Hydroinformatics opening new horizons: Union of computational hydraulics and artificial intelligence

Abstract

The strengthening dominance of electronics over past several decades has resulted in an almost total dominance of digital representation of the hydro-environment. Recognizing these developments, Mike Abbott introduced the notion of electronic encapsulation of information and knowledge. The act of encapsulating information and knowledge changes the very nature of the information and knowledge involved. Suffice to say that electronic encapsulation must also change the way in which an engineer accesses and uses the available information and knowledge. There is a shift in paradigm away from a model-based approach to a more data-based approach. Adopting such foundational considerations as its core, hydroinformatics opened research to the latest IT developments in the fields of artificial intelligence (including machine learning, evolutionary algorithms and artificial neural networks), artificial life, cellular or finite-state automata and other, previously unrelated sciences and technologies. Through studying and exploiting elements of these, at first sight unrelated, sciences, hydroinformatics produced new and innovative solutions to hydraulic and hydrological problems, as represented by real-time control and diagnosis, real-time forecasting, calibration of numerical models, data analysis and parameter estimation. In particular, the new approaches are able to generate important components of physically based, modelling systems by inducing models or sub-models of individual physical processes based only upon measured data. These (sub)models may then replace whole systems of complex, non-linear, differential equations that would otherwise require great skills from the modeller to calibrate, and powerful computing devices to solve. This chapter captures evolution of data science and AI within the field of hydroinformatics and provides an outline of the present state-of-affairs together with some ideas for future directions. The chapter outlines tangible solutions that have been applied by the hydroinformatics community to address specific challenges in hydro-environmental systems.