Center for Advanced Water Research

Antibiotic resistant bacteria (ARB) are emitted from both point and non-point sources such as wastewater treatment plants (WWTP) and farming sites. Accordingly, the level of antibiotic resistance found in surface waters was shown to be linked to the proximity and size of WWTP in several studies. Nevertheless, there is considerable uncertainty about the spread of antibiotic resistances in river networks at longer temporal/spatial scales. Using computer simulations, hypothetical landscapes will be subdivided into hot and cold spots for antibiotic resistances and at a highly abstracted level, spread and survival of ARB within and between hot and cold spots e.g on a landscape or catchment level will be analyzed. We intend to test that a geographic network structure could affect the probability of the spread and survival of ARBs. Distances between hot spots will also be analysed employing delaunay triangulations. This network approach will be compared to current algorithms used in epidemiology. The objective of this project is to gain new insights into the spread and survival of ARB in meso-scale streams (several 10 to 100 km²) through process-oriented model simulations. The project specifically aims at studying the role surface (dead zone) and/or sub-surface (hyporheic) transient storage for spread, survival and reproduction of relevant bacterial strains. Ideally, characteristic environmental conditions (e.g. regarding residence time, temperature, nutrients, etc.) are identified that increase the probability of ARB spread, survival and reproduction in river networks.

• Berendonk, T. U., Manaia, C. M., Merlin, C., et al. (2015). Tackling antibiotic resistance: the environmental framework. Nature Reviews Microbiology 13(5): 310-317.
• Seiler, C., & Berendonk, T. U. (2007). Heavy metal driven co-selection of antibiotic resistance in soil and water bodies impacted by agriculture and aquaculture. Role and prevalence of antibiosis and the related resistance genes in the environment: 101.

Water, nutrients, and pollutants move through the terrestrial ecosystem via complex networks of flowpaths across all spatial scales. Distributed models are valuable tools to assess the governing networks of surface and subsurface water pathways and associated matter fluxes within watersheds across spatial and temporal scales. Such models can help to identify critical source areas of diffuse water pollution and their connectivity via contrasting flowpaths to the river network within the network of connected hydrological areas. Furthermore, this helps to identify the optimal spatial implementation of best management practices to minimize the impairment of water resources. Within the frame of the proposed PhD project the world-wide applied watershed model SWAT (Soil Water Assessment Tool) will be further developed. The objective is to test and apply the recently developed landscape routing algorithm in SWAT to simulate water, sediment, and associated nutrient/pollutant flow across the network of landscape units (defined by relief, soil, and land use) in mesoscale watersheds. The project will be based on recent model modifications by Sun et al. (2015) and Rathjens et al. (2015). This SWAT version will be further developed by including algorithms for agricultural management and nutrient/pesticide fluxes. The modelling work will be supported by information gained from experimental field work (i.e. event-based sampling and hydrogeochemical analysis of dissolved and suspended materials). The new model will be implemented and tested in two mesoscale watersheds located in Germany (Vereinigte Weißeritz, Saxony) and in the USA (Eagle Creek, Indiana). The two watersheds are characterized by a typical pattern of mixed urban and rural/agricultural land use. The PhD student (based at TU Dresden) will test and apply model routines that provide an enhanced representation of water, sediment, and nutrient/pesticide fluxes. This will allow for a model-based identification of critical source areas for nutrient and pollutant transport/export at the landscape level via the governing networks of surface and subsurface pathways and to suggest best management practices to minimize pressures to water resources.

• Julich, S., L. Breuer & H. G. Frede (2012). Integrating heterogeneous landscape characteristics into watershed scale modelling. Advances in Geosciences 31: 31-38.
• Wahren, F. T., Julich, S., Nunes, J. P., Gonzalez-Pelayo, O., Hawtree, D., Feger, K.-H. & Keizer J. J. (2016). Combining digital soil mapping and hydrological modeling in a data scarce watershed in north-central Portugal. Geoderma 264 B: 350-362.
  

Water networks are systems in which multiple functional units are connected by physical links. For example, in technical water distribution networks, pipes, canals and other connections are shown by edges and fixed junctions (reservoirs, tanks, treatment plants, demand points) and pipe intersections represented by nodes. Natural water systems can also be described as networks consisting of functional units such as sources, reservoirs (surface and groundwater reservoirs), treatment (soil-aquifer recharge), demand nodes (irrigated agriculture) and links that transport water above and below the soil surface. The practice of designing each type of water network separately has generated specific optimization methods in each field. However, under the pressure of water scarcity multiple sources of water – such as desalinated water, treated wastewater and freshwater – with different qualities, quantities and unit production costs have to be considered in order to meet current and future water demand. Furthermore, the behavior of demand of the different consumers changes if many alternative sources of water exist. Therefore, a demand model has to be included into an optimization of the design and operation of a heterogeneous water system. This holds true for the design of small-scale self-sufficient hydrosystems as well as for the design of complex water networks on a regional level. The task of optimization is to identify robust and resilient design and operation strategies, i.e. to design networks such that they exhibit similar reliability under various conditions and to operate the network such, that the original performance is re-achieved after a failure or an overload. Solving the optimization problem of design and operation of a heterogeneous water network and to assess its robustness and resilience is challenging, particularly under different sources of uncertainty (e.g. climate, soil conditions and demand). To treat these challenges within a single simulation-optimization framework, it is required (i) to integrate different approaches of modeling of the single water networks into a parsimonious, unified network model, e.g. using a neural network approach as surrogate model; (ii) to provide a numerical solution for the resulting non-convex, mixed-integer, nonlinear optimization problem; (iii) to analyze the tradeoffs that occur between costs, productivity, robustness and reorganizing capability under conditions of change of boundary conditions for specific small-scale and more complex water networks; and (iv) to formulate a tractable probabilistic framework which avoids the considerable computational effort of Monte Carlo simulations since higher quantiles of performance (90% and above) are of interest.

• Subagadis, Y., Grundmann, J., Schütze, N. & Schmitz, G. (2014). An integrated approach to conceptualise hydrological and socio–economic interaction for supporting management decisions of coupled groundwater-agricultural systems. Environmental Earth Sciences 72(12): 4917-4933.
• Pham Van, T., Grundmann, J., Schütze, N. & Schmitz, G. (2012). Multi-objective design of a water distribution network using simulation-based optimization. In: Proceedings of the 14th Water Distribution Systems Analysis Conference, 2012.

In sewers, load peaks develop due to stormwater runoff. Particulate material is eroded from the catchments surface and from the sewer sediments causing higher particles concentrations and thus load increase over-proportional to the flow increase. Particles from the catchments surface are loaded with e.g. heavy metals and PAHs and are thus a major source of pollution in both combined and separate sewer systems. Dissolved compounds, e.g. ammonium originating mainly from urine, are pushed out from the combined sewer system with only little dilution in the initial phase of an event. As a consequence of the dynamic erosion and push-out effects, so-called shock loads develop for the wastewater treatment plant (WWTP) as well as for the receiving water via combined sewer overflow (CSO) or stormwater discharge. CSO and stormwater discharge typically develop in a phase where the receiving water still exhibits the base flow rate that prevailed before the rain event started, as the rain-runoff process is much faster in the urban catchment than in the river basin. Schindler et al. (2010) showed that rain events of moderate intensity may cause more serious and critical pollution impacts to receiving waters than extreme events.

While the dynamics of matter transport is rather well understood and described, the influence of the network structure on the extent of the load peak formation is unclear. Preliminary simulations showed that in wide and short sewer networks, a predefined rain event will create a more pronounced load peak than in long and narrow sewer networks. Obviously, the network structure matters and requires further evaluation. Departing from sewer network archetypes, structures developed with an algorithm and, ultimately, real networks will be investigated with regard to their influence on load dynamics and possible acute impacts to rivers, hampering their resilience. The questions addressed are the following:

• How does the sewer network structure affect load peaks?
• Can peaks be reduced by appropriate network structures, such as increasing the node degree, i.e. the mesh intensity?
• How should combined-sewer overflow-structures, stormwater discharge points and retention facilities be distributed in the catchment to minimise the impacts to rivers?
• Evaluate operation and control options related to network analysis in order to reduce effects and thus to better enable resilient behaviour of the affected receiving waters especially after an acute impact
• Develop a common model together with topic 8 to represent the dynamic interactions between the sewer and river networks with special focus on the impact and resilience pattern between them

Publications:

• Zhang, J., Wang J., Hua, P. & Krebs, P. (2015). The qualitative and quantitative source apportionments of polycyclic aromatic hydrocarbons in size dependent road deposited sediment. Science of the Total Environment 505: 90-101.

• Kaeseberg, T., Blumensaat, F., Zhang J. & P. Krebs (2014). Spülstoßeffekt gelöster und feststoffgebundener Antibiotika und resultierendes Umweltrisiko bei Mischwasserentlastungen. Prävention und Gesundheitsförderung 9:206-210.

Effective estimates of design precipitation and timely prediction of flash floods and storm water events very much relies on accurate, spatially and temporally dense precipitation observations and forecasts. Fractal characteristics of precipitation are traditionally utilized to explore design precipitation (e.g., possible amounts of heavy rain events for a given duration). However, for convective precipitation maximum values depend on the water budget of the local atmosphere and on feeding from the vicinity. This leads to extremes that depend on a network of moisture transports to the precipitating system. Traditionally, the official rain gauge network provides reliable, official information on perception rates, however, with a coarse spatial resolution. Additionally, official observation networks provide valuable observations about spatial and temporal distribution of precipitation intensity (e.g. RADOLAN by German Weather Service). Recent developments in low cost sensors and information technologies allow for a quick and cost efficient establishment of spatially dense gauging sensor networks, utilizing interested citizens as operators of these sensor networks. However, these networks are less reliable and suffer partly from poorer observation qualities. Here smart approaches are required to

• combine the different networks’ observations as a common virtual gauging sensor network and into information products which support forecasts on local, small scale extreme precipitation events,
• describe the different observation network qualities and uncertainties in such a way, that these description can be considered in further data processing ,
• enhance, define and prototype interface and data encoding specifications to support interoperable data fusion and processing,
• identify fractal characteristics of the spatial and temporal information which could support the derivation of design precipitation for extreme events and of information of small scale extreme precipitation forecasts.

Resulting methods and approaches need to be tested for applying them in different settings and catchments, while their fractal structures allow the transfer to different less densely observed regions. The studies should build on existing approaches such as interface specifications from the OpenGIS Sensor Web Enablement Standards, on (robust) geostatistical methods and recent concepts for sensor networks for environmental monitoring.

• Bernard, L., Mäs, S., Müller, M., Henzen, C. & Brauner, J. (2014). Scientific Geodata Infrastructures: Challenges, Approaches and Directions. International Journal of Digital Earth 7(7): 613-633.
• Kronenberg, R. & Bernhofer, C. (2015). A method to adapt radar-derived precipitation fields for climatological applications. Meteorologische Applications 22(3): 636–649.