NEWS & PRESS

The impressive paradigm shift represented by the digital transformation, which encompasses more and more areas in technology and science, economy and society, could not miss its mission as a fundamental tool to contribute to the improvement of prediction techniques for extreme events linked to flash floods in river channels.

In recent years, the research effort is intensifying in line with the possible applications of certain Artificial Intelligence (AI) techniques in this field, and in particular in hydrological and hydraulic modeling of floods, and thus complement conventional numerical calculation methods (finite elements, finite volumes, in explicit or implicit schemes) and physical models.

As is well known, floods are one of the most frequent and damaging types of natural disasters, especially in developing countries and communities. To cope with the global threat of climate change, AI-based technologies are called to stand out within the set of capabilities available to quantify with greater assurance the hazards, impacts and risks of floods, increasingly influenced by climate change.

The incorporation of this type of techniques in the traditional methods used in short- and very short-term flood forecasting (deterministic numerical models, in which the shallow water flow equation is discretized and solved in two spatial dimensions, or probabilistic models, which estimate the probability density function of the plausible states) will undoubtedly help to overcome the limitations of these methods, such as their intensive demand in computational resources and the high computational times, often too long to achieve their purpose (bear in mind that it is a matter of forecasting in the very short term – time scales of a few hours – the effects downstream of the flood source in order to make decisions in real time, practically).

This brief post is the first part of a series of publications that will appear soon in this medium, with the initial purpose of disseminating information on the state of the art and the current and foreseeable possibilities that AI techniques offer to the agencies and authorities responsible for flood management, and with the ultimate aim of providing a social benefit.

In successive entries, we plan to comment on the results of the latest and most relevant research published worldwide by scientists, technicians and engineers, specialists in both the field of AI (Machine Learning, Data Science, etc.) and in the disciplines directly related to the nature of the events to be studied (Meteorology, Hydrology, Hydraulics, Land Use Planning, etc.), and would refer to some of the applications already available on-line (on mobile devices, among others, and generally freely accessible), which try to transfer the theoretical results of such research to easy-to-use and accessible tools. ), and would refer to some of the applications already available on-line (on mobile devices, among others) and generally of free access, which try to transfer the theoretical results of these investigations to tools of easy use and accessibility, which could play a key role in the anticipation of warnings.