Developing a Novel Clustering-Merging method based on Machine Learning Algorithms

The main objective of the current research is to enhance the detection, clustering, and estimation of Extreme Precipitation Events (EPEs) in arid and semi-arid regions. To achieve this goal, we propose a novel framework called Machine-learning-based Clustering-merging algorithms (ML-CMAs) for satellite-based precipitation products (SPPs). Daily precipitation measurements were utilized for training and evaluating EPEs estimation results. Additionally, Auxiliary Data (AD) such as the Digital Elevation Model (DEM), air temperature, and Precipitation-based Precipitable Water Vapor (PPWV) were employed concurrently. In this research, we conduct a statistical analysis and evaluation of the results associated with each of the four SPPs concerning their estimation of precipitation during the occurrence of EPEs. While examining each of these products individually, we place special emphasis on the outcomes obtained from the proposed methods based on ML-CMAs. Furthermore, we aim to address the critical question of whether these methods will demonstrate the capability to estimate precipitation amounts more accurately than SPPs and enhance statistical indices.

Faculty Supervisor:

Saeid Homayouni

Student:

Partner:

K. N. Toosi University of Technology

Discipline:

Earth science

Sector:

Water; Environmental Science and Technology; Artificial Intelligence

University:

Université du Québec : Institut national de la recherche scientifique

Program: