artículo científico
Smiling in the rain: Seven reasons to be positive about uncertainty in hydrological modelling
Fecha
2012-03-30Autor
Juston, John M.
Kauffeldt, Anna
Quesada Montano, Beatriz
Seibert, Jan
Beven, Keith J.
Westerberg, Ida K.
Institución
Resumen
Over 20 years ago, the notion of ‘positive uncertainty’ was introduced in the field of psychology (Gelatt, 1989). This was coincident to some developments in the hydrological sciences (e.g. Beven and Binley, 1992; Grayson et al., 1992) which laid the foundation for today’s continued efforts to explicitly acknowledge and quantify uncertainties inherent in our modelling efforts. ‘Positive uncertainty’ urges moving beyond deterministic frameworks of the past, but doing so not just by regrettably accepting that uncertainties are inevitable, but by positively thriving in the new perspectives that accompany this recognition (Gelatt, 1989). During a recent small workshop in Sweden – while the rain was pouring down – we discussed some of the positive developments in recent years that have resulted directly from explicit recognitions of uncertainty in hydrological modelling.With the earlier paper of Pappenberger and Beven (2006) in mind, we here summarise that workshop and elaborate seven reasons to be positive about uncertainty in hydrological modelling. How have the hydrological sciences benefitted in the last 20 years from the increased recognition of uncertainties, particularly in the modelling of time series data? In our view, uncertainty estimation is a means to address hydrologic research questions in an honest and robust way. There have also been (and will continue to be) new research questions and opportunities initiated by the recognition of what needs to be done to quantify and then constrain uncertainties. This notion fits intrinsically with the view of modelling as a learning process (Box, 1976), learning about the nature of the uncertainties involved and the possibility of reducing them by improving our models or our data.