The pumping-weir complex at IJmuiden plays an important role in the drainage of excess water in the Western Netherlands. Multiple pumps need replacing as 4 out of 6 pumps near their end-of-life term. The optimal replacement strategy critically hinges on the future required pumping capacity. Yet, currently available models are not suited to assess the effect of sea level rise or extremer precipitation events as they ignore certain complexities of the water system. Preliminary data analysis in this paper showed the sensitivities of the system. The required pumping capacity is sensitive to the ability of free discharging during extreme water events. Yet, it is less susceptible to extremer precipitation events. Further research will aim at including more of the water system’s complexity in the model. Due to the node-like structure and high availability of data, a neural network modelling approach will probably be suitable.

The efficacy of risk models and risk analyses critically hinges on sufficient model evaluation. Nevertheless, the usefulness for the intended purpose is rarely systematically assessed. Poor or even lacking model evaluation of the applied risk models and analyses also troubles the asset management of storm surge barriers in the Netherlands. In practice, obvious flaws, missing failure modes and use that deviates from the original purpose regularly lead to unpleasant surprises, unnecessary costs and avoidable risks.

Here, we introduce new guidelines for the quality control during the development, testing, maintenance and usage of risk analyses of critical hydraulic structures.

First responses among stakeholders are rather positive since the guidelines help modelers and analysts to better understand critics and independent reviewers to structure their comments. However, the efficacy of the guidelines itself also need rigorous evaluation in the coming years. This may prove challenging as the application of the guidelines may also reveal that the organizations that operation the storm surge barriers are currently insufficiently equipped for the rigorous quality control of risk models and risk analyses.