Abstract. Flood events cause significant damage not only on the surface but also underground. Infiltration of surface water into soil, flooding through the urban sewer system and, in consequence, rising groundwater are the main causes of subsurface damage. The modelling of flooding events is an important part of flood risk assessment. The processes of subsurface discharge of infiltrated water necessitate coupled modelling tools of both, surface and subsurface water fluxes. Therefore, codes for surface flooding, for discharge in the sewerage system and for groundwater flow were coupled with each other. A coupling software was used to amalgamate the individual programs in terms of mapping between the different model geometries, time synchronization and data exchange. The coupling of the models was realized on two scales in the Saxon capital of Dresden (Germany). As a result of the coupled modelling it could be shown that surface flooding dominates processes of any flood event. Compared to flood simulations without coupled modelling no substantial changes of the surface inundation area could be determined. Regarding sewerage, the comparison between the influx of groundwater into sewerage and the loading due to infiltration by flood water showed infiltration of surface flood water to be the main reason for sewerage overloading. Concurrent rainfalls can intensify the problem. The infiltration of the sewerage system by rising groundwater contributes only marginally to the loading of the sewerage and the distribution of water by sewerage has only local impacts on groundCorrespondence to: T. Sommer (tsommer@dgfz.de) water rise. However, the localization of risk areas due to rising groundwater requires the consideration of all components of the subsurface water fluxes. The coupled modelling has shown that high groundwater levels are the result of a multi-causal process that occurs before and during the flood event.

Mining pit lakes can form in open cut mining pits that extend below the groundwater table. Final lake surface levels generally represent the greatest risk of pit lake closure to stakeholders through potential to overflow and discharge to regional surface water bodies and groundwater resources. An essential prerequisite for managing this risk is a good understanding of the lake's water budget. Pit lakes in the Collie Coal Basin ,Western Australia form a lake district currently consisting of 13 lakes exceeding a total volume of 200 GL of acid and metalliferous (AMD) degraded water. Given long-term risks for off-site contamination, regulatory agencies often rely on geochemical predictions of future pit lake water quality to evaluate closure strategies that protect the surrounding environment. Using an existing regional groundwater model, we modelled representative pit lake types in the Collie Lake District, southwestern Australia, to determine different regional groundwater abstraction regime effects on pit lake water levels. PITLAKQ was used to model three different lakes representing three distinct lake types identified by conceptual modelling: Historic (around 50 years old), New/Rehabilitated, and New/ Un-rehabilitated (both around 5-15 years old). An accurate representation of the water level-volume relationships was developed before all available data on major hydrological sinks and sources such as groundwater inflow/outflow, surface water inflow/outflow, as well as precipitation and evaporation were considered in lake water budget calculations. Although we found large deviations between measured and calculated water levels we could show reasonable limits for groundwater inflows and outflows by examining different scenarios. Reciprocally, this improved the groundwater model(s) suggesting coupling fine-scale pit lake models with groundwater models to identify the data quality for sinks and sources as an approach for other pit lake models. Our modelling scenarios showed that planned groundwater abstraction regime changes would lead to only limited changes in lake water depth compared to modelling uncertainties resulting from limited available data and the use of a regional groundwater model. This example illustrates pit lake modelling with low data availability still allows useful scenario testing under different operational scenarios.

No abstract