Machine Learning for Peatland Ground Water Level Prediction IoT System

Abstract

Peatlands play a critical role in global carbon storage and climate regulation, making the accurate monitoring of their groundwater levels (GWL) essential for sustainable ecosystem management. This study presents an integrated Internet of Things (IoT) and Machine Learning (ML) system for real-time GWL prediction in peatland environments. The IoT component consists of low- power sensors deployed in the field to continuously collect environmental data such as temperature, humidity, rainfall, and water table depth. These data are transmitted to a cloud-based platform for storage and processing. Several ML models, including Random Forest, Support Vector Regression (SVR), and Long Short-Term Memory (LSTM) networks, are evaluated to predict GWL based on temporal and environmental patterns. Results show that LSTM provides superior predictive performance due to its ability to model sequential dependencies in time-series data. The proposed system enables proactive peatland management by forecasting GWL fluctuations, thereby supporting early intervention strategies to prevent peat degradation, reduce fire risk, and maintain ecological balance.

Introduction

Peatlands are vital ecosystems that act as major carbon sinks and help regulate climate, biodiversity, and water management. Ground Water Level (GWL) is a critical indicator of peatland health, influencing peat decomposition, carbon emissions, and fire risk. Traditional GWL monitoring is labor-intensive and often lacks real-time data, especially in remote areas.

This research proposes an IoT-based system enhanced with Machine Learning (ML) to predict GWL in peatlands in real time. IoT sensors collect environmental data (e.g., water level, soil moisture, temperature, rainfall) which is transmitted to the cloud. ML models, including Random Forest, Support Vector Regression (SVR), and Long Short-Term Memory (LSTM) networks, are trained on this data to forecast GWL fluctuations.

The literature review highlights the importance of peatlands, advances in IoT for environmental sensing, and the success of ML models in hydrological predictions. However, integration of IoT and ML specifically for peatland GWL prediction is limited, and this study addresses that gap by developing and deploying a real-time, end-to-end system.

The methodology covers IoT sensor deployment powered by solar energy, real-time data collection and cloud storage, data preprocessing (cleaning, normalization, feature engineering), ML model development and comparison, and system integration into a user-friendly dashboard for visualization and alerts.

Results demonstrate that the LSTM model outperforms others in prediction accuracy, effectively capturing temporal patterns in GWL. The IoT system reliably collects and transmits data with minimal downtime. The platform successfully predicted GWL rises during heavy rainfall, showcasing its practical value for sustainable peatland management and fire risk mitigation.

Conclusion

This study demonstrates the effectiveness of integrating Internet of Things (IoT) technology with Machine Learning (ML) techniques to predict Ground Water Level (GWL) in peatland ecosystems. The proposed system enables real-time environmental data collection through IoT sensors and leverages advanced ML models to generate accurate GWL forecasts. Among the evaluated models, Long Short-Term Memory (LSTM) networks showed superior performance in capturing temporal dependencies and producing reliable predictions By automating data collection and prediction, the system supports proactive peatland management, reduces the risk of peat fires, and helps maintain ecosystem health. Furthermore, the modular design allows for scalability and adaptability to different geographic areas and environmental conditions. This integrated approach offers a practical and innovative solution for sustainable peatland monitoring and can be expanded to other environmental applications in the future. Future work will focus on improving model robustness, integrating satellite data, and incorporating user feedback into the system to enhance decision-making and ecological outcomes.

References

[1] Joosten, H., Tapio-Biström, M. L., & Tol, S. (2012). Peatlands—Guidance for climate change mitigation through conservation, rehabilitation and sustainable use. Food and Agriculture Organization of the United Nations (FAO). [2] Wijaya, A. P., Sari, D. P., &Santoso, S. (2020). IoT- based peatland water table monitoring system with solar energy. IOP Conference Series: Earth and Environmental Science, 484(1), 012086. [3] Ki?i, Ö. (2013). River flow forecasting and estimation using different artificial neural network techniques. Hydrological Research, 44(1), 89–102. [4] Wang, Y., Yu, Y., Chen, Y., & Li, Y. (2019). Prediction of groundwater levels using an integrated model based on convolutional neural network and long short-term memory. Water, 11(9), 1891 [5] Kratzert, F., Klotz, D., Brenner, C., Schulz, K., & Herrnegger, M. (2018). Rainfall–runoff modelling using Long Short-Term Memory (LSTM) networks. Hydrology and Earth System Sciences, 22(11), 6005–6022. https://doi.org/10.5194/hess-22- 6005-2018. [6] Hooijer, A., Page, S., Canadell, J. G., Silvius, M., Kwadijk, J., Wösten, H., & Jauhiainen, J. (2010). Current and future CO? emissions from drained peatlands in Southeast Asia. Biogeosciences, 7(5), 1505–1514. https://doi.org/10.5194/bg-7-1505- 2010

Copyright

Copyright © 2025 S. Govindasamy, G. Rajesh Vasanthan, N. Sabarivasan, K . Sakthivel, J. Sridhar. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.