Department Systems Analysis, Integrated Assessment and Modelling

In densely populated areas, surface waters are affected by many sources of pollution. Besides classical pollutants like nutrients and organic matter that lead to eutrophication, micropollutants from various point- and non-point sources are getting more attention by water quality managers. For cost-effective management an integrated assessment is needed that takes into account all relevant pollutants and all sources of pollution within a catchment. Due to the difficulty of identifying and quantifying sources of pollution and the need for considering long-term changes in boundary conditions, typically substantial uncertainty exists about the consequences of potential management alternatives to improve surface water quality. We therefore need integrated assessment methods that are able to deal with multiple objectives and account for various sources of uncertainty.
This paper aims to contribute to integrated, prospective water management by combining a) multi-criteria decision support methods to structure the decision process and quantify preferences, b) integrated water quality modelling to predict consequences of management alternatives accounting for uncertainty, and c) scenario planning to consider uncertainty from potential future climate and socio-economic developments, to evaluate the future cost-effectiveness of water quality management alternatives at the catchment scale. It aims to demonstrate the usefulness of multi-attribute value functions for water quality assessment to i) propagate uncertainties throughout the entire assessment procedure, ii) facilitate the aggregation of multiple objectives while avoiding discretization errors when using categories for sub-objectives, iii) transparently communicate the results. We show how to use such multi-attribute value functions for model-based decision support in water quality management.
We showcase the procedure for the Mönchaltorfer Aa catchment on the Swiss Plateau. We evaluate ten different water quality management alternatives, including current practice, that tackle macro- and micropollutants from a wide spectrum of agricultural and urban sources. We evaluate costs and water quality effects of the alternatives under four different socio-economic scenarios for the horizon 2050 under present and future climate projections and visualize their uncertainty. While the performance of alternatives is catchment specific, the methods can be transferred to other places and other management situations. Results confirm the need for cross-sectoral coordination of different management actions and interdisciplinary collaboration to support the development of prospective strategies to improve water quality.

Increasing demand for food is driving a worldwidetrend of agricultural input intensification. However, there is nocomprehensive knowledge about the interrelations betweenpotential yield gains and environmental trade-offs that wouldenable the identification of regions where input-driven intensificationcould achieve higher yields, yet with minimal environmentalimpacts. We explore ways of enhancing global yields, while avoidingsignificant nitrogen (N) emissions (N_e) by exploring a range of Nand irrigation management scenarios. The simulated responses ofyields and N_e to increased N inputs (N_in) and irrigation show highspatial variations due to differences in current agricultural inputsand agro-climatic conditions. Nitrogen use efficiency (NUE) ofyield gains is negatively correlated with incremental N_e due to N_in additions. Avoiding further intensification in regions wherehigh fractions of climatic yield potentials, ≥ 80%, are already achieved is key to maintain good NUE. Depending on theintensification scenarios, relative increases in N_e could be reduced by 0.3−29.6% of the baseline N_e with this intensificationstrategy as compared to indiscriminate further intensification, at the cost of a loss of yield increases by 0.2−16.7% of the baselineyields. In addition, irrigation water requirements and N_in would dramatically decrease by considering this intensificationstrategy.

Freshwater ecosystems are increasingly under threat as they are confronted with multiple anthropogenic impairments. This calls for comprehensive management strategies to counteract, or even prevent, long-term impacts on habitats and their biodiversity, as well as on their ecological functions and services. The basis for the efficientmanagement and effective conservation of any ecosystem is sufficient knowledge on the state of the systemand its response to external influence factors. In freshwater ecosystems, state information is currently drawnfrom ecological assessments at the reach or site scale. While these assessments are essential, they are not sufficientto assess the expected outcome of different river restoration strategies, because they do not account for importantcharacteristics of the whole river network, such as habitat connectivity or headwater reachability. This isof particular importance for the spatial prioritization of restoration measures. River restoration could be supportedbest by integrative catchment-scale ecological assessments that are sensitive to the spatial arrangementof river reaches and barriers. Assessments at this scale are of increasing interest to environmental managersand conservation practitioners to prioritize restoration measures or to locate areas worth protecting. We presentan approach based on decision support methods that integrates abiotic and biotic ecological assessments at thereach-scale and aggregates them spatially to describe the ecological state of entire catchments. This aggregationis based on spatial criteria that represent important ecological catchment properties, such as fish migration potential,resilience, fragmentation and habitat diversity in a spatially explicit way.We identify the most promisingassessment criteria from different alternatives based on theoretical considerations and a comparison with biologicalindicators. Potential applications are discussed, particularly for supporting the strategic, long-term planningand spatial prioritization of restoration measures.

International food trade entails virtual water flows across trading partners. It has been proposed to attenuate regional water scarcity by importing water‐intensive commodities from water‐abundant regions. In addition to alleviating water scarcity in virtual water importing countries, existing studies have reported that food trade also generates global water savings. However, little is known how these global water savings may alleviate water scarcity, which is more relevant to the sustainable use of water resources than only assessing the savings. In this paper, we conducted a comprehensive review on studies of water savings and losses associated with food trade on different spatial scales. We found that the concept of global water savings is built on the disparities in water productivity across countries, whereas the regional water savings measure the inflows of virtual water trade. The significance of water savings is dimmed by the fact that the savings are often not driven by water scarcity. Meanwhile, lacking policy relevance impairs the usefulness of water saving accounting. Future studies should link water savings to alleviating water scarcity at various levels. The water saving accounting needs to go to finer scale, for example, to subnational and river basin scales, to support real water resource management. In the meantime, interdisciplinary efforts are necessary to enhance the water savings as a holistic measure for addressing water scarcity on regional and global scales.

This study investigates Bayesian signature-domain inference of hydrological models usingApproximate Bayesian Computation (ABC) algorithms, and compares it to ‘‘traditional’’ time-domain inference.Our focus is on the quantification of predictive uncertainty in the streamflow time series and onunderstanding the information content of particular combinations of signatures. A combination of syntheticand real data experiments using conceptual rainfall-runoff models is employed. Synthetic experiments demonstrate:(i) the general consistency of signature and time-domain inferences, (ii) the ability to estimatestreamflow error model parameters (reliably quantify streamflow uncertainty) even when calibrating in thesignature domain, and (iii) the potential robustness of signature-domain inference when the (probabilistic)hydrological model is misspecified (e.g., by unaccounted timing errors). The experiments also suggest limitationsof the signature-domain approach in terms of information loss when general (nonsufficient) statisticsare used, and increased computational costs incurred by the ABC implementation. Real data experimentsconfirm the viability of Bayesian signature-domain inference and its general consistency with time-domaininference in terms of predictive uncertainty quantification. In addition, we demonstrate the utility of theflashiness index for the estimation of streamflow error parameters, and show that signatures based on theFlow Duration Curve alone are insufficient to calibrate parameters controlling streamflow dynamics. Overall,the study further establishes signature-domain inference (implemented using ABC) as a promising methodfor comparing the information content of hydrological signatures, for prediction under data-scarce conditions,and, under certain circumstances, for mitigating the impact of deficiencies in the formulation of thepredictive model.