This issue of Water Management includes four papers that cover a broad range of topics, from the analysis of weather extremes to three-dimensional hydrodynamic modelling. Such a variety of topics reveals the undeniable complexity that characterises the management of water on Earth. In this respect, identifying the correct scenarios for extreme rainfall events (Darch et al., 2016) and the deterministic or chaotic behaviour of hydrological processes (Kędra, 2016) is tightly related to the optimal design of reservoirs (Celeste, 2016) and, at a more detailed scale, to the interaction between hydrodynamics and sedimentation that can strongly affect the management of shallow reservoirs (Esmaeili et al., 2016).The first paper (Darch et al., 2016) analyses the use of UK climate projections (UKCP09) weather generation (WG) to infer changes in the magnitude of short-duration extreme rainfall events. The problem is undoubtedly relevant as the UK experienced a two-fold increase in the magnitude of extreme rainfall events in the last 50 years, and UKCP09 projections show a further increase in rainfall and extreme events in winter. With the goal of defining meaningful weather series and to analyse their possible variations, WG can be used as a form of statistical downscaling. Two methodologies are proposed for this kind of analysis: (1) the definition of design storms for events of varying duration and frequency; and (2) the time series method to select a manageable number of WG series to be used as an input for data-and time-intensive models, for instance for flood risk modelling. The proposed methodology contributes to define reliable scenarios with important practical applications, and can be particularly useful for the design of urban drainage and sewer systems.In the second paper (Kędra, 2016), the non-linearity in the response of hydrological systems is discussed considering the daily discharge of the Krzyworzeka Stream (Poland) as a case study. The author shows how it is possible to assess the nature (deterministic versus stochastic) of the process by means of a combination of different methods. Two standard linear (autocorrelation function, ACF, and Fourier analysis) and two non-linear ('averaged false neighbours', AFN, and '0-1 test') methods were applied to the data relative to the Krzyworzeka Stream, and to three other synthetic test cases (periodic signal with noise, white noise and red noise). The analysis of the results of the four methods allows for clearly identifying the nature of the four time series, and in particular to show that the measured discharge of the Krzyworzeka Stream was deterministic, but with chaotic dynamics. Such a conceptual framework may be used to quantitatively measure the relative degree of deterministic behaviour and of random features in hydrological and other geophysical processes, thus supporting the choice of the most adequate predictive model.