Irfan et al. (2026) Forecasting of global water usage in agriculture and total global consumption by using the Bi-GRU model

Identification

• Journal: Scientific Reports
• Year: 2026
• Date: 2026-03-26
• Authors: Muhammad Irfan, Javed Rashid, Javeria Bibi, Shiza Amir, Kamal M. Othman, AbdulGuddoos S. A. Gaid
• DOI: 10.1038/s41598-026-44885-8

Research Groups

• Department of Mathematics, University of Okara, Okara, Pakistan
• Information Technology Services, University of Okara, Okara, Pakistan
• Department of Electrical Engineering, College of Engineering and Architecture, Umm Al-Qura University, Makkah, Saudi Arabia
• Department of Communication & Computer Engineering, Faculty of Engineering & Information Technology, Taiz University, Taiz, Yemen

Short Summary

This study develops and applies a Bidirectional Gated Recurrent Unit (Bi-GRU) model to forecast global Total Water Consumption (TWC) and Agricultural Water Use (AWU), demonstrating its superior accuracy compared to other deep learning models for effective water resource management.

Objective

• To accurately predict global Total Water Consumption (TWC) and Agricultural Water Use (AWU) to support the development of efficient water-management strategies.

Study Configuration

• Spatial Scale: Global
• Temporal Scale: Data from 2000 to 2024 for model training and evaluation, with the objective of long-term forecasting.

Methodology and Data

• Models used: Bidirectional Gated Recurrent Unit (Bi-GRU). Its performance was compared against LSTM, Deep-AR, Mega-CRN, and TFT models.
• Data sources: Global Water Consumption Dataset (2000–2024), available from Kaggle.

Main Results

• The Bi-GRU model demonstrated superior performance in forecasting global TWC and AWU compared to LSTM, Deep-AR, Mega-CRN, and TFT models.
• Key performance metrics for the Bi-GRU model:
  • Mean Absolute Percentage Error (MAPE): 0.2257
  • Mean Squared Error (MSE): 0.0049
  • Mean Absolute Error (MAE): 0.0587
• The higher precision achieved by the Bi-GRU model facilitates a more accurate evaluation of global water consumption, particularly in the agricultural and industrial sectors.

Contributions

• Introduction and validation of the Bi-GRU model for forecasting global Total Water Consumption and Agricultural Water Use, leveraging its bidirectional temporal learning capability for improved modeling of long-term and multivariate dependencies.
• Comprehensive comparison of Bi-GRU against several state-of-the-art deep learning models (LSTM, Deep-AR, Mega-CRN, TFT), establishing its superior accuracy for this specific forecasting task.
• Provides a robust tool for more precise evaluation of global water consumption, thereby enabling the development of more efficient and informed global water-management schemes.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Citation

@article{Irfan2026Forecasting,
  author = {Irfan, Muhammad and Rashid, Javed and Bibi, Javeria and Amir, Shiza and Othman, Kamal M. and Gaid, AbdulGuddoos S. A.},
  title = {Forecasting of global water usage in agriculture and total global consumption by using the Bi-GRU model},
  journal = {Scientific Reports},
  year = {2026},
  doi = {10.1038/s41598-026-44885-8},
  url = {https://doi.org/10.1038/s41598-026-44885-8}
}