报告题目：“Artificial Intelligence in Earth Sciences,Dos and Don’ts”

报告时间：2025年1月13日15:30-17:30

报告地点：地理资源所A901

Vahid Nourani,Ph.D.

Professor of Civil & Environmental Eng.

Director for Center of Excellence in Hydroinformatics and Institute of Environment

Faculty of Civil Eng.,Univ. of Tabriz,Iran

Adjunct Prof.,Near East Univ. (Turkey) and Charles Darwin Univ. (Australia)

BIO:

As an expert in the field of Hydro-environmental Informatics, Prof. Nourani has been affiliated to Faculty of Civil Engineering, University of Tabriz (Iran), University of Minnesota (USA), Near East University (Turkey) and Charles Darwin University (Australia). In his career, 21 Ph.D. and 181 M.S. students have been graduated under his technical supervision as well as 7 Postdocs. His research interests include hydro-geological modeling, Artificial Intelligence applications, Hydroinformatics and computational hydraulics. His researches outcomes include over 300 Journal articles, 3 books, 21 book chapters, more than 200 conference presentations and 8 international projects. Currently he is Editor-in-Chief for Journal of Civil & Env. Eng., Associate Editor for Journal of Hydrology, Journal of Hydroinformatics and AQUA. He is director of Excellence Center in Hydroinformatics and director for the Institute of Environment at Tabriz University. In 2023, Prof. Nourani was ranked as 9035th among world's top 2% scientists reported by the Stanford with the global position of 29th in Environmental Engineering, 1st in Iran. According to ScholarGPS, Prof. Nourani is ranked in top 0.05% scientists with 5-year global rank of 2104 and the global positions of 2nd in Hydrology and 16th in AI. He is a fellow of CAS- PIFI (China), JSPS (Japan) andTUBITAK(Turkey).

Abstract:

The rapid development of AI technology has unlocked new opportunities in Earth sciences, offering innovative ways to model complex processes and phenomena. However, like all mathematical tools, AI models have limitations, particularly due to their inherent "black-box" nature. These limitations necessitate cautious and well-informed applications to prevent potential shortcomings. To address these challenges, it is essential for users to possess a deep understanding of the physical principles underlying the processes. Reliable interpretation of results is crucial for ensuring that AI-driven insights are scientifically sound and actionable. This lecture aims to i) present critical perspectives on the application of AI in Earth sciences, highlighting potential pitfalls and limitations, ii) Explain, in straightforward language, how AI models are trained and deployed for modeling complex processes—especially in cases where physically based models are infeasible or insufficient, iii) Provide practical advice and recommendations for achieving reliable, high-performance outcomes from AI-driven models. By combining AI tools with domain expertise, scientists can leverage their potential while maintaining the integrity and reliability of Earth science research.