Geological Survey of Denmark and Greenland Bulletin 28, 2013, 33-36 33 Assessing urban groundwater table response to climate change and increased stormwater infiltration Mark T. Randall, Lars Troldborg, Jens Christian Refsgaard and Jacob B. Kidmose The global climate is expected to show continued warming throughout the coming century. As a direct consequence of higher temperatures, the hydrological cycle will undergo significant changes in the spatial and temporal distribution of precipitation and evapotranspiration. In addition to more frequent and severe droughts and floods, climate change can affect groundwater recharge rates and groundwater table elevation (Bates et al. 2008). Some previous studies of climate change impact on groundwater have suggested alarming reductions in ground- water recharge and lowering of water tables. Other studies, especially those focusing on regions of higher latitudes, have indicated a potential rise in water tables due to increased pre- cipitation and recharge (Scibek & Allen 2006; Woldeamlak et al. 2007). In addition to changes in precipitation patterns, a shift in stormwater infrastructure design may also alter the hydro- logic cycle of urban areas. In recent years, there has been a growing trend towards adoption of low-impact development practices managing stormwater runoff. These practices aim to mitigate the impacts of urbanisation such as increased runoff volume, higher peak runoff flows, lowered water tables and reduced water quality (Prince George’s County 1999). In contrast to conventional stormwater infrastruc- ture, which is designed to rapidly collect and convey runoff, low-impact development practices are designed to slow run- off, remove pollutants and evapotranspirate and infiltrate runoff locally. In recent years, numerous modelling studies have inves- tigated the potential impact of stormwater infiltration on groundwater levels. Gobel et al. (2004) used a combination of models (GwNeu, HYDRUS-2D, SPRING) to demonstrate that the installation of infiltration practices across an urban catchment area in Germany could raise the groundwater sur- face by up to 2.3 m in some locations. In another catchment scale study, Maimone et al. (2011) used the modelling code DYNFLOW to show that the future groundwater table may eventually stabilise up to 1.5 m higher than its current level in parts of Philadelphia, if the city’s plan to alter 40% of its impervious areas into so-called ‘green’ stormwater recharge areas is completed. Thompson (2010) used HYDRUS-2D to demonstrate that a stormwater infiltration basin could cause up to 1.3 m of localised groundwater mounding. In yet an- other study, Endreny & Collins (2009) used MODFLOW to show that rain gardens installed throughout a residential catchment area could raise the steady-state groundwater ta- ble by up to 1.1 m. The studies mentioned above have investigated ground- water level response to either changes in climate or storm- water management infrastructure. However, to the authors’ knowledge no studies have investigated the concurrent ef- fects of both alterations on the urban hydrologic cycle. In urban areas, it is necessary to determine the potential mag- nitude of the combined impact, as a steep rise in ground- water level can damage building foundations and subsurface infrastructure due to flooding and buoyancy forces (Gobel et al. 2004; Vázquez-Suñé et al. 2005). This study aims to assess the potential changes in groundwater response caused by both increased precipitation and widespread instalment of stormwater infiltration infrastructure in the city of Silke- borg, Denmark, using the MIKE SHE model. Change of groundwater level at the planned location of a new motorway in Silkeborg is the focus of this study as portions of the con- Fig. 1. The Silkeborg study area and the proposed course of the motorway. Inset: the location of Silkeborg in Jylland. © 2013 GEUS. Geological Survey of Denmark and Greenland Bulletin 28, 33–36. Open access: www.geus.dk/publications/bull 1 kmEnd of pipe recharge area Pervious area Impervious area Motorway Jylland Silkeborg c. 9°40´E c. 56°11´N 3434 struction are expected to come critically close to the present high groundwater table in that area. Knowledge of the mag- nitude of potential groundwater changes is essential because improved drainage measures and increased use of concrete will significantly raise the costs of the new motorway. Study area The city of Silkeborg has a population of c. 43 000 inhabit- ants and is located in the central part of Jylland, Denmark (Fig. 1). The focus of this study is just north of the river Gudenåen, where a portion of the new motorway will be constructed c. 6 m below the present terrain surface. The surficial geology is dominated by coarse-grained, postgla- cial, sandy sediments that form an upper unconfined aquifer with a vertical extent of 10–15 m. The average precipitation in Silkeborg during the period 1961–1990 was 903 mm per year, and the average potential evapotranspiration was 546 mm per year. The average monthly temperature during that period was 15.2°C in July/August and −0.3°C in January/ February (Kidmose et al. 2013). Methods Hydrological models – MIKE SHE is a deterministic, fully- distributed and physically based model software capable of simulating surface and subsurface hydrological processes. The Danish National Water Resources Model (DK-model) is based on MIKE SHE and incorporates national data on geology, soil type, land use, topography, river network ge- ometry, water abstraction and climate. The Silkeborg model is a 100 m grid local model using hydraulic head boundary conditions from the 500 m grid DK-model. A 9.2 km2 area within the 103 km2 Silkeborg model, which encompasses the new motorway construction and the greater part of the ur- banised surroundings, was chosen for the current study (Fig. 1). Details on the development, calibration and validation of the DK-model and the Silkeborg model are found in Højberg et al. (2013) and Kidmose et al. (2013), respectively. Six dif- ferent model scenarios have been evaluated (Table 1). Stormwater infiltration modelling – The Silkeborg study area consists of 65.5% pervious and 34.5% impervious cover. In the scenarios with conventional drainage stormwater infra- structure (i.e. the ‘CD’ scenarios), 100% of the precipitation on impervious cells was routed directly to the river system (Fig. 2A). Precipitation on impervious cells had one time step (i.e. one day) to infiltrate or evapotranspirate. At the end of the time step, any water in excess of a detention storage of 4.7 mm (based on calibration results) was routed overland to adjacent cells based on topography. It is assumed that the CD-2010 scenario is representative of Silkeborg’s current cli- mate and stormwater conditions. In the end of pipe recharge (EPR) scenarios (Fig. 2B), it was assumed that 10.7% (34 ha) of the city’s pervious area has been turned into end of pipe stormwater infiltration ponds (Figs 1, 2). Model cells which were assumed to contain in- Fig. 2. Three model scenarios for stormwater drainage infrastructure. Scenario name Climate data input Stormwater infrastructure CD-2010 Recorded 1991–2010 Conventional drainage to river system EPR-2010 Recorded 1991–2010 End of pipe infiltration ponds LAR-2010 Recorded 1991–2010 Local area recharge CD-2100 Projected 2081–2100 Conventional drainage to river system EPR-2100 Projected 2081–2100 End of pipe infiltration ponds LAR-2100 Projected 2081–2100 Local area recharge Table 1. Summary of model scenarios A B C Conventional drainage End of pipe recharge Local area recharge 35 filtration ponds were assigned detention storage of 500 mm to represent the storage depth of the pond. In the EPR sce- narios, precipitation which would normally be applied to impervious cells was reduced to zero, and the equivalent vol- ume of precipitation was instead evenly distributed over the infiltration pond cells via an increase in precipitation applied to those cells. In the EPR scenarios there were 9.3 times as much impervious drainage area as infiltration pond area, so the infiltration pond model cells had 1030% (i.e. 100% + 9.3 × 100%) of the actual rainfall applied to them. This method of manipulating precipitation to simulate the collection of stormwater in specialised infiltration areas on a city-wide scale is similar to the modelling strategy used by Holman- Dodds et al. (2003). The local area recharge (LAR) scenarios represent a sys- tem where stormwater is managed at the level of individual plots through any combination of infiltration practices, each no more than tens of metres across. It was assumed that infil- tration possibilities are numerous and located in close prox- imity so that at the scale of the model, each cell effectively behaves as a pervious cell. Therefore, all paved areas were given properties identical to the pervious areas with infiltra- tion rates controlled by the underlying soils. Climate input – Precipitation, temperature and evapotran- spiration data from the Danish Meteorological Institute from 1991 to 2010 were used as input to the ‘2010’ model scenarios. The input climate data for the ‘2100’ scenarios were generated by applying correction factors based on nine climate model projections from the ENSEMBLES project (Christensen et al. 2009) to present-day climate data. Fur- ther information on the Delta Change downscaling method used can be found in Seaby et al. (2013). To generate the re- sults for each of the three ‘2100’ infrastructure scenarios, the model was run nine times (once for each of the nine climate model projections), and the results averaged. Results Water table elevation – Average groundwater elevations along the area planned for the motorway construction were extracted from the MIKE SHE model results (Fig. 3). Areas where the solid black line (i.e. the motorway surface) drops below the water table indicate portions of the motorway which could be flooded by groundwater. In the CD-2010 scenario, a stretch of 160 m of motorway is below the average water table. In the CD-2100 scenario, the average groundwater table elevation is raised by 0.08 m, and the length of motorway surface at risk is extended to 180 m. Hundreds of metres of the proposed motorway are potentially flooded in the LAR-2010 and the LAR-2100 scenarios where the average water table rose 0.48 and 0.55 m above CD-2010 levels, respectively. The highest average water tables of 1.15 and 1.19 m above CD-2010 occur in the EPR-2010 and EPR-2100 scenarios, which would both put a stretch of nearly 1 km of the proposed motorway at risk. The results indicate that the impact of climate change (i.e. the difference between the ‘2010’ and the ‘2100’ scenarios) is small compared to the impact of extensive implementa- tion of either local area or end of pipe stormwater infiltra- tion practices. Only average water tables are presented here to compare the relative impacts of different model scenarios. However, maximum water tables could put much longer sec- tions of the motorway at risk and will therefore be considered in the final design of the motorway. Water balance – Average yearly volumes of precipitation, evapotranspiration, recharge and overland flow were calcu- lated for the 1991–2010 time period for each stormwater in- Fig. 3. Average modelled groundwater table elevations along the 2000 m of projected motorway at Silkeborg. The results are relative to CD-2010. Model scenario Mean (mm/year, 1991–2010) Precipitation Evapotranspiration Recharge Overland flow Baseflow CD 911 319 304 292 8 LAR 911 441 463 11 15 EPR 911 311 588 19 29 Table 2. Catchment water balances for different stormwater infrastructure scenarios LAR-2010 LAR-2100 EPR-2010 EPR-2100 Motorway surface CD-2100 NW SE –0.5 0.5 1.0 1.5 3.0 2.5 2.0 0.0G ro un dw at er t ab le ( m ) 0 500 1000 1500 2000 Distance along subsurface motorway stretch (m) 3636 frastructure scenario using MIKE SHE’s water balance tool (see Table 2). Evapotranspiration was greater in the LAR sce- nario, due to the much larger evaporation surface available. Recharge was much higher in both infiltration scenarios than in the CD scenario. Overland flow, or the volume of water which flows directly into the river system, was very small in both the infiltration scenarios in comparison to the CD scenario which routed all water from impervious areas into the nearest stream. Baseflow was highest in the EPR scenario, followed by the LAR scenario and finally the CD scenario, as would be expected based on the relative recharge volumes in these scenarios. Summary and conclusions Previous studies have reported groundwater level rise due to either climate change (Scibek & Allen 2006; Woldeamlak et al. 2007) or stormwater infiltration practices (Gobel et al. 2004; Maimone et al. 2011). However, these two changes to the urban hydrologic cycle are typically not assessed in an in- tegrated way as in this study. The modelling results present- ed in this paper are within the ranges of the above studies, i.e. tens of centimetres due to climate change and potentially more than 1 m due to the widespread adoption of stormwater infiltration practices. However, these results are specific to the Silkeborg motorway and it is expected that the relative magnitude of the impact due to climate change and storm- water infiltration could vary greatly under different climatic and geological regimes. Stormwater infiltration practices are often regarded as a form of climate change adaptation in the field of stormwater management as they can help to accommodate the higher intensity and larger volume precipitation events expected in the future. However, as the results of this study indicate, these same practices amplify other problems associated with climate change (i.e. groundwater table rise). The study clearly shows the need for integrated research of urban hydrology, and communication between hydrogeologists, stormwater engineers, planners and policy makers. Acknowledgement We thank the Danish Road Directorate for funding this study. References Bates, B., Kundzewicz, Z., Wu, S. & Palutikof, J. 2008: Climate change and water. Intergovernmental Panel on Climate Change, Technical Pa- per 6, 200 pp. Geneva: IPCC. Christensen, J.H., Rummukainen, M. & Lenderink, G. 2009: Formula- tion of very-high-resolution regional climate model ensembles for Eu- rope [Research Theme 3]. ENSEMBLES: Climate change and its im- pacts: summary of research and results from the ENSEMBLES project, 47–58. Exeter, UK: Meteorological Office Hadley Centre. Endreny, T. & Collins, V. 2009: Implications of bioretention basin spa- tial arrangements on stormwater recharge and groundwater mounding. Ecological Engineering 35, 670–677. Gobel, P. et al. 2004: Near-natural stormwater management and its effects on the water budget and groundwater surface in urban areas taking ac- count of the hydrogeological conditions. Journal of Hydrology 299, 267–283. Højberg, A.L., Troldborg, L., Stiesen, S., Christensen, B.B.S. & Henrik- sen H.J. 2013: Stakeholder driven update and improvement of a na- tional water resources model. Environmental Modelling & Software 40, 202–213. Holman-Dodds, J.K., Bradley, A.A. & Potter, K.W. 2003: Evaluation of hydrologic benefits of infiltration based urban storm water man- agement. Journal of the American Water Resources Association 39, 205–215. Kidmose, J., Refsgaard, J.C., Troldborg, L., Seaby, L.P. & Escrivà, M.M. 2013: Climate change impact on groundwater levels: ensemble mod- elling of extreme values. Hydrology and Earth System Sciences 17, 1619–1634. Maimone, M., O’Rourke, D.E., Knighton, J.O. & Thomas, C.P. 2011: Potential impacts of extensive stormwater infiltration in Philadelphia. Environmental Engineer 14, 29–39. Prince George’s County 1999: Low-impact development design strategies: an integrated design approach, 150 pp. Prince George’s County, MD: Department of Environmental Resources. http://water.epa.gov/pol- waste/green/upload/lidnatl.pdf Scibek, J. & Allen, D. 2006: Comparing modelled responses of two high- permeability, unconfined aquifers to predicted climate change. Global and Planetary Change 50, 50–62. Seaby, L.P., Refsgaard, J.C., Sonnenborg, T.O., Stisen, S., Christensen, J.H. & Jensen, K.H. 2013: Assessment of robustness and significance of climate change signals for an ensemble of distribution-based scaled climate projections. Journal of Hydrology, http://dx.doi.org/10.1016/j. jhydrol.2013.02.015 Thompson, A., Nimmer, M. & Misra, D. 2010: Effects of variations in hy- drogeological parameters on water-table mounding in sandy loam and loamy sand soils beneath stormwater infiltration basins. Hydrogeology Journal 18, 501–508. Vázquez-Suñé, E., Sanchez-Vila, X. & Carrera, J. 2005: Introductory re- view of specific factors influencing urban groundwater, an emerging branch of hydrogeology, with reference to Barcelona, Spain. Hydroge- ology Journal 13, 522–533. Woldeamlak, S., Batelaan, O. & de Smedt, F. 2007: Effects of climate change on the groundwater system in the Grote-Nete catchment, Bel- gium. Hydrogeology Journal 15, 891–901. Authors’ addresses M.T.R., Computational Hydraulics International, 147 Wyndham Street North, Suite 202, Guelph, Ontario, N1H 4E9 Canada. E-mail: mark@chiwater.com L.T., J.C.R. & J.B.K., Geological Survey of Denmark and Greenland, Øster Voldgade 10, DK-1350 Copenhagen K, Denmark. http://dx.doi.org/10.1016/j.jhydrol.2013.02.015 http://dx.doi.org/10.1016/j.jhydrol.2013.02.015