Panta Rhei - Non-stationarity in planning, designing, and operating urban drainage systems

The past one to two decades have witnessed the most dramatic changes in urban water management since the sewer systems were introduced in the 1850ies. There are hence many threats to a business-as-usual approach. Climatic changes caused by anthropogenic activities increase the occurrence rate and size of water related extremes, in northern Europe in particular leading to an increased risk of flooding.

Increasing the infrastructure capacity will lead to dry habitats, which is conflicting with the goal of maintaining liveable cities. The capital costs related to maintaining infrastructure are already high and increased funding will most likely only be allocated if the services are more visible in the urban space.

Hence the framing of urban water management is moving from a purely technical domain of ensuring hygienic barriers between polluted water and humans into a very complex framing of delivering many types of services. It is in this context that the urban pluvial flood risk must be considered. So while the key aspect of the thesis is technical the underlying theme is setting the stage for engaging with other stakeholder.

Only this way can we achieve what is needed and ensure that the impacts of flooding can be mitigated by proper planning and management. Urban pluvial flood risk has experienced a leap of knowledge within the last decade of research, on which this thesis focuses. The thesis presents the work that I have been involved in since 2008 reflecting both on the achievements that have been made by me and other scholars during this period as well as the work that still needs to be undertaken.

The key hazard of pluvial risk is extreme convective precipitation. In 2008 it was recognized that climatic changes would change precipitation patterns significantly, but how this would impact extreme precipitation was poorly understood. Today, it is well known that extreme precipitation will become more frequent and more severe, especially for high return periods and for sub-daily durations.

We have shown that for emission scenarios corresponding to 3 – 4°C increase in year 2100 compared to pre-industrial levels an increase of 20-40% can be expected over the next century in Denmark. The mechanisms are explored by using climate change models where the ability to model convective thunderstorms are poor.

Hence, the changes may also become much larger (or lower). We have also shown that higher temperature increases will lead to higher increases of extreme precipitation and vice versa. The results have been implemented into Danish design guidelines by recommending to account for the anticipated changes in precipitation extremes over the technical lifetime of the infrastructure and to make worst case assessments of the changes in case of critical urban functions being impacted.

The recommended changes in design intensities are lower than the observed changes during the past 40 years. We have shown that the observed change is primarily a result of natural variation of precipitation rather than an impact of emissions of greenhouse gasses. In the current climate, severe urban pluvial flooding will not occur concurrently with the other important water hazard in Denmark, sea surges.

We have shown that there nevertheless is a positive correlation between these hazards. Climatic changes will thus increase the severity of the hazards individually, but the rate of concurrent events will increase even more. This will exacerbate the impacts of these hazards severely in case this is not considered when designing water infrastructure.

The understanding of vulnerability and exposure has also improved significantly, which mean that we can today quantify urban pluvial flood risk well enough to make reliable decisions on feasible adaptation options. The description of vulnerability has improved by better access to data, because insurance providers see a benefit in providing good data for the modelling and, alas, much more data because of many recent floods.

The framing of vulnerability remains an issue because many vulnerabilities are difficult to express in monetary terms. Exposure to hazards is rather easy to model for present climate, but making projections are deeply uncertain because the future exposure is driven by urban development, which is deeply uncertain.

Therefore the exploration of future risks must be done by means of explorative modelling. This is a challenge to engineers, since it hinders traditional design of infrastructure, because no unique optimal solution can be identified based on a clear set of assumptions. The deep uncertainties should not hinder or delay decisions.

On the contrary, the increasing risk implies that in general decisions should be taken before they are recognized by the general society as being needed. The thesis suggests different ways of modelling and presenting information about present and future flood risk in ways that make it more comprehensible to other stakeholders.

It is the hope that this will help society in obtaining a suitable amount of information enabling good decisions on flood risk management by facilitating better communication between flood risk managers and other stakeholders. Historically flood risk has remained almost constant when normalized with economic development, i.e. in absolute value it has been rising exponentially.

We have very strong evidence that in the future also the hazards will increase at least linearly, perhaps even exponentially. The result will be a very high increase in flood risk unless structural changes in urban water management is implemented. The development of a comprehensive urban flood risk management framework that incorporates future changes into current decision-making as presented in this this thesis is a necessary step towards considering these risks as a result of human decisions and actions rather than something we should consider as a result of “natural” processes and hence something that should be accepted as-is.