Preface—Special Issue: Facets of Uncertainty

Uncertainty is the only certainty we have, taught Pliny the Elder 20 centuries ago. How far have we come today, in that modern science would confirm or contradict this classical conundrum? This Special Issue of Hydrological Sciences Journal is dedicated particularly to contributions that could shed some light on any of the many facets of uncertainty.

The title of the Special Issue has been inspired by the name of a triple conference held on Kos Island (Greece) in October of 2013, which also represents its main source of contributions. The triple conference, convened by Demetris Koutsoyiannis, hosted the Statistical Hydrology (STAHY) Workshop of the International Association of Hydrological Sciences (IAHS), the Leonardo Conference of the European Geosciences Union (EGU), and the third edition of the Hydrofractals Conference, all of which focused, this time, on the most diverse aspects of uncertainty in natural processes.

The different facets involving the role of uncertainty in modelling natural processes include the heuristics of its presence (regarded by many as an opponent of science, eventually to-be-eliminated, but perhaps being intrinsic to nature, as stated by the standard view of quantum mechanics or implied by statistical thermophysics and the theory of dynamical systems), its pragmatic treatment (estimation and ways to include it into mathematical models), as well as its propagation in model predictions. The paper by Dimitriadis et al. (2016a) compares the predictability of a die’s motion with that of hydrometeorological processes, and challenges the false dichotomy between determinism and randomness, showing how deterministic approaches can be combined with uncertainty estimation.

A particular section (Sutcliffe et al. 2016, O’Connell et al. 2016) celebrates the legacy of Harold Edwin Hurst, the empirical discoverer of long-range dependence in natural processes. It is not surprising that this discovery occurred precisely in the field of hydrology, where it managed to grant a parsimonious description to an otherwise complex process. The fertility of the seed that Hurst planted is shown by the variety of areas in which similar scaling behaviours with statistically estimated exponents are being employed, from structure functions in turbulence to climatic processes, as well as by the great minds who have concerned themselves with its study, Kolmogorov, Wiener, Mandelbrot and many others.

Historically, it has been precisely the existence of such, initially regarded as “anomalous”, exponents characterizing the behaviour of many natural processes that prompted the development of the notions of fractal sets and fractal measures. These eventually grew into the more general theory of monofractal and multifractal scaling, whose contribution to the development of hydroscience has largely yet to be written. This Special Issue adds to the body of knowledge on this topic with three papers. The first contribution (Tsonis, 2016) is a concept paper introducing randomness as a property of nature, while the stochastic representation with uncertainty of the self-similarity structures of turbulence and hydrometeorological processes, e.g. fractal dimension and Hurst coefficient, is covered by Dimitriadis et al. (2016b). Finally, Gires et al. (2016) discuss the use of the theory of multifractals as a tool to improve knowledge and understanding of hydrometeorological processes (specifically, rainfall as a space−time process).

The Leonardo Conference, in turn, offered an opportunity for the development of applications and new methodology, as well as revisiting the underlying philosophical basis for uncertainty analysis. The contribution by Beven (2016), inspired by his Leonardo Lecture, highlights the rich history and current challenges of uncertainty analysis, having to deal with various types of uncertainty that he defines. Nearing et al. (2016) put the debate about hydrological uncertainty analysis into the context of the historical literature on epistemology and the contemporary axiomatically derived interpretation of probability theory as extended logic. Tyralla and Schumann (2016) present a new approach for dealing with model structural uncertainty. Mylopoulos and Sidiropoulos (2016) consider the management of an aquifer under uncertain hydrogeology of the subsurface and propose a stochastic optimization approach, while Mukolwe et al. (2016) deal with uncertainties on the surface, studying the influence of various elevation datasets on accuracy of flooding models. Kabeya et al. (2016) quantify the effects of forest harvesting versus climate on streamflow variability and trend, while Koga et al. (2016) present an urban-scale hydrological application accounting for the uncertainty in land-use delineation. Maftei et al. (2016) investigate the presence of long-range dependence in flow time series from a sparsely studied region in Romania. Finally, Di Baldassarre et al. (2016) consider the seventh facet of uncertainty, unknown unknowns, and the surprises that can arise in coupled socio-hydrological systems.

As work on this Special Issue comes to a close, we cannot avoid a look back to the island of Kos, venue of the “Facets of Uncertainty” triple conference and homeland of Hippocrates, the pioneer of Western Medicine and a paragon of mitigating human suffering, only to find that the paradise we came to know during the conference, has since been overshadowed by the uncountable tragedies of those who, having been uprooted from their homeland by war, have come to seek shelter on the island. As researchers in the realm of uncertainty, it is precisely the latter that allows us to be hopeful for an end to the humanitarian crisis, as it is uncertainty itself that allows for unexpected change (Koutsoyiannis 2010, 2013), and enables the existence of such marvellous human traits as optimism and free will, whose inner workings we still have to understand in depth.