Identifying and formulating uncertainties in Urban Water Management

Dealing with specific problems, the environmental engineer as well as any other technical professional has to analyze or predict a certain real system behavior. With respect to his objective he chooses to model this behavior with a suitable level of abstraction (model structure), using the corresponding model parameters and input variables (data). All three factors are uncertain. Taking this fact into account is either done implicitly (safety factors, values from literature, experience, intuition etc.) or explicitly, meaning that every uncertain input variable or parameter is not regarded as a fixed value but rather represented by a probability distribution of its value. In the latter case all uncertainties are directly considered in the calculation of the model output. Thus, the outcome will not be of any fixed value either, apparently representing a very reliable result. However, in most cases the incorporated uncertainties are insufficiently evaluated – a fact that may question the whole approach of taking uncertainties explicitly into account.

This Ph.D. project focuses on model inputs – variables and parameters. A strategy will be developed, which quantifies relevant uncertainties when applying models from the area of urban water management. The following questions shall be answered with respect to each analyzed model:

What sources of uncertainty have an impact on the model outcome?
What is the context behind these uncertainties?
Are there any correlations between input variables or model parameters?
How can the uncertainties and their interrelations be described quantitatively?
How are data, expert knowledge and experience implemented properly in this description?