A dike is designed for an extreme event which greatly exceed the situation under daily circumstances. The expected loading during such an event is, however, difficult to estimate. Inherent uncertainties in nature and epistemic uncertainties in models and statistical data impede a clear verdict about the exact loading on a dike. Consequently, incorporating uncertainties in a dike design can have a great impact. One of the aims, following the Delta Decisions of 2015, is to improve the way uncertainties are incorporated in future dike assessment and design models. Conventionally, models become more computationally intensive with the incorporation of each additional uncertainties. Therefore, it is useful to only account for most influential uncertainties. Until new models are established, the Ontwerpinstrumentarium (OI2014) is released for the transitional period. The guideline explains for which extreme event a dike needs to be designed. Looking in more detail to the OI2104, questions can be asked about the underlying assumptions. In this thesis, a modified method of the OI2014 is exhibited by looking into a dike’s location specific dominant failure mechanisms. In so, a better cost-optimal dike design can be realized. Also, a sensitivity is carried out in which currently recognized uncertainties are investigated using the PC-Ring model. In addition, a new statistical wind uncertainty model is built, wherein uncertainty is integrated a priori under the assumption that correlations between random variables do not play a significant role. The model’s implications are evaluated using the Hydra-Zoet model. From this research three main findings can be deduced. Firstly, the proposed methods of the OI2014 do not lead to a cost-optimal dike design. Secondly, addition of statistical wind uncertainty leads to a significant increase of hydraulic boundary conditions for lake systems. Lastly, the prior integration of any random variable has many computational benefits.