Comparison of interpolation methods in order to improve watertable prediction using deep learning methods

Articles in Press

Document Type : Research Paper

Authors

1 Faculty of Agriculture - Tabriz University

2 2- Assistant Professor, Department of Water Engineering, Faculty of Agriculture, University of Tabriz, Iran

3 Professor, Department of Water Engineering, Faculty of Agriculture, University of Tabriz, Iran

10.22059/jwim.2024.372424.1145

Abstract

Groundwater resources are an important factor in managing and maintaining water that is used for drinking water, irrigation, and other purposes. Forecasting and controlling the fluctuations of the underground water level is very important, especially in areas where water shortage is severe. However, these areas have significant water needs, and on the other hand, they need cost-effective strategies to use underground water resources, so that approaches and decisions can be taken to prevent excessive loss in them. In this study, the Azarshahr Plain, which has recently faced a significant drop in the underground water level, was examined monthly from 2017 to 2018. Also, in order to complete the data that was not measured for any reason, kriging interpolation methods and M5P algorithm were used. By analyzing each method, the M5P method with the minimum root mean square error of 1.83 meters and correlation coefficient of 0.975 was the best. It had the function. On the other hand, the deep learning (DL) method was used to predict the underground water level. This method is acceptable with an error of 1.408 meters and an accuracy of 88%, and it can be used in future research.

Keywords

Main Subjects

Water and Irrigation Management

Articles in Press, Accepted Manuscript
Available Online from 10 June 2024

Files

Share

How to cite

Statistics