Modeling human-water-systems: Towards a comprehensive and spatially distributed assessment of co-evolutions for river basins in Central Europe

In the context of medium to long term river basin and flood risk management there is a growing need to improve the understanding of and the feedbacks between the driving forces "climate and socio-economy" and water systems, e.g. by combining scenarios of the global and regional climate change with those of developments of the society. Although progress has been made in the recent years, several methodological challenges still exist for the development of supra-regional Water Management Strategies. Those should take into account the social and environmental requirements across national, catchment, and sectoral boundaries in order to anticipate possible regional water problems or specific water usages. Progress refers to improvements in the availability of homogeneous hydro-meteorological and hydrological observation and modelling products, including refined scenarios in climate and impacts in the (natural) hydrological system. Main challenges refer e.g. to the availability of long time series of data and indicators representing the socio-economic part of the human-water-system as well as its integration into hydrological models at appropriate spatial and temporal scales. We make use of a variety of data resources to illustrate interrelationships between different constituents of the human-water-systems. Taking water storage for energy production as an example we present a first analysis on the co-evolution of socio-economic and hydrological indicators. The findings will serve as a starting point and validation target for the development of conceptual, but fully coupled sociohydrological models for selected sectors and regions. These models will be used to generate integrated scenarios of the climate and socio-economic change. Further steps leading to large scale, cross-sectoral and cross-catchment assessments will be discussed.


Introduction
The adequate availability of water is an important determinant for the socio-economic development -with water representing either an essential raw material or a threat e.g. in the case of flood events. At the same time the socio-economic development affects the water resources and the environment. Extensive reviews of challenges as well as guiding principles of developing integrated sociohydrological models for (i) system understanding (ii) forecasting and prediction as well as (iii) policy and decision-making have recently been given by Sivapalan & Blöschl (2015) and Blair & Buytaert (2016). With this study we intend to assess the need and the feasibility of socio-hydrological approaches for solving practical management questions (cf. iii). In the context of medium to long term river basin and flood risk management there is a growing need to improve the understanding of and the feedbacks between the water and the human component. Among others, scenarios of global and regional climate change need to be combined with those of developments of the society to come to "adaptive scenarios" instead of "climate driven impact scenarios". As a first step of an ongoing study we explore relevant human-water loops in a major part of Central Europe (Fig. 1). In this paper (i) regions where human interference with the water cycle is particularly strong will be highlighted by a hydrological modelling approach, and (ii) time scales and potential drivers of a human-water system will be explored using the Swiss water power sector as a case study. The spatial dimensions of water management -Redistribution of benefits and risks (7th International Water Resources Management Conference of ICWRS, Bochum IAHS -18-20 May 2016) 2 Study area Our study area covers a major part of Central Europe (approx. 800.000 km 2 ), namely the German river catchments including their international upstream parts (Fig. 1). In a global perspective this area is usually not mentioned among the regions facing water ressource problems. However, there are strong regional disparities of the natural water ressources, which roughly follow a southwestnortheast gradient with the Alps in the south being an important 'water tower'. There is also a distinct seasonal contrast in natural water availability. Under the influence of the Alps (extending to the upper Rhine and southern tributaries of the Danube) runoff is usually higher in the summer months, while in the remainder of the area winter runoff is higher. Climate change may lead to stronger seasonal contrasts of water availability in the latter case (e.g. Schneider et al., 2013;Nilson et al., 2014).  Nov.1976to Oct. 2005. Locations of dams and mining were compiled from various sources (among others Hydrological Atlas of Germany, HAD; Global Reservoir and Dam Database, GRanD; Lehner et al., 2011). Arrow marks the region selected as a case study Also, the water demand is unequally distributed regionally, being concentrated in a number of large agglomeration areas. The socio-economic systems in Central Europe are optimized to and dependent on relatively high and stable water resources. Consequently, the region is vulnerable to changes of water availability. Water resources and water use are thus regularly monitored on a national and continental level. Also, the "German Adaptation Strategy to Climate Change" introduces several indicators describing water ressorces (UBA 2015).

Data and Methods
Main challenges in socio-hydrological studies refer to the availability of long time series of data and indicators representing the socio-economic part of the human-water-system as well as the level of conceptualization and complexity at appropriate spatial and temporal scales that allow integrating them into hydrological models. We start assessing these challenges by identifying regions in our study area where human interference with the water cycle is particularly strong. The spatial dimensions of water management -Redistribution of benefits and risks (7th International Water Resources Management Conference of ICWRS, Bochum IAHS -18-20 May 2016) For that purpose, a hydrological model is applied for Central Europe ("LARSIM-ME") using a new homogeneous hydro-meteorological data product (DWD-BFG HYRAS) as input. The model has been parameterised by regionalization of model parameters from catchments that show no/weak human influence. We assume the simulated runoff to reflect natural conditions and compare them with observed runoff data and with an inventory of dams and mining areas the model domain.
Starting from this version of the water balance model, human influence is currently successively being implemented. This will be done using operation plans of dams and water transfers, pumping volumes from open pit mines and water abstraction rules. Acquisition of this kind of information over a large region as the overall model domain is an intricate task because of the diversity of institutions and sectors involved. In some cases detailed information is often not available at all because it is retained due to privacy reasons. Therefore, we evaluate methods to synthesize operation data for our model domain based on information from selected case studies that show an extraordinary good data situation. On top of data availability, each case study has to fulfil two prerequisites: (i) the sector has to be dependent on at least one component of the water balance (ii) the regional water balance is modified by the sector with at least one management measure.
In the following section we choose the case study "Water Power in Switzerland" as an example to highlight available data and to display time scales and drivers of change. Data resources are given in Table 1. Analyses with regard to co-evolution processes will be performed by comparing time series of these variables with water resources data as well as meteorological and hydrological data. The spatial dimensions of water management -Redistribution of benefits and risks (7th International Water Resources Management Conference of ICWRS, Bochum IAHS -18-20 May 2016) natural flow conditions. The dark grey regions show particular strong deviations between simulations and observations. These regions are most frequent in the River Elbe catchment, the Alpine subcatchments of the Rivers Danube and Rhine, and in some catchments of tributaries of the Lower River Rhine. The spatial distribution of clusters of mining activities and water storages match well these regions (Fig. 1) indicating important modifications of the natural hydrological system in Central Europe (transfer systems are not shown). Obviously, water balance simulations aiming to match the observed runoff require additional representation of human activities.

Case study "Water Power in Switzerland"
Dams are mainly constructed and operated in Switzerland for energy supply. In the context of water ressources monitoring, developments of water storage capacity of dams as well as of management rules are of main interest. Fig. 2 shows a long term perspective on water storage capacities in Switzerland, runoff at River Rhine gauging station Basel near the Swiss/German border as well potential climatic and socio-economic drivers of change. Available data allows this assessment on an annual basis for the period 1910 to 2012.  Table 1) At the beginning of the 20 th century storage capacities and electricity consumption were very low. At this time water power was used more or less mechanically in the vicinity of the rivers (Swiss Energy Council, 1987). In the following decades storage capacities and water power production increased continuously until the 1970 as a consequence of several developments: (i) a Swiss law targeting at the utilization of water power was released in 1918 (ii) technical innovations allowed the construction of larger dams and (iii) improved distribution of electricity over large distances. The latter also preconditioned the general electrification of the society and economy, which were both steadily growing. Since the mid of the 1970ies, there was no further increase of storage capacity and water power production. At the same time nuclear power was prioritized to meet the still growing The spatial dimensions of water management -Redistribution of benefits and risks (7th International Water Resources Management Conference of ICWRS, Bochum IAHS -18-20 May 2016) electricity demand (Swiss Energy Council, 2015). It may be noteworthy, that after the mid 1970ies onwards the annual minimum runoff at gauge Basel hardly did ever drop below 400 m 3 /s (cf. arrow in Fig. 2). This can to a large degree be attributed to the storage capacity implemented.
From 1995 onwards water power production data shows only minor long term increases (due to optimizaztion of existing power plants), but a stronger year-to-year variability. This is at least in part due to a change of the data source and the way the data were calculated (cf. references given in Table 1). Nevertheless, these data may give an indication that -although socio-economical drivers dominated the centennial development -meteorological and hydrological conditions affect water power production variability on a year-to-year scale. This coincidence is, however not very clear on the annual scale.
In a monthly view (Fig. 3), the individual components of this human-water system show very systematic annual cycles. Storages are filled during the summer when precipitation and snow melt water runoff is high. The maximum level of water storage is usually reached in September or October. During the winter season, when runoff is typically low, water is released. The water power production shows a vague pattern with one weak peak during the high runoff period in summertime and one stronger peak during the release period in wintertime. The overall consumption of electricity shows -on top of a long term increase up to the year 2006 -a rather regular pattern with a single major peak during the heating season in winter. The amplitudes of most variables vary only moderately and the pattern persists remarkably stable over the years. One exception is the winter of 2006/2007 which was extraordinary warm. This coincides with higher than average minimum water levels of storages due to lower electricity consumption (less heating).  The spatial dimensions of water management -Redistribution of benefits and risks (7th International Water Resources Management Conference of ICWRS, Bochum IAHS -18-20 May 2016)

Discussion and Conclusions
Clearly, potential water availability is the backbone of activity in the water power sector. However, technological developments were the main drivers for the onset and the cessation of the growth of water power production and the related modifications of the natural water system through reservoirs. The increase of water power production took place over three to four decades. Variability of natural drivers as reflected e.g. by air temperature and runoff manifests as second order variability in water power time series. This is visible on a year-to-year basis, but much more on a monthly timescale.
The steady seasonal pattern of reservoir operation represents a feature that could be described by comparatively simple models. This would allow us to generate storage data for the past as well as for climate impact studies. However, the long term socio-economy-driven developments tell us, that the pattern may be stationary only for limited time. Global climate change, a change of energy mix in Europe, and the ongoing deregulation of the energy market in Switzerland are possible candidates to cause changes in the operation of Swiss dams. Estimating the consequences of these so far unobserved phenomena remains a challenge. Similar case studies are underway in other regions of our model domain and examining other sectors such as urban water supply, agriculture, navigation etc. We attempt to build simple conceptual models representing those sectors which interfere most with the water cycle in Central Europe. By combining these models with the water balance model we try to provide explicit information on spatial and temporal scales that are suitable for supra-regional water resources management.