and many more!

To my family and friends back home and elsewhere in the world, thanks for always being there and supporting me no matter what. You hold a special place in my heart. Last, but certainly not least, thank you Helga (also the cover artist!) for all your love, support and our adventures together. I could definitely not have done any of this without you!

51

Sammanfattning på svenska (Summary in Swedish)

Hur olika delar av landskapet påverkar vattenbalansen och bidrar till avrin-ningen har länge varit en central fråga inom hydrologin. Även om det är känt att landskapsegenskaper, såsom topografi, geologi och vegetation påverkar vattenbalansen och avrinningen, saknas fortfarande detaljerad kunskap om hur landskapets egenskaper påverkar hydrologiska processer. Sådan kunskap kan öka vår förståelse av hur avrinningsområden lagrar vatten och hur avrin-ning bildas. Den kan underlätta uppskattning av avrinning i områden utan mätningar och ge förbättrat underlag för förvaltningen av vattenresurser, näringsämnen och föroreningar.

Hydrologiska system är komplexa och heterogena vad gäller processer och respons, både i tid och rum. I stället för att studera komplexiten som finns i olika områden var för sig, kan man söka efter enkla principer som kan beskriva komplexiteten i olika avrinningsområden och deras hydrologiska funktion. Ett sätt att uppnå detta är att analysera mönster i landskapet för att hitta enkla begrepp som kan beskriva den hydrologiska heterogenitet vi ob-serverar. Hydrologiska egenskaper som kan ingå i en sådan analys är vatten-lagring och avrinningsbildning. Avrinningen i ett vattendrag integrerar olika processer inom dess avrinningsområde, och avrinningsdata är därför en god grund för en sådan analys.

Det boreala landskapet kan beskrivas som en mosaik av olika landskaps-element såsom våtmarker, sjöar och skog. Det primära syftet med denna avhandling var att utnyttja den rumsliga variationen i avrinning inom ett landskap för att förbättra förståelsen för vad som styr variabilitet i avrin-ningsbildningen. Detta gjordes genom fortlöpande mätningar av avrinning i olika delar av landskapet, kvantifiering av variabilitet och sökande efter mönster som kan beskriva variabiliteten. Avhandlingen använde data från Krycklans avrinningsområde i norra Sverige, ca 50 km från Umeå och Ös-tersjöns kust. Vattenföringsdata från 14 delavrinningsområden (0.12-68 km2) inom Krycklan, med varierande landskapskarakteristik såsom olika jordarter, vegetation och topografi, användes för att undersöka rumsliga mönster hos avrinningen över olika tidsperioder.

52

Resultaten visar att variationen i specifick avrinning (avrinning per area-enhet) mellan närliggande avrinningsområden var stor även vad gäller års-värden. Variationen var i samma storleksordning som effekterna av stora skogsavverkningar och av förutspådda effekter av det kommande seklets förväntade klimatförändringar. Avrinning från olika områden var starkt rela-terad till deras landskapsegenskaper. Avrinningsområden med stor andel våtmarker kunde ha 40-80% större årsavrinning än avrinningsområden med tät skog på moränjordar. Särskilt under vårens snösmältning och andra blöta perioder hade våtmarksområden en hög avrinning jämfört med resten av landskapet. Detta kan till stor del förklaras av hög ackumulering av snö un-der vintern i dessa våtmarksområden jämfört med tät skog, liksom skillnader i avrinningsbildning. Skogsområden antas ha en större lagringskapacitet genom att en större andel smältvatten kan lagras i marken. Våtmarker är nära vattenmättnad och i kombination med en frusen yta under våren kan mindre vatten lagras och en högre andel av snösmältning och regn rinner av.

Under torrare perioder, visade djupare sedimentjordar förmåga att upp-rätthålla en högre avrinning än både skog på moränjord och våtmarker. Detta är sannolikt relaterat till djupare och mer genomsläppliga jordarter, vilka ger högre grundvattenbildning och ett stabilare utflöde av lagrat vatten till vat-tendragen. Dessa områden fungerar som en hydrologisk buffert, där toppflö-den dämpas och utflödet av vatten är mer stabilt över tid. Resultaten visar också att lagringen som bidrar till avrinningen i dessa större vattendrag på-verkas mindre av avdunstning än de med ytligare flödessystemen i morän och våtmarker. Det kan ytterligare öka mängden vatten tillgängligt för avrin-ning under torrare perioder.

Variationen i avrinning mellan olika områden är viktig att tänka på när man studerar klimatförändringars effekter på avrinning. Med de förväntade klimatförändringarna (en ökad regnmängd med 17%, en temperaturökning på 3.7 grader och 13% ökad evapotranspiration mot slutet av århundradet) visade beräkningar generellt ökad årlig avrinning. Ökningen var 20% för Krycklans utlopp, men variationen var stor mellan olika delavrinningsområ-den. Till exempel visade avrinningsområden med mer skog en större ökning av den årliga avrinningen och basflöden under sommaren än områden som domineras av våtmark. Förväntade effekter kan således variera betydligt beroende på vilka avrinningsområden vi använder som referens för klimat-studier.

Den stora variationen i avrinning mellan olika avrinningsområden är även viktig för vår förståelse av geokemiska massbalanser och för exempelvis export av kol från det boreala landskapet, vilken spelar en viktig roll i den globala kolcykeln. Tidigare har antagits att den specifika avrinningen inte varierar mellan delavrinningsområden i Krycklan. Denna avhandlings resul-

53

tat tyder på stor variation och visar att rumsliga mönster av till exempel ex-port av lösta ämnen som kol från tidigare studier kan behöva omvärderas.

Avhandlingen har givit en bättre förståelse av avrinningens rumsliga vari-ation under olika årstider och i olika tidsskalor och dess samband med områ-desegenskaper. Tillsammans visar de resultat som presenteras att mönstren för avrinning visar stor variation i landskapet. Variationen har stark koppling till landskapsegenskaper, och denna koppling ger en grund för en bättre för-ståelse av den heterogenitet som vi observerar hos avrinningen i boreala landskap.

54

References

Ågren, A., Buffam, I., Berggren, M., Bishop, K.H., Jansson, M., Laudon, H., 2008. Dissolved organic carbon characteristics in boreal streams in a forest-wetland gradient during the transition between winter and summer. Journal of Geophysical Research 113, 1–11. doi:10.1029/2007JG000674

Ågren, A.M., Buffam, I., Cooper, D.M., Tiwari, T., Evans, C.D., Laudon, H., 2014. Can the heterogeneity in stream dissolved organic carbon be explained by contributing landscape elements? Biogeosciences 11, 1199–1213. doi:10.5194/bg-11-1199-2014

Andréasson, J., Bergström, S., Carlsson, B., Graham, L.P., Lindström, G., 2004. Hydrological Change – Climate Change Impact Simulations for Sweden. AMBIO: A Journal of the Human Environment 33, 228–234. doi:10.1579/0044-7447-33.4.228

Atkinson, S.E., Sivapalan, M., Viney, N.R., Woods, R.A., 2003. Predicting space-time variability of hourly streamflow and the role of climate seasonality: Mahurangi Catchment, New Zealand. Hydrological Processes 17, 2171–2193. doi:10.1002/hyp.1327

Barr, A., Black, T.A., McCaughey, H., 2009. Climatic and Phenological Controls of the Carbon and Energy Balances of Three Contrasting Boreal Forest Ecosystems in Western Canada, in: Noormets, A. (Ed.), Phenology of Ecosystem Processes. Springer New York, New York, NY, pp. 3–34. doi:10.1007/978-1-4419-0026-5_1

Bart, R., Hope, A., 2014. Inter-seasonal variability in baseflow recession rates: The role of aquifer antecedent storage in central California watersheds. Journal of Hydrology 519, 205–213. doi:10.1016/j.jhydrol.2014.07.020

Bergström, S., 1976. Development and application of a conceptual runoff model for Scandinavian catchments. SMHI Reports RHO, No. 7, Norrköping, Sweden.

Bergström, S., Carlsson, B., Gardelin, M., Lindström, G., Petterson, A., Rummukainen, M., 2001. Climate change impacts on runoff in Sweden - Assessments by global climate models, dynamical downscalling and hydrological modelling. Climate Research 16, 101–112. doi:10.3354/cr016101

Betts, A.K., Ball, J.H., McCaughey, J.H., 2001. Near-surface climate in the boreal forest. Journal of Geophysical Research 106, 33,529-33,541. doi:10.1029/2001JD900047

55

Bishop, K., Dambrine, E., 1995. Localization of tree water uptake in Scots pine and Norway spruce with hydroiogical tracers. Canadian Journal of Forest Research 25, 286–297. doi:10.1139/x95-033

Bishop, K., Seibert, J., 2015. A primer for hydrology: the beguiling simplicity of Water’s journey from rain to stream at 30. Hydrological Processes 29, 3443–3446. doi:10.1002/hyp.10516

Bishop, K., Seibert, J., Köhler, S., Laudon, H., 2004. Resolving the Double Paradox of rapidly mobilized old water with highly variable responses in runoff chemistry. Hydrological Processes 18, 185–189. doi:10.1002/hyp.5209

Bishop, K., Seibert, J., Nyberg, L., Rodhe, A., 2011. Water storage in a till catchment. II: Implications of transmissivity feedback for flow paths and turnover times. Hydrological Processes 25, 3950–3959. doi:10.1002/hyp.8355

Bishop, K.H., 1991. Episodic increases in stream acidity, catchment flow pathways and hydrograph separation. PhD thesis, Cambridge University, Dept. of Geography, Cambridge.

Blöschl, G., Sivapalan, M., 1995. Scale issues in hydrological modelling: A review. Hydrological Processes 9, 251–290. doi:10.1002/hyp.3360090305

Blöschl, G., Sivapalan, M., Wagener, T., Viglione, A., Savenije, H.H.G. (Eds.), 2013. Runoff Prediction in Ungauged Basins. Synthesis across Processes, Places and Scales. Cambridge University Press.

Bogaart, P.W., van der Velde, Y., Lyon, S.W., Dekker, S.C., 2015. Stream flow recession patterns can help unravel the role of climate and humans in landscape co-evolution. Hydrology and Earth System Sciences Discussions 12, 9865–9913. doi:10.5194/hessd-12-9865-2015

Brauer, C.C., Teuling, A.J., Torfs, P.J.J.F., Uijlenhoet, R., 2013. Investigating storage-discharge relations in a lowland catchment using hydrograph fitting, recession analysis, and soil moisture data. Water Resources Research 49, 4257–4264. doi:10.1002/wrcr.20320

Brutsaert, W., Nieber, J.L., 1977. Regionalized drought flow hydrographs from a mature glaciated plateau. Water Resources Research 13, 637–643. doi:10.1029/WR013i003p00637

Buffam, I., Laudon, H., Temnerud, J., Mörth, C.-M., Bishop, K., 2007. Landscape-scale variability of acidity and dissolved organic carbon during spring flood in a boreal stream network. Journal of Geophysical Research 112, G01022. doi:10.1029/2006JG000218

Buttle, J.M., Eimers, M.C., 2009. Scaling and physiographic controls on streamflow behaviour on the Precambrian Shield, south-central Ontario. Journal of Hydrology 374, 360–372. doi:10.1016/j.jhydrol.2009.06.036

Chen, B., Krajewski, W., 2016. Analysing individual recession events: sensitivity of parameter determination to the calculation procedure. Hydrological Sciences Journal 6667, 1–15. doi:10.1080/02626667.2016.1170940

56

Clark, M.P., Rupp, D.E., Woods, R.A., Tromp-van Meerveld, H.J., Peters, N.E., Freer, J.E., 2009. Consistency between hydrological models and field observations: linking processes at the hillslope scale to hydrological responses at the watershed scale. Hydrological Processes 23, 311–319. doi:10.1002/hyp.7154

Conrad, O., Bechtel, B., Bock, M., Dietrich, H., Fischer, E., Gerlitz, L., Wehberg, J., Wichmann, V., Böhner, J., 2015. System for Automated Geoscientific Analyses (SAGA) v. 2.1.4. Geoscientific Model Development 8, 1991–2007. doi:10.5194/gmd-8-1991-2015

Daniels, J., Thunholm, B., 2014. Rikstäckande jorddjupsmodell. SGU-rapport 2014:14, Uppsala, Sweden.

de Wit, M.J.M., van den Hurk, B., Warmerdam, P.M.M., Torfs, P.J.J.F., Roulin, E., van Deursen, W.P.A., 2007. Impact of climate change on low-flows in the river Meuse. Climatic Change 82, 351–372. doi:10.1007/s10584-006-9195-2

Dixon, R.K., Solomon, A.M., Brown, S., Houghton, R.A., Trexier, M.C., Wisniewski, J., 1994. Carbon Pools and Flux of Global Forest Ecosystems. Science 263, 185–190. doi:10.1126/science.263.5144.185

Dooge, J.C.I., 1986. Looking for hydrologic laws. Water Resources Research 22, 46S–58S. doi:10.1029/WR022i09Sp0046S

Doyle, M.W., Stanley, E.H., Strayer, D.L., Jacobson, R.B., Schmidt, J.C., 2005. Effective discharge analysis of ecological processes in streams. Water Resources Research 41, W11411. doi:10.1029/2005WR004222

Dunne, T., 1983. Relation of field studies and modeling in the prediction of storm runoff. Journal of Hydrology 65, 25–48. doi:10.1016/0022-1694(83)90209-3

Eder, G., Sivapalan, M., Nachtnebel, H.P., 2003. Modelling water balances in an Alpine catchment through exploitation of emergent properties over changing time scales. Hydrological Processes 17, 2125–2149. doi:10.1002/hyp.1325

Eriksson, L., Johansson, E., Kettaneh-Wold, N., Trygg, J., Wikström, C., Wold, S., 2006. Multi- and megavariate data analysis. Part 1 Basic principles and applications, 2nd ed. MKS Umetrics AB, Umeå.

Gaál, L., Szolgay, J., Kohnová, S., Parajka, J., Merz, R., Viglione, A., Blöschl, G., 2012. Flood timescales: Understanding the interplay of climate and catchment processes through comparative hydrology. Water Resources Research 48, W04511. doi:10.1029/2011WR011509

Gibson, J.J., Edwards, T.W.D., Prowse, T.D., 1993. Runoff Generation in a High Boreal Wetland. Nordic Hydrology 24, 213–224.

Grabs, T., Seibert, J., Bishop, K.H., Laudon, H., 2009. Modeling spatial patterns of saturated areas: A comparison of the topographic wetness index and a dynamic distributed model. Journal of Hydrology 373, 15–23. doi:10.1016/j.jhydrol.2009.03.031

Grayson, R.B., Western, A.W., Chiew, F.H.S., Blöschl, G., 1997. Preferred states in spatial soil moisture patterns: Local and nonlocal controls. Water Resources Research 33, 2897–2908. doi:10.1029/97WR02174

57

Grelle, A., Lindroth, A., Mölder, M., 1999. Seasonal variation of boreal forest surface conductance and evaporation. Agricultural and Forest Meteorology 98–99, 563–578. doi:10.1016/S0168-1923(99)00124-0

Grip, H., 2015. Sweden’s first forest hydrology field study 1905-1926: contemporary relevance of inherited conclusions and data from the Rokliden Hillslope. Hydrological Processes 29, 3616–3631. doi:10.1002/hyp.10420

Grip, H., Bishop, K.H., 1990. Chemical dynamics of an acid stream rich in dissolved organics, in: Mason, B.J. (Ed.), The Surface Water Acidification Programme. Royal Society, London, pp. 75–84.

Grip, H., Rodhe, A., 1985. Vattnets väg från regn till bäck. Forskningsrådens förlagstjänst, Stockholm.

Harman, C.J., Sivapalan, M., Kumar, P., 2009. Power law catchment-scale recessions arising from heterogeneous linear small-scale dynamics. Water Resources Research 45, W09404. doi:10.1029/2008WR007392

Hayashi, M., Quinton, W.L., Pietroniro, A., Gibson, J.J., 2004. Hydrologic functions of wetlands in a discontinuous permafrost basin indicated by isotopic and chemical signatures. Journal of Hydrology 296, 81–97. doi:10.1016/j.jhydrol.2004.03.020

Hewlett, J.D., Hibbert, A.R., 1967. Factors affecting the response of small watersheds to precipitation in humid areas, in: Sopper, W.E., Lull, H.W. (Eds.), International Symposium on Forest Hydrology. Pergamon, New York, pp. 275–290.

Hoover, M.D., Hursh, C.R., 1943. Influence of topography and soil-depth on runoff from forest land. Transactions, American Geophysical Union 24, 693–697. doi:10.1029/TR024i002p00693

Hrachowitz, M., Savenije, H.H.G., Blöschl, G., McDonnell, J.J., Sivapalan, M., Pomeroy, J.W., Arheimer, B., Blume, T., Clark, M.P., Ehret, U., Fenicia, F., Freer, J.E., Gelfan, A., Gupta, H.V., Hughes, D.A., Hut, R.W., Montanari, A., Pande, S., Tetzlaff, D., Troch, P.A., Uhlenbrook, S., Wagener, T., Winsemius, H.C., Woods, R.A., Zehe, E., Cudennec, C., 2013. A decade of Predictions in Ungauged Basins (PUB)—a review. Hydrological Sciences Journal 58, 1198–1255. doi:10.1080/02626667.2013.803183

Humphreys, E.R., Lafleur, P.M., Flanagan, L.B., Hedstrom, N., Syed, K.H., Glenn, A.J., Granger, R., 2006. Summer carbon dioxide and water vapor fluxes across a range of northern peatlands. Journal of Geophysical Research: Biogeosciences 111, G04011. doi:10.1029/2005JG000111

Ide, J., Finér, L., Laurén, A., Piirainen, S., Launiainen, S., 2013. Effects of clear-cutting on annual and seasonal runoff from a boreal forest catchment in eastern Finland. Forest Ecology and Management 304, 482–491. doi:10.1016/j.foreco.2013.05.051

Jencso, K.G., McGlynn, B.L., 2011. Hierarchical controls on runoff generation: Topographically driven hydrologic connectivity, geology, and vegetation. Water Resources Research 47, W11527.

58

doi:10.1029/2011WR010666 Jónsdóttir, J.F., Uvo, C.B., Clarke, R.T., 2008. Filling gaps in measured

discharge series with model-generated series. Journal of Hydrologic Engineering 13, 905–909. doi:10.1061/(ASCE)1084-0699(2008)13:9(905)

Kasurinen, V., Alfredsen, K., Kolari, P., Mammarella, I., Alekseychik, P., Rinne, J., Vesala, T., Bernier, P., Boike, J., Langer, M., Belelli Marchesini, L., van Huissteden, K., Dolman, H., Sachs, T., Ohta, T., Varlagin, A., Rocha, A., Arain, A., Oechel, W., Lund, M., Grelle, A., Lindroth, A., Black, A., Aurela, M., Laurila, T., Lohila, A., Berninger, F., 2014. Latent heat exchange in the boreal and arctic biomes. Global Change Biology 20, 3439–3456. doi:10.1111/gcb.12640

Kellner, E., 2001. Surface energy fluxes and control of evapotranspiration from a Swedish Sphagnum mire. Agricultural and Forest Meteorology 110, 101–123. doi:10.1016/S0168-1923(01)00283-0

Kirchner, J.W., 2009. Catchments as simple dynamical systems: Catchment characterization, rainfall-runoff modeling, and doing hydrology backward. Water Resources Research 45, 1–34. doi:10.1029/2008WR006912

Kirnbauer, R., Blöschl, G., Haas, P., Müller, G., Merz, B., 2005. Identifying Space-time Patterns of Runoff Generation: A Case Study from the Löhnersbach Catchment, Austrian Alps, in: Huber, U.M., Bugmann, H.K.M., Reasoner, M.A. (Eds.), Global Change and Mountain Regions. Springer Netherlands, Dordrecht, pp. 309–320. doi:10.1007/1-4020-3508-X_31

Knutsson, G., Fagerlind, T., 1977. Grundvattentillgångar i Sverige. Report 9, Swedish Geological Survey, Stockholm.

Krakauer, N.Y., Temimi, M., 2011. Stream recession curves and storage variability in small watersheds. Hydrology and Earth System Sciences 15, 2377–2389. doi:10.5194/hess-15-2377-2011

Kuraś, P.K., Weiler, M., Alila, Y., 2008. The spatiotemporal variability of runoff generation and groundwater dynamics in a snow-dominated catchment. Journal of Hydrology 352, 50–66. doi:10.1016/j.jhydrol.2007.12.021

Laudon, H., Berggren, M., Ågren, A., Buffam, I., Bishop, K.H., Grabs, T., Jansson, M., Köhler, S., 2011. Patterns and dynamics of dissolved organic carbon (DOC) in boreal streams: The role of processes, connectivity, and scaling. Ecosystems 14, 880–893. doi:10.1007/s10021-011-9452-8

Laudon, H., Ottosson Löfvenius, M., 2016. Adding snow to the picture - providing complementary winter precipitation data to the Krycklan Catchment Study database. Hydrological Processes 30, 2413–2416. doi:10.1002/hyp.10753

Laudon, H., Seibert, J., Köhler, S., Bishop, K.H., 2004. Hydrological flow paths during snowmelt: Congruence between hydrometric measurements and oxygen 18 in meltwater, soil water, and runoff.

59

Water Resources Research 40, 1–9. doi:10.1029/2003WR002455 Laudon, H., Sjoblom, V., Buffam, I., Seibert, J., Morth, M., 2007. The role

of catchment scale and landscape characteristics for runoff generation of boreal streams. Journal of Hydrology 344, 198–209. doi:10.1016/j.jhydrol.2007.07.010

Laudon, H., Taberman, I., Ågren, A., Futter, M., Ottosson-Löfvenius, M., Bishop, K., 2013. The Krycklan Catchment Study-A flagship infrastructure for hydrology, biogeochemistry, and climate research in the boreal landscape. Water Resources Research 49, 7154–7158. doi:10.1002/wrcr.20520

Lidman, F., Köhler, S.J., Mörth, C.M., Laudon, H., 2014. Metal transport in the boreal landscape-the role of wetlands and the affinity for organic matter. Environmental Science and Technology 48, 3783–3790. doi:10.1021/es4045506

Lundberg, A., Koivusalo, H., 2003. Estimating winter evaporation in boreal forests with operational snow course data. Hydrological Processes 17, 1479–1493. doi:10.1002/hyp.1179

Lundin, L., 1982. Mark- och grundvatten i moränmark och marktypens betydelse for avrinningen. PhD thesis, UNGI Report 56, Uppsala Univ., Dept Phys. Geogr., Uppsala, Sweden.

Lyon, S.W., Laudon, H., Seibert, J., Mörth, M., Tetzlaff, D., Bishop, K.H., 2010. Controls on snowmelt water mean transit times in northern boreal catchments. Hydrological Processes 24, 1672–1684. doi:10.1002/hyp.7577

Lyon, S.W., Nathanson, M., Spans, A., Grabs, T., Laudon, H., Temnerud, J., Bishop, K.H., Seibert, J., 2012. Specific discharge variability in a boreal landscape. Water Resources Research 48, W08506. doi:10.1029/2011WR011073

McClain, M.E., Boyer, E.W., Dent, C.L., Gergel, S.E., Grimm, N.B., Groffman, P.M., Hart, S.C., Harvey, J.W., Johnston, C.A., Mayorga, E., McDowell, W.H., Pinay, G., 2003. Biogeochemical Hot Spots and Hot Moments at the Interface of Terrestrial and Aquatic Ecosystems. Ecosystems 6, 301–312. doi:10.1007/s10021-003-0161-9

McDonnell, J.J., Sivapalan, M., Vaché, K., Dunn, S., Grant, G., Haggerty, R., Hinz, C., Hooper, R., Kirchner, J., Roderick, M.L., Selker, J., Weiler, M., 2007. Moving beyond heterogeneity and process complexity: A new vision for watershed hydrology. Water Resources Research 43, W07301. doi:10.1029/2006WR005467

McDonnell, J.J., Woods, R., 2004. On the need for catchment classification. Journal of Hydrology 299, 2–3. doi:10.1016/j.jhydrol.2004.09.003

McGlynn, B., McDonnell, J., Stewart, M., Seibert, J., 2003. On the relationships between catchment scale and streamwater mean residence time. Hydrological Processes 17, 175–181. doi:10.1002/hyp.5085

McGuire, K.J., McDonnell, J.J., Weiler, M., Kendall, C., McGlynn, B.L., Welker, J.M., Seibert, J., 2005. The role of topography on catchment-scale water residence time. Water Resources Research 41, W05002.

60

doi:10.1029/2004WR003657 McMillan, H., Gueguen, M., Grimon, E., Woods, R., Clark, M., Rupp, D.E.,

2014. Spatial variability of hydrological processes and model structure diagnostics in a 50 km2 catchment. Hydrological Processes 28, 4896–4913. doi:10.1002/hyp.9988

Mellander, P.E., Bishop, K., Lundmark, T., 2004. The influence of soil temperature on transpiration: A plot scale manipulation in a young Scots pine stand. Forest Ecology and Management 195, 15–28. doi:10.1016/j.foreco.2004.02.051

Nicolson, J.A., 1988. Water and chemical budgets for terrestrial basins at the Turkey Lakes Watershed. Canadian Journal of Fisheries and Aquatic Sciences 45, 88–95. doi:10.1139/f88-271

Nilsson, M., Sagerfors, J., Buffam, I., Laudon, H., Eriksson, T., Grelle, A., Klemedtsson, L., Weslien, P., Lindroth, A., 2008. Contemporary carbon accumulation in a boreal oligotrophic minerogenic mire - a significant sink after accounting for all C-fluxes. Global Change Biology 14, 2317–2332. doi:10.1111/j.1365-2486.2008.01654.x

Nippgen, F., McGlynn, B.L., Marshall, L.A., Emanuel, R.E., 2011. Landscape structure and climate influences on hydrologic response. Water Resources Research 47, W12528. doi:10.1029/2011WR011161

Nyberg, L., 1995. Water flow path interactions with soil hydraulic properties in till soil at Gårdsjön, Sweden. Journal of Hydrology 170, 255–275. doi:10.1016/0022-1694(94)02667-Z

Nyberg, L., Stähli, M., Mellander, P.E., Bishop, K.H., 2001. Soil frost effects on soil water and runoff dynamics along a boreal forest transect: 1. Field investigations. Hydrological Processes 15, 909–926. doi:10.1002/hyp.256

O’Callaghan, J.F., Mark, D.M., 1984. The extraction of drainage networks from digital elevation data. Computer Vision, Graphics, and Image Processing 27, 323–344. doi:10.1016/S0734-189X(84)80047-X

Patnaik, S., Biswal, B., Kumar, D.N., Sivakumar, B., 2015. Effect of catchment characteristics on the relationship between past discharge and the power law recession coefficient. Journal of Hydrology 528, 321–328. doi:10.1016/j.jhydrol.2015.06.032

Payn, R.A., Gooseff, M.N., McGlynn, B.L., Bencala, K.E., Wondzell, S.M., 2012. Exploring changes in the spatial distribution of stream baseflow generation during a seasonal recession. Water Resources Research 48, 1–15. doi:10.1029/2011WR011552

Payn, R.A., Gooseff, M.N., McGlynn, B.L., Bencala, K.E., Wondzell, S.M., 2009. Channel water balance and exchange with subsurface flow along a mountain headwater stream in Montana, United States. Water Resources Research 45. doi:10.1029/2008WR007644

Peichl, M., Sagerfors, J., Lindroth, A., Buffam, I., Grelle, A., Klemedtsson, L., Laudon, H., Nilsson, M.B., 2013. Energy exchange and water budget partitioning in a boreal minerogenic mire. Journal of Geophysical Research: Biogeosciences 118, 1–13.

61

doi:10.1029/2012JG002073 Peralta-Tapia, A., Sponseller, R.A., Ågren, A., Tetzlaff, D., Soulsby, C.,

Laudon, H., 2015a. Scale-dependent groundwater contributions influence patterns of winter baseflow stream chemistry in boreal catchments. Journal of Geophysical Research: Biogeosciences 120, 847–858. doi:10.1002/2014JG002878

Peralta-Tapia, A., Sponseller, R.A., Tetzlaff, D., Soulsby, C., Laudon, H., 2015b. Connecting precipitation inputs and soil flow pathways to stream water in contrasting boreal catchments. Hydrological Processes 29, 3546–3555. doi:10.1002/hyp.10300

Pinay, G., Peiffer, S., De Dreuzy, J.R., Krause, S., Hannah, D.M., Fleckenstein, J.H., Sebilo, M., Bishop, K., Hubert-Moy, L., 2015. Upscaling nitrogen removal capacity from local hotspots to low stream orders’ drainage basins. Ecosystems 18, 1101–1120. doi:10.1007/s10021-015-9878-5

Pomeroy, J.W., Schmidt, R.A., 1993. The use of fractal geometry in modelling intercepted snow accumulation and sublimation, in: 50th Eastern Snow Conference 8-10 June 1993. Quebec City, pp. 1–10.

Prepas, E.E., Burke, J.M., Whitson, I.R., Putz, G., Smith, D.W., 2006. Associations between watershed characteristics, runoff, and stream water quality: hypothesis development for watershed disturbance experiments and modelling in the Forest Watershed and Riparian Disturbance (FORWARD) project. Journal of Environmental Engineering and Science 5, S27–S37. doi:10.1139/s05-033

Quinton, W.L., Hayashi, M., Pietroniro, A., 2003. Connectivity and storage functions of channel fens and flat bogs in northern basins. Hydrological Processes 17, 3665–3684. doi:10.1002/hyp.1369

Rodhe, A., 1989. On the generation of stream runoff in till soils. Nordic Hydrology 20, 1–8.

Rodhe, A., 1987. The origin of streamwater traced by oxygen-18. PhD thesis, UNGI Report Series A no. 41, Uppsala Univ., Dept Phys. Geogr., Uppsala, Sweden.

Rodhe, A., 1981. Spring flood: meltwater or groundwater? Nordic Hydrology 12, 21–30.

Rosén, K., 1984. Effect of clear-felling on runoff in two small watersheds in central Sweden. Forest Ecology and Management 9, 267–281.

Sayama, T., McDonnell, J.J., Dhakal, A., Sullivan, K., 2011. How much water can a watershed store? Hydrological Processes 25, 3899–3908. doi:10.1002/hyp.8288

Scharlemann, J.P., Tanner, E.V., Hiederer, R., Kapos, V., 2014. Global soil carbon: understanding and managing the largest terrestrial carbon pool. Carbon Management 5, 81–91. doi:10.4155/cmt.13.77

Schelker, J., Kuglerová, L., Eklöf, K., Bishop, K.H., Laudon, H., 2013. Hydrological effects of clear-cutting in a boreal forest – Snowpack dynamics, snowmelt and streamflow responses. Journal of Hydrology 484, 105–114. doi:10.1016/j.jhydrol.2013.01.015

62

Seibert, J., 2000. Multi-criteria calibration of a conceptual runoff model using a genetic algorithm. Hydrology and Earth System Sciences 4, 215–224. doi:10.5194/hess-4-215-2000

Seibert, J., Bishop, K., Rodhe, A., McDonnell, J.J., 2003. Groundwater dynamics along a hillslope: A test of the steady state hypothesis. Water Resources Research 39, 1014. doi:10.1029/2002WR001404

Seibert, J., McGlynn, B.L., 2007. A new triangular multiple flow direction algorithm for computing upslope areas from gridded digital elevation models. Water Resources Research 43, 1–8. doi:10.1029/2006WR005128

Seibert, J., Vis, M.J.P., 2012. Teaching hydrological modeling with a user-friendly catchment-runoff-model software package. Hydrology and Earth System Sciences 16, 3315–3325. doi:10.5194/hess-16-3315-2012

Shaman, J., Stieglitz, M., Burns, D., 2004. Are big basins just the sum of small catchments? Hydrological Processes 18, 3195–3206. doi:10.1002/hyp.5739

Shaw, S.B., 2015. Investigating the linkage between streamflow recession rates and channel network contraction in a mesoscale catchment in New York state. Hydrological Processes 492, 479–492. doi:10.1002/hyp.10626

Shaw, S.B., Riha, S.J., 2012. Examining individual recession events instead of a data cloud: Using a modified interpretation of dQ/dt–Q streamflow recession in glaciated watersheds to better inform models of low flow. Journal of Hydrology 434–435, 46–54. doi:10.1016/j.jhydrol.2012.02.034

Sivapalan, M., 2005. Pattern, process and function: Elements of a unified theory of hydrology at the catchment scale, in: Anderson, M.G., McDonnell, J.J. (Eds.), Encyclopedia of Hydrological Sciences. John Wiley & Sons, Ltd, New York, pp. 193–219. doi:10.1002/0470848944

Sivapalan, M., Takeuchi, K., Franks, S.W., Gupta, V.K., Karambiri, H., Lakshmi, V., Liang, X., McDonnell, J.J., Mendiondo, E.M., O’Connell, P.E., Oki, T., Pomeroy, J.W., Schertzer, D., Uhlenbrook, S., Zehe, E., 2003. IAHS Decade on Predictions in Ungauged Basins (PUB), 2003–2012: Shaping an exciting future for the hydrological sciences. Hydrological Sciences Journal 48, 857–880. doi:10.1623/hysj.48.6.857.51421

Sørensen, R., Ring, E., Meili, M., Högbom, L., Seibert, J., Grabs, T., Laudon, H., Bishop, K., 2009. Forest Harvest Increases Runoff Most during Low Flows in Two Boreal Streams. AMBIO: A Journal of the Human Environment 38, 357–363. doi:10.1579/0044-7447-38.7.357

Soulsby, C., Tetzlaff, D., Rodgers, P., Dunn, S., Waldron, S., 2006. Runoff processes, stream water residence times and controlling landscape characteristics in a mesoscale catchment: An initial evaluation. Journal of Hydrology 325, 197–221. doi:10.1016/j.jhydrol.2005.10.024

Spearman, C., 1904. The proof and measurement of association between two things. The American journal of psychology 15, 72–101.

63

Szilagyi, J., Gribovszki, Z., Kalicz, P., 2007. Estimation of catchment-scale evapotranspiration from baseflow recession data: Numerical model and practical application results. Journal of Hydrology 336, 206–217. doi:10.1016/j.jhydrol.2007.01.004

Tague, C., Grant, G.E., 2004. A geological framework for interpreting the low-flow regimes of Cascade streams, Willamette River Basin, Oregon. Water Resources Research 40, W04303. doi:10.1029/2003WR002629

Tallaksen, L., 1995. A review of baseflow recession analysis. Journal of Hydrology 165, 349–370. doi:10.1016/0022-1694(94)02540-R

Temnerud, J., Seibert, J., Jansson, M., Bishop, K.H., 2007. Spatial variation in discharge and concentrations of organic carbon in a catchment network of boreal streams in northern Sweden. Journal of Hydrology 342, 72–87. doi:10.1016/j.jhydrol.2007.05.015

Tetzlaff, D., Seibert, J., Soulsby, C., 2009. Inter-catchment comparison to assess the influence of topography and soils on catchment transit times in a geomorphic province; the Cairngorm mountains, Scotland. Hydrological Processes 23, 1874–1886. doi:10.1002/hyp.7318

Tetzlaff, D., Soulsby, C., Youngson, A.F., Gibbins, C., Bacon, P.J., Malcolm, I.A., Langan, S., 2005. Variability in stream discharge and temperature: a preliminary assessment of the implications for juvenile and spawning Atlantic salmon. Hydrology and Earth System Sciences 9, 193–208. doi:10.5194/hess-9-193-2005

Teuling, A.J., Lehner, I., Kirchner, J.W., Seneviratne, S.I., 2010. Catchments as simple dynamical systems: Experience from a Swiss prealpine catchment. Water Resources Research 46, 1–15. doi:10.1029/2009WR008777

Tiwari, T., Laudon, H., Beven, K., Ågren, A.M., 2014. Downstream changes in DOC: Inferring contributions in the face of model uncertainties. Water Resources Research 50, 514–525. doi:10.1002/2013WR014275

Tranvik, L.J., Downing, J.A., Cotner, J.B., Loiselle, S.A., Striegl, R.G., Ballatore, T.J., Dillon, P., Finlay, K., Fortino, K., Knoll, L.B., Kortelainen, P.L., Kutser, T., Larsen, S., Laurion, I., Leech, D.M., McCallister, S.L., McKnight, D.M., Melack, J.M., Overholt, E., Porter, J.A., Prairie, Y., Renwick, W.H., Roland, F., Sherman, B.S., Schindler, D.W., Sobek, S., Tremblay, A., Vanni, M.J., Verschoor, A.M., von Wachenfeldt, E., Weyhenmeyer, G.A., 2009. Lakes and reservoirs as regulators of carbon cycling and climate. Limnology and Oceanography 54, 2298–2314. doi:10.4319/lo.2009.54.6_part_2.2298

van der Velde, Y., Lyon, S.W., Destouni, G., 2013. Data-driven regionalization of river discharges and emergent land cover-evapotranspiration relationships across Sweden. Journal of Geophysical Research: Atmospheres 118, 2576–2587. doi:10.1002/jgrd.50224

Vedin, H., 1995. Lufttemperatur, årstiderna, in: Raab, B., Vedin, H. (Eds.), Sveriges Nationalatlas: Klimat, Sjöar Och Vattendrag. Bra Böcker, Stockholm, p. 176.

Wagener, T., Sivapalan, M., Troch, P., Woods, R., 2007. Catchment

64

classification and hydrologic similarity. Geography Compass 1, 901–931. doi:10.1111/j.1749-8198.2007.00039.x

Wallin, M.B., Grabs, T., Buffam, I., Laudon, H., Ågren, A., Öquist, M.G., Bishop, K., 2013. Evasion of CO 2 from streams - The dominant component of the carbon export through the aquatic conduit in a boreal landscape. Global Change Biology 19, 785–797. doi:10.1111/gcb.12083

Woods, R., 2005. Hydrologic Concepts of Variability and Scale, in: Anderson, M.G., McDonnell, J.J. (Eds.), Encyclopedia of Hydrological Sciences. John Wiley & Sons, Ltd, New York, pp. 23–40. doi:10.1002/0470848944

Yanai, R.D., Tokuchi, N., Campbell, J.L., Green, M.B., Matsuzaki, E., Laseter, S.N., Brown, C.L., Bailey, A.S., Lyons, P., Levine, C.R., Buso, D.C., Likens, G.E., Knoepp, J.D., Fukushima, K., 2015. Sources of uncertainty in estimating stream solute export from headwater catchments at three sites. Hydrological Processes 29, 1793–1805. doi:10.1002/hyp.10265

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology 1420

Editor: The Dean of the Faculty of Science and Technology

A doctoral dissertation from the Faculty of Science andTechnology, Uppsala University, is usually a summary of anumber of papers. A few copies of the complete dissertationare kept at major Swedish research libraries, while thesummary alone is distributed internationally throughthe series Digital Comprehensive Summaries of UppsalaDissertations from the Faculty of Science and Technology.(Prior to January, 2005, the series was published under thetitle “Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Science and Technology”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-302400

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2016