1

Needs and means to advance science, policy and management

understanding of the freshwater system – A synthesis report

Georgia Destouni1*, Shilpa M. Asokan1, Anna Augustsson2, Berit Balfors3, Arvid Bring1, Fernando Jaramillo1, Jerker Jarsjö1, Emma Johansson1,4, John Juston1, Lea Levi1,3,5, Bo Olofsson3, Carmen Prieto1, Andrew Quin1, Mats Åström2 &Vladimir Cvetkovic3 1Department of Physical Geography, Bolin Centre for Climate Research Stockholm University SE-106 91 Stockholm Sweden 2Department of Biology and Environmental Science Linnaeus University SE-39182 Kalmar Sweden 3Department of Sustainable Development, Environmental Science and Engineering The Royal Institute of Technology SE-100 44, Stockholm Sweden 4Swedish Nuclear Fuel and Waste Management Co Box 250101 24 Stockholm Sweden 5Department of Applied Hydraulics Faculty of Civil Engineering, Architecture and Geodesy University of Split 21 000 Split Croatia * Corresponding author Telephone: +46 8 164785 Mobile phone: +46 704532366 E-mail: georgia.destouni@natgeo.su.se

Citation: Destouni G, Asokan SM, Augustsson A, Balfors B, Bring A, Jaramillo F, Jarsjö J, Johansson E, Juston J, Levi L, Olofsson B, Prieto C, Quin A, Åström M, Cvetkovic V (2015). Needs and means to advance science, policy and management understanding of the freshwater system – A synthesis report. Research project: Climate-land-water changes and integrated water resource management in coastal regions (KLIV), of Stockholm University, The Royal Institute of Technology and Linnaeus University, Sweden, http://urn.kb.se/resolve?urn=urn:nbn:se:su:diva-117549.

2

Abstract Fragmented and inconsistent understanding of the freshwater system limits our ability to achieve water security and sustainability under the human-driven changes occurring in the Anthropocene. To advance system-level understanding of freshwater, gaps and inconsistencies in knowledge, data, representations and links of processes and subsystems need to be identified and bridged under consideration of the freshwater system as a continuous whole. Based on such identification, a freshwater system conceptualization is developed in this report, which emphasizes four essential, yet often neglected system aspects:

i) Distinction of coastal divergent catchments.

ii) Four main zones (surface, subsurface, coastal, observation) of different types of freshwater change.

iii) Water pathways as system-coupling agents that link and partition water change

among the four change zones.

iv) Direct interactions with the anthroposphere as integral system pathways across the change zones.

We explain and exemplify some key implications of these aspects, identifying in the process also distinct patterns of human-driven changes in large-scale water fluxes and nutrient loads. The present conceptualization provides a basis for common inter- and trans-disciplinary understanding and systematic characterization of the freshwater system function and its changes, and of approaches to their modeling and monitoring. This can be viewed and used as a unifying checklist that can advance science, policy and management of freshwater and related environmental changes across various scales and world regions. Keywords: Freshwater system; Social-ecological system; Catchment; Human-driven change; Change propagation; Water pathways; Solute travel times; Water fluxes; Nutrient loads.

3

1. Introduction For freshwater security, management and adaptation in the Anthropocene epoch (Crutzen and Stoermer, 2000 and Crutzen, 2002), it is increasingly important to distinguish main drivers of change in freshwater availability (Milly et al., 2005, Destouni et al., 2013, Berghuis et al., 2014 and Jaramillo et al., 2014), freshwater quality (Banks et al., 1997, Darracq et al., 2008, Fick et al., 2009, Fergusson and Gleeson, 2012 and Åkesson et al., 2014) and related risks of drought, flooding, and human and ecological health (Persson and Destouni, 2009, Vörösmarty et al., 2010, Karlsson et al., 2011, Törnqvist et al., 2011, Ledger et al., 2013, Nilsson et al., 2013 and Destouni and Verrot, 2014). It is also increasingly important to be able to follow and project the evolution and impacts of such changes as they propagate in space and time with the ever-flowing water along its various flow pathways (Baresel and Destouni, 2005, Destouni et al., 2010a, Cvetkovic et al., 2012, Jarsjö et al., 2012, Berghuis et al., 2013, Botter et al., 2013, Destouni et al., 2013, Montanari et al., 2013 and Kiel and Cardenas, 2014). However, our ability to meet these challenges is limited by fragmented and inconsistent understanding, monitoring and handling of the freshwater system across science disciplines, policy and management. Even fundamental closure of water and constituent mass balances may still constitute a major challenge for freshwater systems in the landscape (e.g. Beven, 2006), and there are calls for new unifying theory (e.g. Sivapalan, 2005), paradigm shift (McDonnell et al., 2007) and new catchment conceptualization (Tetzlaff et al., 2008) in order to advance system understanding and be able to meet major challenges. To achieve this, the advancement of freshwater system understanding needs to be inter- and trans-disciplinary, across science, policy and management (Sivapalan et al., 2012). However, in contrast to such unification needs and calls, different water-related disciplines and competent authorities still mostly study and manage some limited selection from all existing physical freshwater components and prevailing societal freshwater uses and effects. Physical components include soil water, groundwater, lakes, wetlands, individual streams and their hyporheic water, stream networks, and engineered water systems. Societal sectors that use and affect freshwater include agriculture, industry, households, and utilities for energy, water and sanitation. The various freshwater components and societal uses and effects are all connected in the landscape, hydrologically (through water and constituent mass balances) and hydraulically (through water pressure, momentum and pathway continuity across component interfaces). Their disconnect in science, management and policy is thus artificial and may lead to development of crucial gaps in the data, knowledge, measures and policy needed to secure freshwater security and sustainability (Hannerz et al., 2006, Destouni et al., 2008, Bring and Destouni, 2009, McDonnell et al., 2010, Karlsson et al., 2011, Prieto and Destouni, 2011, Azcárate et al., 2013, Nilsson et al., 2013 and Åkesson et al., 2014). As one example of fragmented freshwater management and policy, consider the case of Sweden. Good knowledge base and extensive long-term data are well developed for Swedish management of some freshwater aspects, such as surface water flows, nutrient concentrations in surface water and change-driving conditions of atmospheric climate change (Basu et al., 2010, Destouni et al., 2013 and van der Velde at al., 2013). Overall, however, management and policies relating to freshwater are highly fragmented among competent authorities (Government Offices of Sweden, 2014). For example, the Swedish Agency for Marine and Water Management is responsible for surface water, while the Geological Survey of Sweden is responsible for groundwater, and the Swedish Environmental Protection Agency is

4

responsible for water-affecting land management. Responsibilities for national freshwater monitoring are also divided among yet another set of organizations, for example: the Swedish Meteorological and Hydrological Institute for surface water flows, the Swedish University of Agricultural Sciences for surface water quality, and the Geological Survey of Sweden for groundwater flow and quality. Furthermore, the Ministry of the Environment, with overall responsibility for water and the seas, has charge of only the surface-water focused competent authorities: the Swedish Agency for Marine and Water Management, the Swedish Meteorological and Hydrological Institute, and the Swedish Environmental Protection Agency. For the activities of the Geological Survey of Sweden, it is instead the Ministry of Enterprise, Energy and Communications, with regional growth as a main area of responsibility, which has overall charge. As discussed above more generally, the fragmentation of water-relevant policy and management in Sweden does indeed create an artificial freshwater disconnect with crucial data and knowledge gap implications, especially regarding groundwater quality, variability and change and for linking these to surface and coastal water quality problems and their possible solutions (Baresel et al., 2006, Destouni et al., 2008 and 2010a, Baresel and Destouni, 2009 and Åkesson et al., 2014). The fast-changing Arctic region represents another example of evident freshwater policy and management gaps. These regard, for instance, the application of strategic environmental assessment to Arctic freshwater changes (Azcárate et al., 2013), and the monitoring of waterborne nutrient and pollutant spreading (Bring and Destouni, 2009), of water security indicators (Nilsson et al., 2013), and of climate change (Bring and Destouni, 2013) and ecosystem shifts (Karlsson et al., 2011) linked to freshwater change in the Arctic. In this region, as in the above national example of Sweden, these gaps exist in contrast to good monitoring coverage of some water aspects, such as Arctic river discharges to the Arctic Ocean (Arctic-HYDRA consortium, 2010). For some identified small hotspot areas in the Arctic, sufficient river discharge information is available in relevant combination with water-related data on permafrost degradation and ecosystem regime shifts (Karlsson et al., 2011). Study of such hotspot areas indicates a high knowledge value that may be gained by improved cross-disciplinary understanding of the freshwater system based on combined water-related datasets (Karlsson et al., 2011 and 2012 and Törnqvist et al., 2014). In contrast, lack and inconsistencies of essential freshwater-related data for large parts of the Arctic may mislead interpretations of the Arctic freshwater system, its changes and their drivers and impacts (Bring and Destouni, 2009, 2011, 2013 and 2014, Karlsson et al., 2011, Azcárate et al., 2013 and Nilsson et al., 2013). The present report summarizes and synthesizes the identifications and bridging contributions of co-authors with regard to key freshwater gaps and inconsistencies, which may be crucial for system-level understanding and management of the function and changes of the freshwater system across scales and in different world regions. As a basis for this synthesis, we first exemplify and structure some freshwater representation gaps and inconsistencies that exist in traditional water cycle conceptualizations, as these may fundamentally underlie fragmented views of the freshwater system among science disciplines and societal sectors. Furthermore, with regard to bridging gaps and inconsistencies, we use the present synthesis to develop and propose a unifying freshwater system conceptualization, which considers, addresses and visualizes this system as an integrated continuous whole. We finally explain some key new aspects of this conceptualization and exemplify some main implications of these aspects by synthesis and meta-analysis of some relevant literature results.

5

2. Freshwater system representation and conceptualization We discuss here some examples of freshwater representations in traditional and fundamental water cycle conceptualizations considered within advanced Earth system science literature (Jacobson et al., 2000). These are:

- The global water balance conceptualization (and quantification) of Shiklomanov and Sokolov (1983; Figure 6.3 in Jacobson et al., 2000).

- The process conceptualization of the global hydrological cycle of Horton (1931; Figure 6.4 in Jacobson et al., 2000)

- The surface process conceptualization with focus on evaporation and transpiration

from vegetation of Dickinson (1984; Figure 6.5 in Jacobson et al., 2000).

- The conceptualization of main cyclic processes and fluxes between major Earth reservoirs of Jacobson et al. (2000; their Figure 1.1).

- The conceptualization (and quantification) of processes connecting global

biogeochemical cycles and climate of Jacobson et al. (2000; their Fig. 19-1).

- The conceptualization of flux changes relative to pre-industrial time for radiatively important trace species in the atmosphere of Jacobson et al. (2000; their Table 19-1).

As freshwater gaps and inconsistencies among these examples, we consider key freshwater aspects that are not considered or shown, or that are considered, shown or interpreted differently across these fundamental conceptualizations; see further information and details in Appendix: Supplementary Material (SM) Section SM1. Based on such freshwater gaps found in traditional water cycle conceptualizations, we develop and propose a new conceptualization, focusing on the freshwater system as a continuous whole. To bridge key freshwater gaps and inconsistencies, we believe that freshwater must be accounted for as a continuous system, on par and interacting as a whole with other key segments of the Earth System (anthroposphere-technosphere, atmosphere, biosphere, cryosphere, lithosphere-pedosphere, marine hydrosphere; the anthroposphere-technosphere and lithosphere-pedosphere will in the following be simply referred to as anthroposphere and lithosphere, respectively). A conceptualization of the freshwater system should then not be confused with any particular quantification model of this system. Several hydrological, climate and other types of quantification models exist, which may include and account for multiple components and process interactions of the freshwater system. Each model thus represents certain quantification potential for these components and interactions, as accounted for in model equations and parameterizations. Modelers choose what part of that potential to actually realize in specific model applications, and model documentation may describe in some detail some or all of the freshwater components and interactions that each model incorporates. However, even the latter does not typically describe all of the freshwater features included in a model as parts of a continuous whole freshwater system. The present conceptualization aims to provide a complete freshwater system representation,

6

which may be thought of as a visualized checklist of the main freshwater components and interaction pathways that couple this system into an integrated whole, link it with other key segments of the Earth System. This aim is analogous to that of traditional water cycle conceptualizations discussed above, which also provide some complete system representation, however not of the freshwater system, but rather of the Earth’s total water cycle or some selected parts of it within other Earth segments. To complement these, the present conceptualization does focus on the freshwater system and its continuity and ability to propagate environmental change through the whole system and from/to other Earth System segments. The freshwater system integration and focus are important because the water flow and waterborne propagation of environmental change can transform water availability and quality as well as water-related human, societal and ecosystem risks over a wide range of spatiotemporal scales. We propose that consideration of change propagation through the whole freshwater system is needed to advance a common understanding among water scientists, managers and policy makers of multiple freshwater changes occurring in the Anthropocene. With a common system-understanding basis, key gaps that do exist in freshwater data, knowledge, policy and management should become more clearly and consistently identifiable by different actors, who can then act more in concert toward bridging the gaps. To aid in such gap-bridging development, we depart here from the proposed conceptualization to more closely explain some essential yet typically neglected aspects of the freshwater system. By synthesis and meta-analysis of relevant literature results we also quantitatively exemplify some key implications of these aspects, in particular regarding changes in one freshwater component that are further propagated across the whole system, on wide-ranging temporal and spatial scales. Where relevant, the sources of data and methods used for these quantitative examples are given in associated result figure captions, with Appendix: Supplementary Material Section SM2 and Tables SM.1-3 providing further information and details.

7

3. Examples of traditional water cycle conceptualizations Through analysis of freshwater representations in traditional water cycle conceptualizations we identify three key aspects that these generally do not consider or show for the freshwater system. (1) First, they do not explicitly show the interactions of freshwater with the anthroposphere, even though freshwater is deeply embedded in and interacts with all human activities and socio-economic sectors (Sivapalan et al., 2012 and Montanari et al., 2013). (2) Second, they also do not show the fundamental catchment-wise organization of freshwater fluxes, even though catchments are recommended or even mandated spatial units for water resource management under global change (Pahl-Wostl, 2007 and UNECE, 2009) and the topographical integration of water and constituent fluxes in hydrological catchments offers substantial advantage for water balance closure (Karlsson et al., 2012, Destouni et al., 2013, van der Velde et al., 2013, Törnqvist et al., 2014 and Jaramillo and Destouni, 2014). (3) Third, they do not visualize the interactions between freshwater and seawater, which are also typically also not monitored as part of common water and environmental management (Destouni et al., 2008) and thereby subject to particularly large uncertainties (Prieto and Destouni, 2011) while threatening water availability, quality and security (Boman et al., 2010, Ferguson and Gleeson, 2012 and Mazi et al., 2013) for the many people living in the particularly densely populated coastal parts of the world (Brown et al., 2002). Furthermore, traditional water cycle conceptualizations inconsistently represent and visualize the following two types of freshwater system interactions: (4) Surface and subsurface water interactions. (5) Subsurface water interactions with the lithosphere. Even though the latter, lithosphere, is not in itself a freshwater component, the subsurface water is embedded in it, and the continuous interactions and feedbacks between water and the solid matrix and its biological material are crucial for water flow and waterborne pollutant and biogeochemical transport toward downgradient subsurface, surface and coastal waters (Cole et al., 2007, Battin et al., 2009, Basu et al., 2010, Destouni et al., 2010a, Lyon et al., 2010, Cvetkovic et al., 2012 and Jantze et al., 2013). Overall, such fundamental freshwater representation omissions and inconsistencies lead to contradictory interpretations of the freshwater system across different disciplinary and environmental perspectives, as discussed further and in more detail in SM Section SM1.

8

4 Focused conceptualization of the freshwater system The present conceptualization shows the freshwater system as a catchment-wise organized, yet scale-free system (Figure 1), emphasizing four essential system aspects that are typically neglected in traditional water cycle conceptualizations:

(i) A distinction is made of coastal divergent catchments, in addition to traditional convergent hydrological catchments.

(ii) Four main zones of different types of water change are emphasized: the land surface (S), the subsurface (SS), the coastal interface (C), the human observation and monitoring system (O).

(iii) The pathways of water flow and waterborne material and energy transport through

the four change zones couple the freshwater system.

(iv) Multiple interactions with the anthroposphere as integral pathways and parts of the freshwater system.

In the following sections, the four aspects (i)-(iv) are further explained and exemplified.

Figure 1. Schematic conceptualization of the continuous freshwater system. The pathways of water flow and waterborne material and energy transport (arrows) physically couple the freshwater system, and are catchment-wise organized in two types of catchments: convergent (light gray) and divergent (dark gray). The water pathways interact and are partitioned in four main zones of water change (schematized as blue filled circles, marked with: S-surface, SS-subsurface, C-coast, O-observation). Developed from a similar conceptualization reported by Destouni et al. (2014).

9

5 Distinction of two catchment types (aspect i) Two types of catchments are distinguished by the human choices of hydrological observation and monitoring systems. These are:

• The traditional hydrological catchments (light gray, Figure 1), which are convergent as they topographically channel and drain their water through distinct outlet/monitoring points.

• The remaining land areas (dark gray, Figure 1), which are divergent catchments as they drain their typically unmonitored water through extended coastline stretches.

A main novel aspect pointed out by this distinction is that of divergent catchments. Even though relatively small in area, coastal divergent catchments are often particularly densely populated (Brown et al., 2002). The freshwater of these catchments therefore receives and propagates disproportionally large impacts of human-driven change, such as nutrient and pollutant inputs to the catchment that propagate to loads into the sea (Destouni et al., 2008). Furthermore, human drawdown of groundwater level and climate-driven rise of sea level combine in enhancing seawater intrusion into coastal groundwater (Ferguson and Gleeson, 2012 and Mazi et al., 2013), which in turn enhances the recirculating seawater component of submarine groundwater discharge and associated transport of land-originated pollutants to the sea (Destouni and Prieto, 2003 and Destouni et al., 2008). Human drawdown of coastal groundwater also activates oxidation of metal sulphides inherent in soil (Burton et al., 2009 and Boman et al., 2010), which may cause waterborne acidification and heavy-metal pollution of both inland and coastal waters (Sundström et al., 2002). With regard to understanding and controlling these coastal zone processes and their changes, seawater intrusion may be relatively well understood and investigated locally, but system-level links to other water and environmental changes in coastal divergent catchments of different scales are not. The particularly high population density and multiple human pressures that prevail in coastal divergent catchments imply a key need to consider, monitor and account for the freshwater changes in these catchments. However, the coastal divergent catchments are typically not distinguished and their contributions of water flow and waterborne constituent transport to the sea are often neglected in water monitoring (Hannerz and Destouni, 2006, Destouni et al., 2008, Prieto and Destouni, 2011 and Bring and Destouni, 2009 and 2014). Furthermore, a key point of the catchment-wise conceptualization of the freshwater system (Figure 1) is that it is scale-free. This is so because convergent catchments (subcatchments, watersheds, river basins, drainage basins, in nested, serial or parallel combinations) can be defined for any land point and associated catchment scale of interest. Relevant convergent-divergent catchment combinations can also be defined for any coastline of interest. Using catchments as spatial units for analyzing and discretizing water and environmental change analysis is advantageous because water pathways are indeed catchment-wise organized, and this organization implies that there is fundamental closure of water and constituent flux balances over catchments (Darracq et al., 2008, Lyon et al., 2010, Karlsson et al., 2012, Destouni et al., 2013, van der Velde et al., 2013, Törnqvist et al., 2014 and Jaramillo and Destouni, 2014). In addition, catchments are also the recommended or even mandated spatial units for modern water resource management (Pahl-Wostl, 2007 and UNECE, 2009).

10

Furthermore, the catchment-wise conceptualization of the freshwater system (Figure 1) is general because it includes all main types of intra- and inter-system pathways of water, in addition to being scale-free. This generalization provides a common basis for identifying key differences between specific catchments of different scales and in different parts of the world. Not least, specific catchments will exhibit different relative pathway strength at each of the four change zones (e.g., of freshwater interaction with the anthroposphere or cryosphere). Such differences can further be used for systematic characterization and typology of freshwater system function and its change in different geographic locations, regions and continents, as discussed further in the following subsection.

11

6 Four main change zones (aspect ii) The freshwater conceptualization in Figure 1 distinguishes four main zones of different types of freshwater change: A) The land surface (S) where liquid and frozen freshwater on land interacts with water and vapor in the atmosphere. B) The subsurface (SS) where most of the liquid water on land resides and flows as groundwater, which interacts with up-gradient soil water and down-gradient surface and coastal water, as well as with the solid lithosphere and frozen subsurface water (with permafrost being a main component of the land cryosphere). C) The coastal interface (C) where subsurface, surface, liquid and frozen freshwater interacts and mixes with seawater. D) The human observation system (O), where water changes are observed and monitored. Through all of these zones, freshwater also interacts directly with the anthroposphere, and directly or indirectly with the terrestrial and coastal-marine ecosystems that affect or are affected by the freshwater conditions and fluxes. Water pathways in Figure 1 (arrows) visualize the physical coupling of the freshwater system itself, the social coupling of the freshwater system to the anthroposphere (Vörösmarty et al., 2010, Sivapalan et al., 2012 and Montanari et al., 2013) and the ecological coupling of the freshwater system to various parts of the biosphere (Vörösmarty et al., 2010, Karlsson et al., 2011 and Ledger et al., 2013). The pathway interactions with the anthroposphere (red, Figure 1) represent direct human water use (e.g., for irrigation or hydropower) as well as effects on the freshwater system of human use of land and other natural resources. The effects are both on freshwater availability (Jarsjö et al., 2012, Destouni et al., 2013, Seitz et al., 2013 and Jaramillo and Destouni, 2014) and on freshwater quality (Banks et al., 1997, Baresel and Destouni, 2005, Darracq et al., 2008, Fick et al., 2009, Fergusson and Gleeson, 2012 and Åkesson et al., 2014). In general, the pathways of water flow and waterborne material and energy transport propagate environmental change through the freshwater system and to or from other Earth System segments (Persson and Destouni, 2009, Destouni et al., 2010a,b, Vörösmarty et al., 2010, Karlsson et al., 2011, Törnqvist et al., 2011, Ledger et al., 2013, Nilsson et al., 2013 and Destouni and Verrot, 2014). Furthermore, human choices for water observation (red dashed line in Figure 1) do also represent important freshwater interactions with the anthroposphere. These observation choices affect the spatiotemporal coverage of available data for understanding the freshwater system function, its change and its relation with other environmental changes (Destouni et al., 2008, Karlsson et al., 2011 and Bring and Destouni, 2009, 2013 and 2014). Overall, the conceptualization in Figure 1 shows that water and waterborne constituent fluxes through the four change-zones link the water and related environmental changes that occur in the various change zones. Our observation choices further determine what we can see and understand of these changes.

12

Figure 1 further shows that the total water flux and the flux changes that propagate through the freshwater system are partitioned into fractions, which follow the different pathways through each change zone (as exemplified in Table 1). This partitioning also includes fractions of monitored and unmonitored pathways of water, as determined by the human observation system. In general, the variation of water pathway fractions across catchments provides a basis for systematic catchment and freshwater function typology. Such typology may distinguish characteristic flow and transport partitioning behavior in catchments along climate, land-use, water-use, population, socio-economic and other spatiotemporal gradients of interest. It may also be used for comparing and distinguishing the freshwater components and pathway-partitioning possibilities included in different quantification models. Table 1 exemplifies these typology possibilities, with reference to such fraction quantifications reported in the literature for some catchments or catchment conditions. Table 1. Examples of possible catchment and freshwater function typology. The typology exemplification is based on the pathway partitioning at the conceptualized four change zones (Figure 3), with reference to such partitioning quantifications reported in the literature for some catchments or temporal catchment states (catchment conditions at different times). Change zone Type of partitioning Catchments (or temporal catchment states) with

different (or changed) characteristics of: S – land surface Precipitation into

evapotranspiration and runoff

Land cover (van der Velde et al., 2013)

Human land-use and water-use (Destouni et al., 2013)

Anthropogenic disturbance level (Jaramillo and Destouni, 2014)

SS – subsurface Groundwater runoff into stream runoff and submarine groundwater discharge (SGD)

Landscape conditions (Bosson et al., 2012)

Climate conditions (Bosson et al., 2012)

Permafrost conditions (Bosson et al., 2012)

S and SS Fractions of solute flux through various pathways at each change zone

Topography (Darracq et al., 2010)

Soil (Persson et al., 2011)

Permafrost (Bosson et al., 2013)

C – coast Surface and subsurface fractions of nutrient or pollutant load to the sea

Population density (Destouni et al., 2008)

Land-use (Destouni et al., 2008)

Freshwater and re-circulated seawater fractions of total SGD

Human use of coastal groundwater (Destouni and Prieto, 2003)

Seasonal flow variability (Prieto and Destouni, 2005)

Tidal sea-level oscillation (Prieto and Destouni, 2005)

O – observation system

Monitored and unmonitored fractions of freshwater and solute fluxes

Climate change observation (Bring and Destouni, 2011 and 2013)

Observation of ecosystem regime shifts (Karlsson et al., 2011)

Observed landscape or socio-economic conditions (Hannerz and Destouni, 2006 and Bring and Destouni, 2009 and 2014)

13

7 Flow and transport pathways (aspect iii) Figure 1 conceptualizes and visualizes the freshwater system in terms of the scattered pathways of water flow and waterborne transport that extend through and connect the whole system. These pathways emphasize water flow as a key catchment-wise integrator and carrier of various environmental changes. Implicit in the pathway concept is then also an inherent temporal pathway correspondence, which may be expressed in terms of the travel times (or residence times) of water or solute fractions following different pathways through each catchment (Cvetkovic et al., 2012, Darracq et al., 2010 and McDonnell et al., 2010). Figure 2 exemplifies results for such travel times along different pathways through a small coastal divergent catchment. These results have been reported separately for different pathways in studies of the same catchment but in different environmental contexts (Darracq et al., 2010, Persson et al. 2011 and Bosson et al., 2013). They travel time distributions associated with different main pathways through the catchment are here synthesized to show a relatively complete picture of the widely variable time scales involved in water flow and transport through even a small catchment. In combination, Figures 1 and 2 illustrate two complementary aspects of water and constituent pathways: the spatial and the temporal aspect of each type of water pathway. The environmental role of water travel (residence) times is crucial in quantifying the physical propagation rates that compete with a broad range of biogeochemical process rates to control resulting dynamics and impacts of water quality and related environmental changes (Steefel et al., 2005, Destouni et al., 2010a, Darracq et al., 2010 and Persson et al., 2011). Such competition over a wide time spectrum can, among others, explain biogeochemical stationarity found in large-scale nutrient concentrations across different parts of the world (Basu et al., 2010). Specifically, water travel times through streams range over time scales of less than a year (light blue curve, Figure 2), while travel times along heterogeneous groundwater pathways may be several decades long on average (red, green, blue curves, Figure 4). Waterborne tracers, nutrients and pollutants that are carried by and follow the groundwater pathways may accumulate along the subsurface pathways, in immobile water, or sorbed in the solid matrix (soil, rock, fracture filling, sediments) due to both slow physical transport and physical and biogeochemical mass transfer and immobilization-remobilization processes (Steefel et al., 2005, Cvetkovic et al., 2012 and Kiel and Cardenas, 2014). Sustained release of waterborne nutrients and metals that have accumulated in the subsurface has indeed been found in the field and been modeled based on available observation data (Baresel and Destouni, 2005 and 2009, Darracq et al., 2008, Olli and Destouni, 2008 and Törnqvist et al., 2011). Such release from subsurface accumulation is relatively slow and steady and can as such explain biogeochemical stationarity reported in the literature for large-scale nutrient concentrations (Basu et al., 2010).

14

Figure 2. Water and ideal tracer travel times along different main pathways through a coastal divergent catchment. (a) The coastal divergent catchment of Forsmark (red outline, 30 km2 in total) in Sweden, with its surface water system, including streams (dark blue lines), lakes (light blue), wetlands (light blue with green stripes) and ten main convergent subcatchments (delineated by black lines) defined by associated stream outlet points at the coast. (b) Travel time distributions, shown with corresponding mean value T and coefficient of variation CV(T), as reported for water/tracer transport in the Forsmark catchment for different pathways: through groundwater to the nearest stream or coast (red: preferential shallow pathways through relatively highly-conductive homogeneous soil -transport scenario 1 in Persson et al. (2011); green: shallow pathways through the whole range of prevailing soil heterogeneity - transport scenario 2 in Persson et al. (2011); blue: deep pathways through both Quaternary Deposits and fractured bedrock - current temperate climate case A in Bosson et al. (2013); and turquoise: surface water pathways from various input points to and through the stream network to the coast - reported in Darracq et al. (2010)). Map is produced using ArcGIS 10.2. The travel time illustration in Figure 2 emphasizes the importance of characterizing and keeping track of all fractions of water and waterborne tracer, nutrients and pollutants that follow pathways with different travel times through each catchment. In addition to such characterization of pathway variability within each catchment, there is also a need to characterize and account for spatial pathway variability among catchments, with regard to the intensity and magnitude of flow, and constituent mass transport and accumulation. To further illustrate this, consider for example nutrient loading from Sweden to the Baltic Sea. Specifically, we synthesize here data over Sweden for: twentieth century hydroclimatic changes driven by changes of human land-use and water-use in the landscape (Destouni et al., 2013), and nutrient concentrations in freshwater discharges to the Baltic Sea (HELCOM, 2009). This data synthesis exemplifies how various human-driven changes and conditions in the landscape combine to create inter-catchment variability in the change propagation pathways of the freshwater system.

15

Figure 3. Inter-catchment variability in Swedish freshwater flux and nutrient load changes from 1901-1930 to 1971-2000. (a) Change in runoff (R) from Sweden to the Baltic Sea through the five Swedish water management districts (based on data from Destouni et al., 2013); the districts span combinations of convergent-divergent catchments draining into different marine basins. (b) Area-average change in precipitation (ΔP) and runoff (ΔRa) (based on data from Destouni et al., 2013) and concentration-weighted average change of R (ΔRC-w) over the entire Swedish catchment of the Baltic Sea; ΔRC-w is based on (c) long-term flow-average concentration of total nitrogen (TN) in the Swedish water districts (data for 1994-2006 from HELCOM, 2009). (d) Implications of R change (in a) and TN concentration (in c) for change in waterborne TN load (ΔLTN) from the Swedish water districts to the Baltic Sea. (e) Estimated change in total nutrient load (ΔL) from Sweden to the Baltic Sea based on ΔP, ΔRa or ΔRC-w. In (e), ΔL estimates are shown for TN and total phosphorus (TP; data for 1994-2006 from HELCOM, 2009) and are normalized by the total load reduction requirement for Sweden in the internationally agreed Baltic Sea Action Plan (BSAP; HELCOM, 2007). Appendix: Supplementary Material Section SM2 and Table SM1 provide further details on data and calculations. Maps are produced using ArcGIS 10.2. Twentieth-century changes in Swedish runoff to the sea (R, area-normalized freshwater discharge) exhibit a pattern of R decrease in the more densely populated southern catchments and R increase only in the less populated northern ones (Figure 3a). This runoff change pattern implies essentially unchanged average runoff from the whole of Sweden to the Baltic Sea (ΔRa, Figure 3b) and occurs in spite of an overall increase in precipitation over Sweden (ΔP, Figure 3b). The ΔR pattern combines with a similar inter-catchment variability pattern also in nutrient concentration levels (e.g., concentration CTN of total nitrogen (TN), Figure 3c) to yield a concentration-weighted change in large-scale runoff (ΔRC-w in Figure 3b based on CTN in Figure 3c) that differs considerably from the large-scale average precipitation change ΔP and the concentration-independent average runoff change ΔRa. The concentration-weighted change ΔRC-w quantifies the average runoff change in the freshwater pathways with the highest nutrient concentrations, which thus carry most of the nutrient mass load to the sea (product of water flow and concentration). In Sweden, these are

16

the pathways through the southern, most highly populated catchments, which thus propagate most of the change in Swedish nutrient loading to the sea (e.g., ΔLTN for TN, Figure 3d). Inter-catchment variability of nutrient concentrations (Figure 3c), which partly determines the nutrient load change pattern, is not just specific for Sweden, but also exhibited more generally by concentrations of both TN and total phosphorous (TP) in freshwater discharges to the sea from the whole Baltic region (Figure 4). Specifically, there is high inter-catchment variability in these concentrations (Figure 4) in spite of their temporal variability being small (biogeochemical stationarity; Basu et al., 2010). There is also high spatial correlation between the inter-catchment variability of concentration and that of population density and farmland area share (Figure 4), indicating the spatial pattern of concentration variability as largely human-driven.

Figure 4. Long-term, flow-averaged nutrient concentrations in water discharges to the Baltic Sea and their correlations with anthropogenic change drivers. Concentrations for total nitrogen (TN; blue) and total phosphorous (TP; red) are plotted versus corresponding population density (main plot) and relative farmland area (upper right insert) for different sub-catchments of the Baltic Sea (upper left insert). The sub-catchments for the population density plot are nested and extend over different countries (black country names in legend), catchment areas of different marine basins of the Baltic Sea (blue basin names in legend), or the whole Baltic Sea drainage basin (red in legend). Further information and data for the studied catchments and correlation discussion are given in Appendix: Supplementary Material Section SM2 and Table SM2. Map is produced using ArcGIS 10.2. Moreover, the Swedish inter-catchment variability in runoff change, which implies a contrast of precipitation increase and runoff decrease for the most highly populated southern Swedish catchments (Figure 3a), is also human-driven (Destouni et al., 2013). This contrast between precipitation and runoff change is consistent with a more general pattern of human-driven runoff change, found across different parts of the world, as illustrated further in the following subsection.

17

For comparison, climate model output of large-scale hydroclimatic change may be highly uncertain with regard to runoff changes, not least if these are mostly human-driven rather than driven by atmospheric climate change (Jarsjö et al., 2012 and Törnqvist et al., 2014). If estimates of large-scale nutrient load changes (ΔL) were to rely directly on climate model output, e.g., of large-scale average precipitation change or runoff change (ΔP or ΔRa based estimates of ΔL, respectively, Figure 3e), they could differ considerably in both magnitude and change direction from ΔL estimates that do account for actual inter-catchment variability of runoff change (ΔRC-w based estimate of ΔL, Figure 3e). Such differences may be critical in a regulatory context. In the Swedish example, discrepancies among various ΔL estimates are as large as the total load reduction requirement for Sweden in the internationally agreed Baltic Sea Action Plan for combating marine eutrophication and hypoxia (HELCOM, 2007) (ΔL is normalized with corresponding reduction requirement in Figure 3e).

18

8 Interactions with the anthroposphere (aspect iv) The freshwater conceptualization in Figure 1 represents interactions with the anthroposphere as integral pathways in the freshwater system, emphasizing that these interactions may be essential for freshwater changes, in addition to such changes driven by atmospheric climate change. This possibility has already been indicated above by observed changes in Swedish precipitation and runoff (Figures 3a-b), in combination with human-related patterns in nutrient concentrations (Figures 3c and 4). In this section, we will further investigate this possibility by comparative analysis of catchment-wise connected hydroclimatic data and observed change behavior in catchments with different known anthropogenic disturbance levels. Consider then such changes across different world regions (Figure 5a). Specifically, the comparison is between relatively well-studied regional catchments or sub-catchments, of which some have anthropogenic disturbance levels that are known to have increased considerably over the 20th century and others have no such reported twentieth century increases. These catchments and sub-catchments are thus considered here because of their previously studied, and therefore to considerable degree independently known, twentieth century developments in hydroclimatic variables and human disturbance levels. The known human disturbances include agricultural intensification, expansion or irrigation, and/or flow regulation and fragmentation related to hydropower developments. For each of the investigated catchments in Figure 5a, the catchment-wise balancing of their average annual water fluxes implies that available observation data for precipitation (P) and runoff (R) can be used to physically constrain estimates of evapotranspiration (ET) as ET≈P-R (Destouni et al., 2013, van der Velde et al., 2013, Jaramillo and Destouni, 2014 and Levi et al., 2015). By the present synthesis of such long-term data, historic change patterns emerge that do indeed differ distinctly between the known disturbed and the other catchments. On average in the disturbed catchments (red, Figures 5b-d), evapotranspiration ET has increased more than precipitation P, leading to decreased average runoff R in spite of the increased average P (Figure 5b). Even individual disturbed catchments with small or no P change exhibit an ET increase of around 50 mm yr-1 (Figure 5c) and, as a consequence, also a corresponding R decrease (Figure 5d). In contrast, ET and R changes in the other catchments (green, Figures 5b-d) are around zero for zero P change, on average (Figure 5b) as well as in individual catchments (Figures 5c-d). Overall, the found differences between known anthropogenically disturbed and other catchments shows how changes in human land-use and water-use propagate through the freshwater system to yield dominant human-driven changes in ET and R on large scales, in addition to such changes driven by atmospheric climate change. The representation of direct freshwater interactions with the anthroposphere in the present conceptualization of the freshwater system is needed to facilitate consideration and characterization of such interactions as possible dominant drivers of freshwater change. The large-scale ET changes (Figures 5b-c) imply corresponding changes in latent heat flux (Destouni et al., 2010b and Asokan et al., 2014), thus illustrating that direct freshwater interactions with the anthroposphere may also drive distinct changes in energy transport to the atmosphere with associated human-driven effects on regional climate change.

19

Figure 5. Historic changes of catchment-wide freshwater fluxes between different climatic periods of the twentieth century. (a) Regions with studied hydrological catchments, as summarized in the panel, with further catchment information and data given in Appendix: Supplementary Material (SM) Section SM2 and Tables SM.3-7. (b) Average value ±1 standard deviation (error bars) of changes in catchment-wise precipitation (P), runoff (R) and evapotranspiration (ET) across the studied catchments; changes are between different time periods of the twentieth century as listed in Appendix: Table SM3. Changes in individual catchments (red: known disturbed, green: other; see further disturbance information in Appendix: Table SM3) are shown in (c) for ET and in (d) for R, versus corresponding changes in P. The Arctic catchment of Olenek appears as an outlier in the set of other catchments (green) in panels (c)-(d) and has been noted as such also in previous studies of Arctic hydroclimatic change (Bring and Destouni, 2011); for comparison, without Olenek, regression lines for the set of other catchments are: y = 0.7225x - 5.9942 (R² = 0.83565) for ET versus P in panel (c) and y = 0.2775x + 5.9942 (R² = 0.42855) for R versus P in panel (d). Map is produced using Wolfram Mathematica.

20

9. Discussion Traditional water cycle conceptualizations and interpretations exhibit freshwater gaps and inconsistencies. These convey a lack of common multi- and trans-disciplinary understanding of the freshwater system and its propagation of water change. To secure sustainable water resources, such gaps and inconsistencies must be recognized and bridged across scientific disciplines, policies and management practices. This need also applies to the monitoring and observation data collected to follow, interpret and project freshwater change. Identification and control of long-term change trends and evolution in the freshwater system are still elusive on system level and across local-regional-global scales, even though physical interaction principles may be relatively well understood for local processes in some freshwater components. Such local understanding may also include local component interactions, e.g. of groundwater and surface water, or freshwater and seawater, however typically without systematic linkages of these interactions to other parts of the freshwater system and its overall changes. Yet identification of system-level changes and feedback loops throughout the whole freshwater system is required for effective monitoring, long-term planning and sustainable solutions to major water resource challenges worldwide. Growing human-driven change pressures in the Anthropocene imply increased risk for un-wanted water resource impacts. They also imply more complex links between direct (e.g., water use, pollution, eutrophication) and indirect (e.g., climate change) human-driven changes in the freshwater system. To secure long-term sustainability, effective collaboration is needed across science disciplines, engineering, planning, management and policy for improved understanding, distinction and control of such human-driven changes and their freshwater propagation. The present conceptualization is designed to help in such collaboration by advancing a common inter- and trans-disciplinary understanding of this life-critical system. Such advancement is facilitated by focused consideration of the freshwater system as a whole, coupled by multiple pathways of change propagation. Freshwater change must thus be followed and interpreted along and among all such main water pathways. Based on these pathways, the present conceptualization provides a basis for systematic characterization and typology of freshwater system function, its changes, and various approaches to model and monitor these within and across catchments, and on local, regional and global scales.

21

10 References Arctic-HYDRA consortium. The Arctic hydrological cycle monitoring, modelling and

assessment programme: Science and implementation plan. ISBN 978-9979-9975-0-4 (2010).

Asokan, S.M. & Destouni, G. Irrigation effects on hydro-climatic change: Basin-wise water balance-constrained quantification and cross-regional comparison. Surveys in Geophysics 35, 879–895 (2014).

Azcárate, J., Balfors, B., Bring, A. & Destouni, G. Strategic environmental assessment and monitoring: Arctic key gaps and bridging pathways. Environ. Res. Lett. 8, 044033 (2013).

Banks, D. et al. Mine-water chemistry: the good, the bad and the ugly. Environmental Geology 32, 157-174 (1997).

Baresel, C. & Destouni, G. Novel quantification of coupled natural and cross-sectoral water and nutrient/pollutant flows for environmental management, Environ. Sci. Technol. 39, 6182 – 6190 (2005).

Baresel, C. & Destouni, G. Diffuse subsurface zinc loads from mining areas in the Dalälven River Basin, Sweden. Hydrology Research 40, 445-453 (2009).

Baresel, C., Destouni, G. & Gren I.M. The influence of metal source uncertainty on cost-effective allocation of mine water pollution abatement in catchments. Journal of Environmental Management 78, 138-148 (2006).

Battin, T. J., Luyssaert, S., Kaplan, L. A., Aufdenkampe, A. K., Richter, A. & Tranvik, L. J. The boundless carbon cycle. Nature Geosci. 2, 598-600 (2009).

Berghuis, W. R., Woods R.A. & Hrachowitz, M. A. Precipitation shift from snow towards rain leads to a decrease in streamflow. Nature Clim. Change 4, 583–586 (2014).

Basu, N. B. et al. Nutrient loads exported from managed catchments reveal emergent biogeochemical stationarity. Geophys. Res. Lett. 37, L23404 (2010).

Beven, K. Searching for the Holy Grail of scientific hydrology: Qt = H(S,R, 1t)A as closure. Hydrology and Earth System Sciences 10, 609–618 (2006).

Boman, A., Fröjdo, S., Backlund, K. & Åström, M.E. Impact of isostatic land uplift and artificial drainage on oxidation of brackish-water sediments rich in metastable iron sulfide. Geochim. Cosmochim. Acta 74, 1268−1281 (2010).

Bosson, E., Sabel, U., Gustafsson, L.G., Sassner, M. & Destouni, G. Influences of shifts in climate, landscape, and permafrost on terrestrial hydrology. Journal of Geophysical Research-Atmospheres 117, D05120 (2012).

Bosson, E., Selroos, J.-O., Stigsson, M., Gustafsson, L.-G. & Destouni, G. Exchange and pathways of deep and shallow groundwater in different climate and permafrost conditions using the Forsmark site, Sweden as an example catchment, Hydrogeology Journal 21, 225–237 (2013).

Botter, G., Basso, S., Rodriguez-Iturbe, I. & Rinaldo, A. 2013. Resilience of river flow regimes. Proc. Natl Acad. Sci. USA 110, 12925-12930 (2013).

Bring, A. & Destouni, G. Hydrological and hydrochemical observation status in the Pan-Arctic drainage basin. Polar Research 28, 327–338 (2009).

Bring, A. & Destouni, G. Relevance of hydro-climatic change projection and monitoring for assessment of water cycle changes in the Arctic. Ambio 40, 361-369 (2011).

Bring, A. & Destouni, G. Hydro-climatic changes and their monitoring in the Arctic: Observation-model comparisons and prioritization options for monitoring development. J. Hydrol. 492, 273-280 (2013).

Bring, A. & Destouni, G. Arctic climate and water change: Model and observation relevance for assessment and adaptation, Surveys in Geophysics 35, 853–877 (2014).

22

Brown, K., Tompkins, E.L. & Adger W.N. Making Waves: Integrating Coastal Conservation and Development, London: Earthscan (2002).

Burton, E. D. et al. Iron-Monosulfide Oxidation in Natural Sediments: Resolving Microbially Mediated S Transformations Using XANES, Electron Microscopy, and Selective Extractions. Environ. Sci. Technol. 43(9): 3128−3134 (2009).

Cole, J.J. et al. Plumbing the Global Carbon Cycle: Integrating Inland Waters into the Terrestrial Carbon Budget. Ecosystems 10, 171–184 (2007).

Crutzen, P. J. & Stoermer, E. F. The ”Anthropocene”. Global Change Newsletter 41, 17–18 (2000).

Crutzen, P. J. Geology of mankind. Nature 415, 23 (2002). Cvetkovic, V., Carstens, C., Selroos, J.-O. & Destouni, G. Water and solute transport along

hydrological pathways. Water Resour. Res. 48, W06537 (2012). Darracq, A., Lindgren, G. & Destouni G. Long-term development of Phosphorus and

Nitrogen loads through the subsurface and surface water systems of drainage basins. Global Biogeochemical Cycles 22, GB3022 (2008).

Darracq, A., Destouni, G., Persson, K., Prieto, C. & Jarsjö, J. Quantification of advective solute travel times and mass transport through hydrological catchments. Environ. Fluid Mech. 10, 103–120 (2010).

Destouni, G. & Prieto, C. On the possibility for generic modeling of submarine groundwater discharge. Biogeochemistry 66, 171–186 (2003).

Destouni, G. & Verrot, L. Screening long-term variability and change of soil moisture in a changing climate. Journal of Hydrology 516, 131-139 (2014).

Destouni, G., Hannerz, F., Prieto, C., Jarsjö, J. & Shibuo, Y. Small unmonitored near-coastal catchment areas yielding large mass loading to the sea. Global Biogeochem. Cy. 22, GB4003 (2008).

Destouni, G., Persson, K., Prieto, C. & Jarsjö, J. General quantification of catchment-scale nutrient and pollutant transport through the subsurface to surface and coastal waters. Environ. Sci. Technol. 44, 2048–2055 (2010a).

Destouni, G., Asokan, S.M. & Jarsjö, J. Inland hydro-climatic interaction: Effects of human water use on regional climate. Geophysical Res. Lett. 37, L18402 (2010b).

Destouni, G., Jaramillo, F. & Prieto, C. Hydroclimatic shifts driven by human water use for food and energy production. Nature Clim. Change 3, 213–217 (2013).

Destouni, G., Asokan, S.M., Augustsson, A., Balfors, B., Bring, A., Jaramillo, F., Jarsjö, J., Johansson, E., Juston, J., Levi, L., Olofsson, B., Prieto, C., Quin, A., Åström M.E. Cvetkovic, V. Biogeochemical Transformation Pathways through the Land-water Geosphere. Abstract ID: 4955, American Geophysical Union (AGU) Fall Meeting, San Francisco, 15-19 December (2014).

Dickinson, R. E. Modeling evapotranspiration for three-dimensional global climate models. Geophysical Monographs (J. E. Hansen and T. Takahashi, eds.) 29, 58-72. American Geophysical Union (1984).

Ferguson, G. & Gleeson, T. Vulnerability of coastal aquifers to groundwater use and climate change. Nature Clim. Change 2, 342–345 (2012).

Fick, J. et al. Contamination of surface, ground, and drinking water from pharmaceutical production. Environmental Toxicology and Chemistry 28, 2522-2527 (2009).

Government Offices of Sweden. Government and Ministries (2014). Date of access 20/12/2014, http://www.government.se/sb/d/576.

Hannerz, F. & Destouni, G. Spatial characterization of the Baltic Sea drainage basin and its unmonitored catchments. Ambio 35, 214-219 (2006).

23

HELCOM. An Approach to Set Country-wise Nutrient Reduction Allocations to Reach Good Marine Environment of the Baltic Sea. Helcom BSAP Eutro Expo. Helsinki, Finland: Helsinki Commission (2007).

HELCOM. Eutrophication in the Baltic Sea: An integrated assessment of the effects of nutrient enrichment in the Baltic Sea Region. BSEP Rep. 115B, Helsinki Comm., Helsinki (2009).

Horton, R. E. The field, scope, and status of the science of hydrology. American Geophysical Union Transactions, Reports and Papers, Hydrology, 189-202 (1931).

Jacobson, M. C., Charlson, R. J., Rodhe, H. & Orians, G. H. (editors). Earth System Science - From Biogeochemical Cycles to Global Change. International Geophysics 72, 1-527, ISBN: 978-0-12-379370-6 (2000).

Jantze, E.J., Lyon, S.W. & Destouni, G. Subsurface release and transport of dissolved carbon in a discontinuous permafrost region. Hydrology and Earth System Sciences 17, 3827–3839 (2013).

Jaramillo, F. & Destouni, G. Developing water change spectra and distinguishing change drivers worldwide. Geophysical Research Letters, doi: 10.1002/2014GL061848 (2014, in press).

Jarsjö, J., Asokan, S.M., Prieto, C., Bring, A. & Destouni, G. Hydrological responses to climate change conditioned by historic alterations of land-use and water-use. Hydrol. Earth Syst. Sci. 16, 1335–1347 (2012).

Karlsson, J. M., Bring, A., Peterson, G. D., Gordon, L. J. & Destouni, G. Opportunities and limitations to detect climate-related regime shifts in inland Arctic ecosystems through eco-hydrological monitoring. Environ. Res. Lett. 6, 014015 (2011).

Karlsson, J.M., Lyon, S.W. & Destouni, G. Thermokarst lake, hydrological flow and water balance indicators of permafrost change in Western Siberia. Journal of Hydrology 464–465, 459–466 (2012).

Kiel, B. A. & Cardenas, M. B. Lateral hyporheic exchange throughout the Mississippi River network. Nature Geosci. 7, 413−417 (2014).

Ledger, M. E., Brown, L. E., Edwards, F. K., Milner, A. M. & Woodward, G. Drought alters the structure and functioning of complex food webs. Nature Clim. Change 3, 223–227 (2013).

Levi L., Jaramillo F., Andričević R. & Destouni, G. Hydroclimatic changes and drivers in the Sava River Catchment and comparison with Swedish catchments. Ambio (2015, in press) DOI 10.1007/s13280-015- 0641-0.

Lyon, S. W., Mörth, M., Humborg, C., Giesler, R. & Destouni, G. The relationship between subsurface hydrology and dissolved carbon fluxes for a sub-arctic catchment, Hydrol. Earth Syst. Sci. 14, 941-950 (2010).

Mazi, K., Koussis, A. D. & Destouni, G. Tipping points for seawater intrusion in coastal aquifers under rising sea level. Environ. Res. Lett. 8, 014001 (2013).

McDonnell, J.J. et al. Moving beyond heterogeneity and process complexity: A new vision for watershed hydrology. Water Resources Research 43, W07301 (2007).

McDonnell, J. J. et al. How old is streamwater? Open questions in catchment transit time conceptualization, modelling and analysis. Hydrol. Process. 24(12), 1745−1754 (2010).

Milly, P. C. D., Dunne, K. A. & Vecchia, A. V. Global pattern of trends in streamflow and water availability in a changing climate. Nature 438, 347–350 (2005).

Montanari, A. et al. “Panta Rhei—Everything Flows”: Change in hydrology and society—The IAHS Scientific Decade 2013–2022. Hydrolog. Sci. J. 58, 1256−1275 (2013).

Nilsson, L. M. et al. A call for urgent monitoring of food and water security based on relevant indicators for the Arctic. Ambio 42, 816–822 (2013).

24

Sivapalan, M., Savenije, H. H. G. & Blöschl, G. Socio-hydrology: A new science of people and water. Hydrol. Process. 26, 1270–1276 (2012).

Olli, G. & Destouni, G. Long-term heavy metal loading to near-shore sediments. Water, Air, and Soil Pollution 192, 105-116 (2008).

Pahl-Wostl, C. Transitions towards adaptive management of water facing climate and global change. Water Resour. Manage. 21, 49–62 (2007).

Persson, K. & Destouni, G. Propagation of water pollution uncertainty and risk from the subsurface to the surface water system of a catchment. Journal of Hydrology 377, 434-444 (2009).

Persson, K., Jarsjö, J. & Destouni, G. Diffuse hydrological mass transport through catchments: scenario analysis of coupled physical and biogeochemical uncertainty effects. Hydrol. Earth Syst. Sci. 15, 3195–3206 (2011).

Prieto, C. & Destouni, G. Quantifying hydrological and tidal influences on groundwater discharges to coastal waters Water Resources Research 41, W12427 (2005).

Prieto, C. & Destouni G. Is submarine groundwater discharge predictable? Geophysical Research Letters 38, L01402 (2011).

Seitz, N.E., Westbrook, C.J., Dubé, M.G. & Squires A.J. Assessing large spatial scale landscape change effects on water quality and quantity response in the lower Athabasca River Basin. Integr. Environ. Assess. Manag. 9, 392-404 (2013).

Shiklomanov, I. A. & Sokolov, A. A. Methodological basis of world water balance investigation and computation. In New Approaches in Water Balance Computations. International Association for Hydrological Sciences Publication No. 148, Proceedings of the Hamburg Symposium (1983).

Sivapalan, M. Pattern, process and function: elements of a unified theory of hydrology at the catchment scale. In Encyclopedia of Hydrological Sciences, Anderson MG (ed.), Wiley: Chichester 193–219 (2005).

Steefel, C. I., DePaolo, D. J. & Lichtner, P. C. Reactive transport modeling: An essential tool and a new research approach for the Earth sciences. Earth and Planetary Science Letters 240, 539-558 (2005).

Sundström, R., Åström, M. & Österholm, P. Comparison of the metal content in acid sulfate soil runoff and industrial effluents in Finland. Environ. Sci. Technol. 36, 4269−4272 (2002).

Tetzlaff, D. et al. Conceptualizing catchment processes: simply too complex? Hydrol. Process. 22, 1727–1730 (2008).

Törnqvist, R., Jarsjö, J. & Karimov B. Health risks from large-scale water pollution: Trends in Central Asia. Environment International 37, 435-442 (2011).

Törnqvist, R. et al. Evolution of the hydro-climate system in the Lake Baikal basin. J. Hydrol. 519, 1953–1962 (2014).

UNECE. Guidance on water and adaptation to climate change. United Nations Economic Commission for Europe, Geneva, p 144 (2009).

Van der Velde, Y., Lyon, S. & Destouni, G. Data-driven regionalization of river discharges and emergent land cover–evapotranspiration relationships across Sweden. Journal of Geophysical Research: Atmospheres 118, 2576–2587 (2013).

Vörösmarty, C. J. et al. Global threats to human water security and river bio- diversity. Nature 467, 555–561 (2010).

Åkesson, M., Sparrenbom, C.J., Dahlqvist, P. & Fraser, S.J. On the scope and management of pesticide pollution of Swedish groundwater resources: The Scanian example. Ambio 44, 226-238 (2015).

25

Acknowledgements Nova R&D (project KLIV), along with the Swedish Research Council (VR, project 2009-3221), The Swedish Research Council Formas (project 2014-43) and Stockholm University’s Strategic Environmental Research Program Ekoklim have supported the research synthesized and summarized in this report. Author contributions

GD has written the report, with particular help from VC, and has led and carried out much of the research based on ideas developed and discussed jointly within the KLIV research group: AA, BB, VC, GD, JJar, BO, AQ and MÅ. Co-authors SA (Aral, Mahanadi), AB (Arctic), FJ (Sweden, water fluxes), EJ (Forsmark, travel times), JJus (Sweden, nutrients), LL (Sava) and CP (Sweden, water fluxes) have contributed data and calculations for different catchment studies. All co-authors have contributed to result interpretations.

26

Appendix - Supplementary Material - Section SM1. Analysis of traditional water cycle conceptualizations.

- Section SM2. Notes for result Figures 3-5.

- Table SM1. Information on changes in land-water fluxes from 1901-1930 to 1971-2000 and their implications for change in waterborne nutrient loading from Sweden to the Baltic Sea; relates to Figure 3.

- Table SM2. Information and data for the catchments in Figure 4, regarding their area, population, farmland area, and long-term flow-averaged concentrations of total nitrogen and total phosphorus in freshwater discharges to the Baltic Sea.

- Table SM3. Information and data for the catchments and associated time periods of change in Figure 5.

27

Section SM1. Analysis of traditional water cycle conceptualizations

Different types of traditional water cycle conceptualizations used in advanced Earth system science literature (Jacobson et al., 2000) were analyzed here: (a) the global water balance conceptualization (and quantification) of Shiklomanov and Sokolov (1983; Figure 6.3 in Jacobson et al., 2000); (b) the process conceptualization of the global hydrological cycle of Horton (1931; Figure 6.4 in Jacobson et al., 2000); (c) the surface process conceptualization with focus on evaporation and transpiration from vegetation of Dickinson (1984; Figure 6.5 in Jacobson et al., 2000); (d) the conceptualization of main cyclic processes and fluxes between major Earth reservoirs of Jacobson et al. (2000; their Figure 1.1); (e) the conceptualization of processes connecting global biogeochemical cycles and climate of Jacobson et al. (2000; their Figure 19-1); (f) the conceptualization (and quantification) of flux changes relative to pre-industrial time for radiatively important trace species in the atmosphere of Jacobson et al. (2000; their Table 19-1). The analysis identified that neither of these considers or visualizes: (1) freshwater interactions with the anthroposphere, (2) the catchment-wise organization of freshwater flows, or (3) freshwater-seawater interactions in coastal regions. Furthermore, with regard to interactions (4) between surface and subsurface freshwater, and (5) between subsurface freshwater and the lithosphere, these are inconsistently considered among the conceptualizations. Specifically, conceptualization (a) of Shiklomanov and Sokolov (1983) shows interaction pathways between subsurface freshwater and surface freshwater, emphasizing a key role of subsurface (soil-groundwater) flow pathways to surface waters (with global average groundwater flow of 43 800 km3/year feeding a total global average river flow of 44 700 km3/year). However, without also representing and visualizing the catchment-wise organization of these pathways, or the subsurface water interactions with the highly heterogeneous lithosphere, this freshwater conceptualization does not convey that most of the groundwater that actually interacts with surface waters may have a very wide range of subsurface turnover (travel, residence) times, from very short to very long, until it reaches a nearest surface or coastal water (see Figure 4), rather than just the single, very long turnover time of 500 years that is used to characterize the entire groundwater reservoir in conceptualization (a). Furthermore, conceptualization (b) of Horton (1931) details processes in different freshwater subsystems without showing any interactions between these or of the subsurface water with the lithosphere. Conceptualization (c) of Dickinson (1984) cuts off and considers only some parts of the freshwater system, at the land surface and root zone, thus missing, for example, key interactions between groundwater, soil moisture and evapotranspiration (Destouni and Verrot, 2014), even though evapotranspiration is of main focus in this particular freshwater conceptualization. Conceptualization (d) of Jacobson et al. (2000) shows some freshwater pathways and freshwater interactions with the lithosphere; however, in direct contrast to

28

conceptualization (c), its focus is on the deeper subsurface, with the groundwater pathway termed here “aquifer recharge” while otherwise not resolving the freshwater system at all beyond the precipitation and evapotranspiration pathways. Such contrasting freshwater representations lead further to contradictory interpretations of the structure and function of the freshwater system across different disciplinary and environmental perspectives. Considering, for example, conceptualization (a) that is included in one context of Jacobson et al. (2000), this shows quantitatively groundwater flow as the major water-feeding pathway into surface water, contributing on average 98% of the total stream and river runoff from land to sea. However, in other contexts of the same Earth System science literature, only lakes and rivers are stated as main transport pathways for dissolved ions, nutrients and organic matter (subsection 6.2.1.4 of Jacobson et al., 2000), and groundwater is stated to have a limited role in conveying weathering products and runoff from the land surface to the oceans (subsection 6.5 of Jacobson et al., 2000). In both of the latter statements, the major contribution of the groundwater pathway in feeding water and its constituents to surface waters is thus neglected. In contrast, in conceptualization (d) that is used in yet another context of the same literature (section 1 of Jacobson et al., 2000), the groundwater pathway is included as necessary link between the freshwater hydrosphere and the lithosphere in order for the latter to feed ions into the former. Furthermore, conceptualization (e), used in another context of Jacobson et al. (2000), states only the terrestrial biosphere as processing carbon on land in addition to industrial carbon output (Figure 19-1 of Jacobson et al., 2000), thus entirely neglecting the inland water system processing of large amounts of organic carbon (Cole et al., 2007 and Battin et al., 2009). Moreover, also studies of carbon cycling that have focused on freshwater have mostly only considered surface water bodies (Cole et al., 2007 and Battin et al., 2009), thus neglecting the subsurface freshwater (soil-groundwater) pathways that largely feed the major amount of water (Shiklomanov and Sokolov, 1983) as well as the dissolved organic and inorganic carbon (Lyon et al., 2010 and Hantze et al., 2013) and other waterborne constituents (Baresel and Destouni, 2005, Darracq et al., 2008, Destouni et al., 2008, 2010a, Basu et al., 2010, Persson and Destouni, 2009, Persson et al., 2011 and Cvetkovic et al., 2012) that this subsurface water carries with it into lakes and rivers. In conceptualization (f), water flux changes from pre-industrial time are further stated as essentially unknown in comparison with flux changes in other radiatively important trace species in the atmosphere (Table 19-1 of Jacobson et al., 2000). In contrast, elsewhere in the same context the same literature states that water changes driven by direct interactions with the anthroposphere are negligible relative to climate-driven changes (subsection 6.8 of Jacobson et al., 2000). The former statement of major uncertainty about water flux changes thus implies some degree of speculation in the latter statement of human-driven water-flux changes being negligible; furthermore, findings in the present study (Figure 9) as well as in other recent studies (Destouni et al., 2010b, 2013, Jarsjö et al., 2012, Asokan et al., 2014 and Jaramillo and Destouni, 2014) contradict the latter statement. In summary, several gaps and inconsistencies are evident across freshwater representations and interpretations in traditional water cycle conceptualizations. The examples from advanced Earth system science literature that were analyzed here combine in showing scientific fragmentation and limited or even contradictory consideration of the structure and function of the freshwater system, and a lack of common multi-disciplinary understanding and conceptualization of this system and its propagation of human-driven change.

29

SM1 References Asokan, S. M. & Destouni, G. Irrigation effects on hydro-climatic change: Basin-wise water

balance-constrained quantification and cross-regional comparison. Surveys in Geophysics, 879–895 (2014).

Baresel, C. & Destouni, G. Novel quantification of coupled natural and cross-sectoral water and nutrient/pollutant flows for environmental management, Environ. Sci. Technol., 6182 – 6190 (2005).

Basu, N. B. et al. Nutrient loads exported from managed catchments reveal emergent biogeochemical stationarity. Geophys. Res. Lett., L23404 (2010).

Battin, T. J., Luyssaert, S., Kaplan, L. A., Aufdenkampe, A. K., Richter, A. & Tranvik, L. J. The boundless carbon cycle. Nature Geosci., 598–600 (2009).

Cole, J.J. et al. Plumbing the Global Carbon Cycle: Integrating Inland Waters into the Terrestrial Carbon Budget. Ecosystems, 171–184 (2007).

Cvetkovic, V., Carstens, C., Selroos, J.-O. & Destouni, G. Water and solute transport along hydrological pathways. Water Resour. Res., W06537 (2012).

Darracq, A., Lindgren, G. & Destouni, G. Long-term development of Phosphorus and Nitrogen loads through the subsurface and surface water systems of drainage basins. Global Biogeochemical Cycles, GB3022 (2008).

Destouni, G. & Verrot, L. Screening long-term variability and change of soil moisture in a changing climate. Journal of Hydrology, 131–139 (2014).

Destouni, G., Hannerz, F., Prieto, C., Jarsjö, J. & Shibuo, Y. Small unmonitored near-coastal catchment areas yielding large mass loading to the sea. Global Biogeochem. Cy., GB4003 (2008).

Destouni, G., Persson, K., Prieto, C. & Jarsjö, J. General quantification of catchment-scale nutrient and pollutant transport through the subsurface to surface and coastal waters. Environ. Sci. Technol., 2048–2055 (2010a).

Destouni, G., Asokan, S. M. & Jarsjö, J. Inland hydro-climatic interaction: Effects of human water use on regional climate. Geophysical Res. Lett., L18402 (2010b).

Destouni, G., Jaramillo, F. & Prieto, C. Hydroclimatic shifts driven by human water use for food and energy production. Nature Clim. Change, 213–217 (2013).

Dickinson, R. E. Modeling evapotranspiration for three-dimensional global climate models. Geophysical Monographs (J. E. Hansen and T. Takahashi, eds.) 29, 58-72. American Geophysical Union (1984).

Horton, R. E. The field, scope, and status of the science of hydrology. American Geophysical Union Transactions, Reports and Papers, Hydrology, 189-202 (1931).

Jacobson, M. C., Charlson, R. J., Rodhe, H. & Orians, G. H. (eds). Earth System Science - From Biogeochemical Cycles to Global Change. International Geophysics ISBN: 978-0-12-379370-6 (2000).

Jantze, E.J., Lyon, S.W. & Destouni, G. Subsurface release and transport of dissolved carbon in a discontinuous permafrost region. Hydrology and Earth System Sciences, 3827–3839 (2013).

30

Jaramillo, F. & Destouni, G. Developing water change spectra and distinguishing change drivers worldwide. Geophysical Research Letters, doi: 10.1002/2014GL061848 (2014, in press).

Jarsjö, J., Asokan, S.M., Prieto, C., Bring, A. & Destouni, G. Hydrological responses to climate change conditioned by historic alterations of land-use and water-use. Hydrol. Earth Syst. Sci., 1335–1347 (2012).

Lyon, S. W., Mörth, M., Humborg, C., Giesler, R. & Destouni, G. The relationship between subsurface hydrology and dissolved carbon fluxes for a sub-arctic catchment, Hydrol. Earth Syst. Sci., 941–950 (2010).

Persson, K. & Destouni, G. Propagation of water pollution uncertainty and risk from the subsurface to the surface water system of a catchment. Journal of Hydrology 377, 434-444 (2009).

Persson, K., Jarsjö, J. & Destouni, G. Diffuse hydrological mass transport through catchments: scenario analysis of coupled physical and biogeochemical uncertainty effects. Hydrol. Earth Syst. Sci., 3195–3206 (2011).

Shiklomanov, I. A. & Sokolov, A. A. Methodological basis of world water balance investigation and computation. In New Approaches in Water Balance Computations. International Association for Hydrological Sciences Publication No. 148, Proceedings of the Hamburg Symposium (1983).

31

Section SM2. Notes for result Figures 3-5

For result Figures 3-4, long-term flow-average concentration of TN and TP was obtained as the district- or catchment-wise ratio of reported (HELCOM, 2009) annual nutrient load and water discharge averaged over the period of available data 1994-2006. For Figure 3, Table SM1 outlines further details. For Figure 4, Table SM2 provides further information and data for the studied Baltic Sea catchments. With regard to Figure 4, the high basin-wise spatial correlation found here across the whole Baltic region and for both TN and TP between the long-term flow-average nutrient concentration and population or farmland area share are in principle logical, even though quantitative specifics may vary for catchments in different regions. Both population density and farmland area relate to human activities (with wastewater discharges from households and industry, and diffuse nutrient leakage from agriculture) yielding excess nutrient inputs to the environment. Furthermore, dominant environmental legacies from historic nutrient inputs of such activities can explain why temporal variability of nutrient concentrations is so small that water discharge (rather than concentration) dynamics controls the inter-annual variability of nutrient loads to the Baltic Sea and other marine systems (Basu et al., 2010), even though the spatial variability of flow-average nutrient concentration is high (Figure 4). For Figure 5, information and data for all studied hydrological catchments are given in Table SM3. Changes in catchment-wise average precipitation (P), runoff (R) and evapotranspiration (ET) are between different time periods of the twentieth century as listed in Table SM3. Sources for reported anthropogenic disturbances in known disturbed catchments are also given in Table SM3.

SM2 References

Basu, N. B. et al. Nutrient loads exported from managed catchments reveal emergent biogeochemical stationarity. Geophys. Res. Lett., L23404 (2010).

HELCOM. Eutrophication in the Baltic Sea: An integrated assessment of the effects of nutrient enrichment in the Baltic Sea Region. BSEP Rep. 115B, Helsinki Comm., Helsinki (2009).

32

Table  SM1  

Information on changes in land-water fluxes from 1901-1930 to 1971-2000 and their implications for change in waterborne nutrient loading from Sweden to the Baltic Sea; relates to Figure 3. The load changes (∆LTN for total nitrogen (TN) and ∆LTP for total phosphorus (TP)) are given in million metric tons per year (MT/year) and their calculation is based on observed change in long-term average precipitation (∆P) or runoff (area-average ∆Ra or concentration-weighted area-average ∆RC-w) from 1901-1930 to 1971-2000 in representative Swedish catchments for each of five Swedish water management districts (combinations of multiple convergent and coastal divergent catchments; see colored fields in maps of Figure 3; P and R data for representative catchments are from Destouni et al. 2013). Each of the five Swedish water management districts represents a Swedish catchment part of a Baltic Sea basin (with two districts located within each of the Gulf of Bothnia and Baltic Proper catchments, and one district within the Danish Straits-Kattegatt catchment; see further Baltic Sea catchments listed in Table SM2). Long-term flow-average nutrient concentration in the water discharge from each Swedish water district to the Baltic Sea is obtained as in Table SM2, using reported data with intra-national resolution from HELCOM (2009). Resulting total load change for the whole of Sweden is in Figure 3e normalized with the internationally agreed load reduction requirement for Sweden (BSAP 2007 in the table, given in MT/year) in the Baltic Sea Action Plan (BSAP) (HELCOM, 2007).

Estimated total Swedish load change

(MT/year)

Total nitrogen with ∆LTN based on: Total phosphorus with ∆LTP based on:

∆P ∆Ra ∆RC-w BSAP 2007 ∆P ∆Ra ∆RC-w BSAP

2007 Accounting for area-average change in precipitation or runoff

19 243 -288

-20 780

601 -9

-290 Weighted average, accounting for combined spatial variability in concentration and runoff change among water districts

-5 128 -196

References:

Destouni, G., Jaramillo, F. & Prieto, C. Hydroclimatic shifts driven by human water use for food and energy production. Nature Clim. Change 3, 213–217 (2013).

HELCOM. Eutrophication in the Baltic Sea: An integrated assessment of the effects of nutrient enrichment in the Baltic Sea Region. BSEP Rep. 115B, Helsinki Comm., Helsinki (2009).

HELCOM. An Approach to Set Country-wise Nutrient Reduction Allocations to Reach Good Marine Environment of the Baltic Sea. Helcom BSAP Eutro Expo. Helsinki, Finland: Helsinki Commission (2007).

33

Table  SM2  

Information and data for the catchments in Figure 4, regarding their area, population, farmland area, and long-term flow-averaged concentrations of total nitrogen (TN) and total phosphorus (TP) in freshwater discharges to the Baltic Sea. *Long-term flow-average concentration is obtained as the catchment-wise ratio of reported annual nutrient load and water discharge averaged over the period of available data 1994-2006.

Catchment

Area within the BSDB (thousand

km2)1

Population within the

BSDB (million)1

Population density (per

km2)

Farmland (incl.

grasslands) % of total

area2

Mean water flow (mm/day)3

TN concen-tration* (mg/l)3

TP concen-tration* (mg/l)3

Denmark 26 4.5 171 66 1.0 6.3 0.21 Estonia 44 1.4 33 30 1.2 1.5 0.06 Finland 301 5.1 17 7 0.6 1.0 0.05 Germany 26 2.8 110 72 0.4 5.5 0.15 Latvia 64 2.4 37 39 1.3 2.6 0.07 Lithuania 65 3.6 56 54 0.9 2.3 0.11 Poland 310 38.6 124 60 0.6 3.3 0.21 Russia 329 9.7 30 12 0.6 1.3 0.07 Sweden 437 8.8 20 6 1.1 0.6 0.02 Gulf of Bothnia 491 3.8 8 1.1 0.5 0.03 Gulf of Finland 428 12.6 29 0.6 0.9 0.07 Gulf of Riga 137 4.0 29 0.6 2.5 0.07 Baltic Proper 572 55.0 96 0.5 2.8 0.15 Danish Straits and Kattegatt 111 8.8 79 1.0 2.7 0.08 Total 1740 84.2 48 0.7 1.5 0.07

References:

1. Hannerz, F. & Destouni, G. Spatial characterization of the Baltic Sea drainage basin and its unmonitored catchments. Ambio 35, 214-219 (2006).

2. HELCOM. The fourth Baltic Sea pollution load compilation (PLC-4). Helsinki, Finland: Helsinki Commission (2004).

3. HELCOM. Eutrophication in the Baltic Sea: An integrated assessment of the effects of nutrient enrichment in the Baltic Sea Region. BSEP Rep. 115B, Helsinki Comm., Helsinki (2009).

34

Table  SM3  

Information and data for the catchments and associated time periods of change in Figure 5. Catchment Area (km2) Time

period 1, 2 Hydro-climatic data and sources detailed in

Known main disturbance type

Circum-polar -disturbed

Churchill 287 000 1961-1990, 1991-2002

Bring and Destouni (2011)1

Hydropower2

Circum-polar -other Kolyma

526 000 1961-1990, 1991-2002

Bring and Destouni (2011)1

Indigirka 305 000 Yana 224 000 Lena 2 430 000 Olenek 198 000 Yenisey 2 440 000 Ob 2 950 000 Pechora 312 000 Sevemaya Dvina 348 000 Yukon 831 386 Mackenzie 1 680 000 Nelson 1 010 000

Swedish - disturbed Vättern 13 283

1901-1930, 1971-2000

Destouni et al. (2013)3

Agriculture3 Norrström 22 650 Agriculture3 Vänern 47 571 Agriculture3 Dalälven 25 058 Hydropower3 Indalsälven 23 842 Hydropower3 Ångermanälven 30 638 Hydropower3 Luleälven 24 949 Hydropower3

Swedish - other Piteälven 10 817 1901-1930,

1971-2000 Destouni et al. (2013)3

Öreälven 2 860 Sava River basin - disturbed

Kozluk 17 847 1931-1960, 1971-2000

Levi et al. (2015)4 Hydropower4

Sava River basin - other Slavonski Brod 54 718 1931-1960,

1964-1993 Levi et al. (2015)4

Mahanadi River basin - disturbed Whole basin 135 084

1901-1930, 1971-2000

Destouni et al. (2013)3 Irrigated agriculture3,5

Aral Sea drainage basin - disturbed Whole basin 1 888 810 1901-1930,

1971-2000 Destouni et al. (2013)3 Irrigated

agriculture3,6 Notes and References:

1. Bring, A., & G. Destouni. Relevance of hydro-climatic change projection and monitoring for assessment of water cycle changes in the Arctic. Ambio 40, 361-369 (2011); Khatanga basin from this paper is not included here due to lack of data for the 1961-1990 period.

2. Déry, S.J., Stieglitz, M., McKenna, E.C. & Wood, E.F. Characteristics and trends of river discharge into Hudson, James, and Ungava Bays, 1964–2000. Journal of Climate 18, 2540–2557 (2005).

3. Destouni, G., Jaramillo, F. & Prieto, C. Hydroclimatic shifts driven by human water use for food and energy production. Nature Clim. Change 3, 213–217 (2013).

4. Levi L., Jaramillo F., Andričević R. & Destouni, G. Hydroclimatic changes and drivers in the Sava River

35

Catchment and comparison with Swedish catchments. Ambio (2015, in press) DOI 10.1007/s13280-015- 0641-0.

5. Asokan, S.M., Jarsjö, J. & Destouni, G. Vapor flux by evapotranspiration: effects of changes in climate, land-use and water-use. J. Geophys. Res.-Atmospheres 115, D24102 (2010).

6. Shibuo, Y., Jarsjö, J. & Destouni, G. Hydrological responses to climate change and irrigation in the Aral Sea drainage basin. Geophys. Res. Lett. 34, L21406 (2007).