In the AllRisk research programme, researcher Wouter Jan Klerk delved into how we improve our management practices on the basis of flood probabilities when reinforcing, managing and maintaining dikes. He focused on three main topics: optimising dike reinforcement projects, investing in reducing uncertainties about dike strength, and risk-based management and maintenance. His doctorate award ceremony is March 28 at Delft University of Technology.

Optimising dike reinforcement projects

In his study, Wouter Jan developed a method to optimise the reinforcement of dike sections (with a length of approximately 20 kilometres) based on the effect of different reinforcement solutions on the flood probability. The method has been applied in practice in the Streefkerk-Ameide-Fort Everdingen (SAFE) dike reinforcement project to determine which parts of this dike section need upgrading in the short term. The comparison with the usual methods demonstrated that using this optimisation method not only leads to a much more economical dike upgrade (reducing costs by 30-40%) but that it also prevents over-extensive dike upgrades and therefore local nuisance.

Reducing uncertainties

Many Dutch dikes have failed assessments because of susceptibility to piping and instability. The fact that we have little certainty about how a specific section of dike will behave under extreme conditions often plays a role here. In his study, Wouter Jan developed a method to substantiate the benefits of dike monitoring and test loads, for example on the basis of infiltration tests, in order to reduce these uncertainties. Not only did it emerge that the benefits are quite substantial, this approach also resulted in a number of practical recommendations that are useful for the efficient use of monitoring and test loads in practice.

Risk-based management and maintenance

The third topic he examined was risk-based management and maintenance. The Dutch dikes are subjected to annual visual inspections by the water authorities, but it was not known how accurate these inspections were. In collaboration with the Rivierenland water authority, this topic was investigated in practice by having a number of dike sections inspected several times by different inspectors. It emerged that inspections provide valuable information but that visual inspections are certainly not perfect, and that damage and defects are sometimes not detected.

So, it was decided to look at the impact on failure probabilities for dike sections: it emerged that the margin associated with visual inspections is so large that it has to be taken into account when determining failure probabilities. At the same time, extra inspections, with or without potentially more accurate inspection methods such as drones, were found to be very promising options for improving risk-based management and maintenance.


The optimisation approach for dike reinforcement operations developed in this study is extremely suitable for programming and prioritising future dike upgrades. It has the potential to make the programming of dike reinforcement much more efficient. Further development should include a closer look at the role of uncertainties and how these can be reduced.

An important area to come to light in this study was the connection between management and maintenance, and the probability of dike failure. The implementation of management and maintenance in practice is far removed from the flood probabilities. However, at the same time, it has proven to be very important.

Substantiating how inspections and maintenance relate to the probabilities gives us a picture of how day-to-day practice can be improved. Improving visual inspections or using drones for inspections, for example, could make an important contribution to complying with the flood probability requirements.

Share this page.