Waqas et al. (2026) CMIP6-Driven Groundwater-Level Projections and Climate Risk Mapping for South Korea Using a Hybrid Deep Learning Framework

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Water
• Year: 2026
• Date: 2026-09-22
• Authors: Muhammad Tayyab Waqas, Sang Kyum Kim
• DOI: 10.3390/w18192358

Research Groups

• Korea Advanced Institute of Science and Technology (KAIST)
• National Groundwater Monitoring Network

Short Summary

This study developed a groundwater-level projection framework based on CMIP6 simulations to assess the impact of climate change on groundwater resources in South Korea.

Objective

• Investigate the relationship between climate forcing and groundwater variability under nonstationary conditions.

Study Configuration

• Spatial Scale: National scale, focusing on South Korea.
• Temporal Scale: 2009–2100, with a focus on the period 2081–2100.

Methodology and Data

• Models used: Hybrid attention-based convolutional neural network-long short-term memory (HACL) model.
• Data sources: National Groundwater Monitoring Network data (199 stations), CMIP6 precipitation and soil-moisture forcings.

Main Results

• The filtered ensemble projected national-average groundwater-level increases of 16.60 m (SSP245), 15.92 m (SSP370), and 17.85 m (SSP585) by 2081–2100.
• The largest increases were observed in the western alluvial lowlands, substantially exceeding historical rates.

Contributions

• This study provides a replicable approach for projecting groundwater-level changes under climate change scenarios, particularly relevant for monsoon-affected regions like South Korea.
• The Climate Groundwater Risk Index (CGRI) was developed to integrate projected changes, ensemble spread, observed variability, and vulnerability.

Funding

• Not specified in the paper.

Citation

@article{Waqas2026CMIP6Driven,
  author = {Waqas, Muhammad Tayyab and Kim, Sang Kyum},
  title = {CMIP6-Driven Groundwater-Level Projections and Climate Risk Mapping for South Korea Using a Hybrid Deep Learning Framework},
  journal = {Water},
  year = {2026},
  doi = {10.3390/w18192358},
  url = {https://doi.org/10.3390/w18192358}
}