Prediction in quantitative hydrological modelling: a small step away from uncertainty

Water use planning is vital, making hydrological modelling crucial for the development of human society. This role of hydrological modelling is rendered even more prominent given current population growth and global warming resulting in more pressure on water resources. Nonetheless, most quantitative hydrological predictions remain fundamentally uncertain leading to uncertain water availability assessment. This is so, partly, because streamflow continues to be used as the main dependent variable, calibration and validation data length and content are arbitrarily selected, and water balance components taken individually are not quantified with the best prediction models. This paper discusses three ideas for stepping further away from uncertainty in quantitative hydrological predictions. Less uncertain hydrological predictions can be achieved through (i) blue water and green water-based hydrological models validation, (ii) blue water and green water-based multi-model evaluation of water resources, and (iii) a conditioned selection of calibration and validation data.