Lake Mälaren is the third largest lake in Sweden (see Figure 1), with 24 cities and municipalities bordering the lakefront, including Stockholm. The lake is the drinking water source for two million people and also serves as a major port (see Figure 2).

The risk of flooding in Lake Mälaren is very high due to milder winters and heavier rainfalls. The winter of 2000 was unseasonably wet and much more water flowed into the lake than could be let out through the sluices that manage the water levels. There was extensive local surface water flooding and flooding nearly reached the central underground station tunnel of Stockholm.

Local disruption was thankfully a one-off at that point in time; however, it is forecast to be less of a one-off in a changed climate. Forecasts show that winter rainfall in Stockholm will increase, there will be greater year-on-year variations in precipitation, and the snowmelt will be greater. So there will be much more water flowing into the lake in the future.

The Study

The Swedish National Security Administration realised the need to prepare for increased flood risk and wanted to understand where the actual risks lie and what the critical and vulnerable points are. They wanted to create a database for guidance on what to do if flooding increases in this area and hired WSP to conduct a study to understand in detail how the area could be affected.

Together with the client, the boundary of the study was defined. The focus was to study the flood consequences of such activities that are important to the functioning of society. The WSP team, in a desk-top and on-site study, looked at a wide variety of installations and service providers – such as banks, hospitals, IT, phones, police - and identified 236 objects – installations, utilities, services, pipes – that are important to society in the flood risk area.

The study revealed that people, industry, and critical infrastructure around Lake Mälaren were in fact more vulnerable than previously thought.

Managing the Process


The WSP team consisted of 30 people of all disciplines and areas of expertise. To collect the data for the database, we identified potentially vulnerable “objects”, then went out and talked to local authorities, people in the relevant companies, and technical officers in the 24 municipalities. Normally such studies are entirely desk-based, so it was a treat for the researchers to make site visits, and these enabled the team’s hypotheses to be tested and confirmed.

More than 400 “objects” (e.g., installations, supply mains, service provision locations) had critical information, but the results needed to be aggregated and a regional analysis performed to report on their status and the recommended actions to be taken. WSP had all the expertise in-house, and this made it easier to coordinate all the work to be done.

We used flooding overlays to identify each object and contacted the operational manager for it, by phone and then in a site visit, to identify important features such as threshold levels, valves, tubing, etc. From this we were able to identify the critical water level, measuring the levels on location with accurate GPS and other such tools. We digitised coordinates to preserve confidentiality. Confidentiality was important and the city of Stockholm requested information security.

For purposes of information security, this article gives only broad, summarized findings and results.

Some Key Findings

  • The risk of flooding in Lake Mälaren is currently high because the inflow to the lake may be higher than the outflow capacity. Until an increase in bottleneck capacity, or prevention and preparedness-raising on a very large scale has been implemented, the risk remains high.
  • Water levels only 0.5 metres above the mean water level of Lake Mälaren begin to impact establishments carrying out important public activities. This can mean great risk or danger to life and health, social functionality, or society’s basic values.
  • Of a total of 236 inventoried socially important objects within the flood prone area, more than 180 items can have consequences that are serious, very serious, or even catastrophic for the objects and thus affect society. Twenty-two of these items provide service to very large parts of the population in various municipalities. The consequences would typically involve the non-delivery of electricity, drinking water, sewerage, or heating.
  • The impact of a rising water level is greatest in the sectors of electricity and local technical support.
  • Many socially important objects are vulnerable for reasons other than flooding. Several of them have critical dependencies on other objects.
  • During a flood event of up to a +1.4 metre rising water level, ground types of agriculture, forestry, and nature conservation areas are affected. At levels above +1.4 metres, settlements, buildings, roads, etc. will be increasingly affected. Large numbers of people will be affected at levels above +1.7 metres.
  • At a “1 in 100 year” event, about a +1.9 metre rising water level and a duration of three weeks, the direct costs of the socially important objects are estimated at approximately €64 million and about 230,000 people will be without service. At the design water level +3.1 metres, the direct costs are estimated at almost €1 billion and more than 600,000 people will be directly affected.
  • During a flood event with high water levels in Lake Mälaren, the problems are likely to be very large, even in the tributaries and within a larger region of the country. There will be competition for society’s collective resources with great challenges and limited capacity as a result.
  • There may be a reason for a number of items that deliver important public services to have steps taken to reduce their vulnerability until an increased outlet capacity from Lake Mälaren is provided.
  • By increasing the outlet capacity and by adopting a different water regulation regime, the high water levels could not occur. Achieving the same protection of the community through prevention and preparedness-raising is not a realistic option.
  • For areas freed up by an increased outlet capacity from Lake Mälaren, municipalities should take into account the long-term risks of flooding in the planning process.


The Swedish Meteorological and Hydrological Institute had done modelling of climate scenarios for Lake Mälaren, which provided us with the lake levels for different annual probabilities. However, WSP did not have to consider return periods or probabilities, but only reviewed the consequences of each 10 centimeter rise in water level. The results did not follow a steady slope. Rather the graph was stepped, with a 10 centimeter rise at one level causing a domino effect - such as electricity loss in one area producing consequences elsewhere. For example, the flooding of an IT server room would prevent a municipality from delivering services; or a pump losing power would cause flooding, electricity outages, or drinking water pollution elsewhere. The next few 10 centimeter steps would typically have no impact.

It was originally thought that the roads would be flooded, and people would be isolated and run out of fuel and food, also that hospitals would be affected by flooding.

What we discovered, contrary to what might have been found from a desk-top study alone, was that the main problem happens before the water reaches the buildings, because the link to electricity and other utilities becomes cut off before then. The critical vulnerable point may well be the electricity cable entering the building 0.5 metres underground, which could get short-circuited well in advance of the buildings flooding.

Most utilities had control of their own building and functions, but assumed they would have electricity to power pumps, IT servers, internet, etc. (although they had generators for some equipment). So they relied on the electricity supply company; and the critical vulnerable local electricity location could be serving, and therefore enabling, several services or installations.

The study results showed the following consequences:

  • Five municipalities would not have drinking water because the pumps were out of action, and the lake itself – the source of drinking water – could become polluted;
  • Critical metro tunnels would flood even with only a small rise in water level;
  • Fuel for electricity/district heating, transported by water, would not be available because the locks were needed for flood control;
  • Key municipal data centres in building basements would be flooded, along with their backup systems; and
  • There would be no sewage treatment in some areas, because receiving waters would be too high.

The direct cost was estimated at €1 billion but the indirect cost would be much higher, and drinking water quality could be compromised. The largest impact would be on the two million people who get their drinking water from the lake; electricity outages from flooding could result in sewage leaking back into the lake’s drinking water supply.


Besides giving the government a comprehensive analysis and assessment of the impact of a flood in Lake Mälaren, the focus of the study has also been to encourage the continued development and management of flood issues. Sluices on Lake Mälaren are now being remodelled, and contingency actions are being taken on some of the most sensitive areas. Negotiations are in progress to design a new lock on the lake in Stockholm, with three times the current capacity. The benefits to society will likely far exceed the costs, but both have yet to be established.

Knowing the risks of flooding today means that the city authorities can take action, and municipalities, utilities, and other national authorities have subsequently asked WSP to help find solutions to all the objects identified as being at risk.

Our research brings two insights for other cities. Firstly, that climate change can impact major infrastructure and not necessarily in the way originally envisaged. Secondly, that the full impact only really becomes clear when you get out the office, involve the engineers, and get into the details of municipal infrastructure design.

More on this subject