K et al. (2026) Forecasting land-use and land-cover change for groundwater sustainability in the Muvattupuzha basin using CA-Markov (2033–2050)

Identification

• Journal: Scientific Reports
• Year: 2026
• Date: 2026-02-13
• Authors: Alagulakshmi K, Sneha Gautam, G. Prince Arulraj, Suneel Kumar Joshi, Chang-Hoi Ho
• DOI: 10.1038/s41598-026-38961-2

Research Groups

• Division of Civil Engineering, Karunya Institute of Technology and Sciences, Coimbatore, India.
• Water Institute, A Centre of Excellence, Karunya Institute of Technology and Sciences, Coimbatore, India.
• Department of Climate and Energy Systems Engineering, Ewha Womans University, Seoul, Republic of Korea.
• Geo Climate Risk Solutions Pvt. Ltd, Visakhapatnam, India.

Short Summary

This study integrates satellite-based land-use analysis with machine learning to demonstrate that built-up areas in the Muvattupuzha basin increased from 12.3% to 44.4% between 2003 and 2023. Using CA-Markov modeling, the research forecasts continued urbanization through 2050 and identifies magnesium, calcium, and alkalinity as the primary drivers of groundwater nitrate contamination.

Objective

• To evaluate historical land-use and land-cover (LULC) changes (2003–2023) and forecast future dynamics (2033–2050) in the Muvattupuzha basin.
• To investigate the relationship between LULC shifts, hydrogeochemical processes, and groundwater quality degradation, specifically focusing on nitrate variability.

Study Configuration

• Spatial Scale: Muvattupuzha river basin, Kerala, India (Total area approximately 2673.67 km²).
• Temporal Scale: Historical analysis from 2003 to 2023; future forecasting for the period 2033–2050.

Methodology and Data

• Models used: CA-Markov (Cellular Automata-Markov) for LULC forecasting; Random Forest (RF), Support Vector Regression (SVR), and XGBoost for nitrate variability prediction; SHapley Additive exPlanations (SHAP) for machine learning model interpretation.
• Data sources: Multi-temporal Landsat satellite imagery (Landsat 5 TM and Landsat 8 OLI/TIRS); hydrogeochemical field sampling data for physicochemical parameters; multivariate statistical tools (PCA and Cluster Analysis).

Main Results

• LULC Dynamics: Built-up land saw a massive expansion from 329.13 km² (12.3%) in 2003 to 1,187.11 km² (44.4%) in 2023, primarily at the expense of agricultural and forested land.
• Future Projections: CA-Markov models predict a continuous increase in built-up and forested areas through 2050, while water bodies and agricultural lands are expected to decline and eventually stabilize.
• Groundwater Quality: Multivariate statistics confirmed that both natural geogenic processes and anthropogenic activities (urbanization/agriculture) control hydrochemistry.
• Nitrate Drivers: Machine learning interpretation via SHAP identified Mg²⁺, Ca²⁺, and alkalinity as the most significant factors influencing nitrate distribution, highlighting the role of buffering and redox-controlled reactions in the aquifer.

Contributions

• Provides a novel integrated framework combining LULC forecasting with advanced machine learning (XGBoost/SHAP) to assess groundwater sustainability.
• Offers quantitative evidence of the rapid transition of a tropical river basin from agricultural/natural cover to an urban-dominated landscape.
• Identifies specific chemical precursors for nitrate contamination, assisting in targeted groundwater management for rapidly urbanizing regions.

Funding

• National Research Foundation of Korea (NRF) funded by the Korean Government (MSIT; RS-2025-00555756).
• Ministry of Education of the Republic of Korea (RS-2018-NR031078).

Citation

@article{K2026Forecasting,
  author = {K, Alagulakshmi and Gautam, Sneha and Arulraj, G. Prince and Joshi, Suneel Kumar and Ho, Chang-Hoi},
  title = {Forecasting land-use and land-cover change for groundwater sustainability in the Muvattupuzha basin using CA-Markov (2033–2050)},
  journal = {Scientific Reports},
  year = {2026},
  doi = {10.1038/s41598-026-38961-2},
  url = {https://doi.org/10.1038/s41598-026-38961-2}
}

Generated by BiblioAssistant using gemini-3-flash-preview (Google API)