Application of Artificial Intelligence and Machine Learning Algorithms in Protection Against Hydrological Disasters

Abstract

In the last decades, the number of disasters has increased significantly, both natural and man-made, at the national and international level. As a result, the importance of forecasting systems has greatly increased. Information technology and computer science, including artificial intelligence and machine learning, have undergone immense development in recent years, hence making it possible to create predictive models with very high accuracy and great generalization capability, even having relatively few resources available. Due to climate change, the weather is becoming increasingly unpredictable, with extreme weather becoming more frequent. This is a serious challenge for countries, including Hungary. Unfortunately, hydrological disasters have become quite regular in our country. For example, we can mention the Danube and Tisza floods of 2006, the Danube flood of 2013, the major flood hazard situation on the Danube of 2024, the 2025 flood on the Kapos River, the 2013 Zemplén flash flood, the 2020 South Zala flash flood, the very severe droughts of 2012 and 2022, the moderately severe drought of 2021, the dry weather conditions of 2025, and the snow disaster of 2013. In 2024, forecasting systems successfully detected the danger in time, and thanks to extraordinary cooperation and the tireless work of experts, the country managed to defend effectively against the flood. Consequently, taking action in time is crucial, along with the development of new, even more accurate forecasting and warning systems, as well as defense mechanisms and population preparedness. In this article, I present the algorithms used in domestic and international research, as well as the forecasting systems operating in Hungary.