Evaluation of Groundwater Storage Anomalies in Major River Basins of India using GRACE Satellite and In-Situ Groundwater Level Data
Loading...
![Thumbnail Image](assets/images/Item.jpg)
Date
2024-06
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
G. B. Pant University of Agriculture & Technology, Pantnagar-263145
Abstract
Groundwater, a significant source of fresh water and often referred to as an
invisible resource, plays a crucial role in sustaining various aspects of human life.
However, excessive exploitation of groundwater can lead to adverse consequences such
as droughts, water scarcity, and hindered social and economic development. Therefore, it
is essential to analyze groundwater levels to effectively manage this vital resource. This
study focuses on understanding groundwater depletion by investigating 14 different
locations across five basins (Ganga, Krishna, Godavari, Mahanadi, and Narmada),
strategically selected to be 300 km apart. In the Ganga Basin, the locations of Badarwas,
Daryaganj, Prayagraj and Bahera were chosen, while Adlur and Amarapuram were
selected in the Krishna Basin. For the Godavari River, Jalgaon, Dhoki, and Akolebazar
selected as the study locations. Bodeli, Handia, and Banjari represented the Narmada
River, and BLD-022-OW and Abhimanpur were selected in the Mahanadi Basin. A
comprehensive dataset spanning 17 years from 2002 to 2019 was utilized, incorporating
GRACE (Gravity Recovery and Climate Experiment) and GLDAS (Global Land Data
Assimilation System) data to calculate groundwater storage anomalies (GWSA). The
Mann-Kendall and Sen's slope methods were utilized to conduct trend analysis. Weaker
correlations were observed between GWSAWELL and GWSAGRACE in the Ganga, Krishna,
Narmada, and Godavari basins. To improve correlation with in-situ measurements
(GWSAWELL), machine learning techniques, namely Artificial Neural Networks (ANN),
Random Forest (RF), and Support Vector Machines (SVM), were employed. Spatial and
temporal trend analysis revealed depletion patterns during specific seasons. Among the
machine learning models, the RF model consistently demonstrated strong correlation
levels (<0.90) with GWSAWELL. This research enhances groundwater estimation and
facilitates effective water resource planning in densely populated areas through the
application of machine learning techniques and analysis of long-term datasets.
Description
Keywords
Citation
Theses of M. Tech