Revision of urban drainage design rules after assessment of climate change impacts on precipitation extremes at Uccle, Belgium
Introduction
The importance of climate change impact studies on urban drainage becomes widely accepted (Butler et al., 2007, Arnbjerg-Nielsen, 2012, Wu et al., 2012). Also urban wastewater studies more frequently take the potential impacts of the changing climatic conditions into account (Astaraie-Imani et al., 2012, Freni et al., 2012, Langeveld et al., 2013). Conducting climate change studies on urban drainage is, however, extremely complicated, mainly because of the small (temporal and spatial) scales of hydrological processes in urban catchments. Climate model simulations, which are the standard means to project future climate conditions, still remain relatively coarse in space and time resolution and are unable to describe accurately the rainfall process at the fine scales of urban drainage systems (Willems et al., 2011, Willems et al., 2012, Arnbjerg-Nielsen et al., in press). Next to the scale difference, current state-of-the-art climate models have limitations in the accuracy of describing precipitation extremes. This is due to a poor description of the non-stationary phenomenon during a convective storm leading to the most extreme events on a local scale (Dibike et al., 2008, Arnbjerg-Nielsen, 2012). To bridge the gaps between the climate model scales and the local urban drainage scales and to account for the inaccuracies in describing precipitation extremes, downscaling methods and bias-correction methods are commonly used in practice (Willems et al., 2011, Willems et al., 2012, Arnbjerg-Nielsen et al., in press). Most promising are the statistical downscaling methods. They apply a statistical model to transfer the coarse scale and bias in the precipitation results of climate models (the “predictor” variables) to the small scale rainfall (the “predictand” variable), and thus involve both bias correction and downscaling. The statistical model so far can only be based on historical data, thus assuming that the transfer from the predictors to the predictands will not significantly change under changing climatic conditions. There is no need to say that the application of statistical downscaling gives rise to additional uncertainties in the climate change impact results (next to the uncertainties induced by the climate models and the urban drainage impact models) (Arnbjerg-Nielsen, 2012). Recent developments attempted to reduce the uncertainty related to the statistical downscaling, making use of rainfall time-scaling laws (Nguyen et al., 2008a, Nguyen et al., 2008b), stochastic rainfall generators (Onof and Arnbjerg-Nielsen, 2009, Sunyer and Madsen, 2009, Sunyer et al., 2012), and transfer functions that depend on climate model process variables such as cloud cover and precipitation type (Olsson et al., 2012).
Despite these new developments, assumptions still have to be made, which might differ from method to method. Given that some assumptions cannot be tested (because related to the unknown future climate conditions), it would be advisable to apply several downscaling methods and compare the results. This was done by Arnbjerg-Nielsen, 2012, Sunyer and Madsen, 2009, Sunyer et al., 2012, and Willems and Vrac (2011), the latter for the specific climate conditions for Belgium, considering downscaling useful for urban drainage impact applications. This paper builds further on the recent research innovations in statistical downscaling, by extending the quantile perturbation based downscaling method, in the version by Willems and Vrac (2011). Whereas in that version the quantile perturbation factors were based on the empirical rainfall quantiles, they are in this paper based on rainfall extreme value distributions. A second new element is the bias correction to climate change factors that are based on an ensemble set of RCM runs, but which are nested in a limited number of GCMs or forced based on a limited number of greenhouse gas emission scenarios. A third new element is the development of high-mean-low tailored climate scenarios. The results are moreover extended from 17 runs with the same GCM and one greenhouse gas emission scenario in Willems and Vrac (2011) to a much larger set of 44 + 69 RCM/GCM runs after three different emission scenarios. Next to these innovations in methodology and results, the final goal of the research presented in this paper was to revise the hydrological design parameters, which are currently used in the guidelines for the design of urban drainage systems in the Flanders region of Belgium. The revision involves extrapolation of the design rainfall statistics, taking into account the current knowledge on future climate change trends till 2100. Uncertainties in these trend projections are assessed after statistically analysing and downscaling a large number of climate model simulation results. Based on the climate scenarios and related changes in rainfall statistics, changes in flood frequencies of sewer systems and overflow frequencies of storage facilities have been quantified. Also the change in storage capacity required to exceed a given overflow return period, has been calculated, for a range of return periods and infiltration or throughflow rates.
Section snippets
Statistical downscaling: quantile perturbation versus weather typing based methods
Willems and Vrac (2011) compared several downscaling methods, each belonging to one of the following two classes: quantile perturbation methods and weather typing based methods. These two classes of methods involve assumptions that are completely different. The quantile perturbation methods apply relative changes to the rainfall intensities, based on the ratios of rainfall quantiles derived from a future period (climate model scenario simulation) over the corresponding quantiles in the
Changes in design rainfall
The current urban drainage design guidelines are based on rainfall statistics, derived for Uccle (Brussels). Given that the rainfall statistics are nearly uniform in the northern flat Flanders region of Belgium, except for the near-coastal zone (Baguis et al., 2009), the Uccle rainfall statistics are considered valid for any inland place in Flanders. At Uccle, a unique dataset of 10-min rainfall intensities is available since 1898 (same instrument and same place). This Uccle series, more
Changes in sewer impacts and storage design rules
While the design storms are most useful to simulate the impact on sewer surcharge and flooding (or for sewer design applications), they are less useful to study the impact on (combined) sewer overflow frequencies, volumes or sediment/pollution loads. They are also less useful in the design of storage or other sewer ancillary facilities. Accurate estimation of sewer overflow frequencies indeed requires long-term simulations (Harremoës, 1988, Rauch et al., 2002, Butler and Davies, 2010). In
On the downscaled climate scenarios
Hydrological design parameters, which are currently used in the guidelines for the design of urban drainage systems have been revised for the Flanders region of Belgium. The revision involved extrapolation of the design rainfall statistics, taking into account the current knowledge on future climate change trends (till 2100). Uncertainties in these trend projections have been assessed after statistically analysing and downscaling by a quantile perturbation tool based on a broad ensemble set of
Acknowledgements
This research is the result of a research project for the Flemish Environment Agency (VMM). It made use of the climate change scenarios prepared within the scope of the CCI-HYDR Project for the Belgian Science Policy Office, which was coordinated by the author in cooperation with the Royal Meteorological Institute (RMI) of Belgium. The author wants to sincerely thank Dr. Pierre Baguis, ir. Emmanuel Roulin and Dr. Victor Ntegeka for their contributions to the project, and Dr. Alexander Bakker of
References (44)
- et al.
Assessing the combined effects of urbanisation and climate change on the river water quality in an integrated urban wastewater system in the UK
J. Environ. Manage.
(2012) - et al.
Bias correction of high resolution regional climate model data
J. Hydrol.
(2012) Stochastic models for estimation of extreme pollution from urban runoff
Water Res.
(1988)- et al.
Climate change and urban wastewater infrastructure: there is more to explore
J. Hydrol.
(2013) - et al.
Resampling of regional climate model output for the simulation of extreme river flows
J. Hydrol.
(2007) - et al.
Applying climate model precipitation scenarios for urban hydrological assessment: a case study in Kalmar City, Sweden
Atmos. Res.
(2009) - et al.
Quantification of anticipated future changes in high resolution design rainfall for urban areas
Atmos. Res.
(2009) - et al.
A comparison of different regional climate models and statistical downscaling methods for extreme rainfall estimation under climate change
Atmos. Res.
(2012) Compound IDF-relationships of extreme precipitation for two seasons and two storm types
J. Hydrol.
(2000)- et al.
Statistical precipitation downscaling for small-scale hydrological impact investigations of climate change
J. Hydrol.
(2011)
Quantification of climate change effects on extreme precipitation used for high resolution hydrologic design
Urban Water J.
Climate change scenarios for precipitation and potential evapotranspiration over central Belgium
Theoret. Appl. Climatol.
Empirical–Statistical Downscaling
Urban Drainage
Sewer storage tank performance under climate change
Water Sci. Technol.
Uncertainty analysis of statistically downscaled temperature and precipitation regimes in Northern Canada
Theor. Appl. Climatol.
Role of modeling uncertainty in the estimation of climate and socioeconomic impact on river water quality
J. Water Resour. Plann. Manage.
Unexpected rise in extreme precipitation caused by a shift in rain type?
Nature Geosci.
Cited by (135)
Identification of the critical factors in flood vulnerability assessment based on an improved DEMATEL method under uncertain environments
2024, International Journal of Disaster Risk ReductionA probabilistic assessment of urban flood risk and impacts of future climate change
2023, Journal of Hydrology