Unknown knowns: a missing piece of the uncertainty quadrant in hydrology

Uncertainties — of different types, sources, and their characterisation methods — have been studied for years in environmental sciences and beyond, until Donald Rumsfeld gave it a new kick and audience. The terms known knowns, known unknowns, and unknown unknowns have become more prevalent to map out different types of uncertainties associated with understanding, modelling, and managing complex systems; even though these terms had been used in scientific literature before Rumsfeld’s remarks. The last piece of this quadrant that is less discussed is unknown knowns, i.e. the knowns that are treated as unknowns, that we have overlooked, forgotten, or ignored. In this work, we would like to emphasis on unknown knowns by discussing a few examples of them in scientific inquiries in general, and hydrological sciences in particular; from socio-economic factors to unconscious biases in formulating and analysing complex human-water systems. For example, in many cases the modellers’ choice of hydrological models is biased toward legacy (of existing models within the repertoire of the modellers) than based on adequacy. While human-water interactions are inherent in catchment processes, most common catchment models do not have a module to account for human activities and policy feedback. Neglecting such basic knowns — unknown knowns — can bias our analyses, decision making, and the narratives established to explain or react to a phenomenon such as drought. For instance, the “Zero Day” narrative of the Cape Town water crisis (around 2017). While the hydro-climatic multi-year drought was a contributing factor to the water shortage, the regional politics and water management also played a major role. Despite dramatic drop in water supply and hence drastic restrictions in water consumption across multiple sectors, water consumption by industrial users with political connections (such as local licensed Coca-Cola manufacturers) remained disproportionally higher. In effect, the zero day was only zero for some and not for everybody. To advance hydrological sciences, particularly to integrate socioeconomic and policy feedback interactions into the mainstream hydrological modelling, hydrologists can learn from such advances in neighbouring disciplines (e.g. sustainability science and system dynamics) and account for less conventional types/sources of uncertainties, especially unknown knowns.