Satellite Imagery and Machine Learning Map Soil Salinity in the Red River Delta
A study co-authored by Huu Duy Nguyen, Dung Duc Tran (HAPRI Senior Research Associate), Edward Park and Minh Cuong Ha combines remote sensing data from Sentinel-2A imagery with machine learning to monitor both the spatial and temporal distribution of soil salinity in Tien Hai District, Thai Binh Province, in Vietnam's Red River Delta. Among the models tested, CatBoost achieved the highest performance, and the mapped shares of non-saline, slightly saline and strongly saline land differed between May 2023 and March 2025.
Context
The authors describe the Red River Delta (RRD) as one of the regions most severely affected by salinization, posing challenges to agricultural productivity, food security, and environmental sustainability. In the study's words, understanding the spatial and temporal distribution of soil salinity is essential for assessing land degradation and developing effective mitigation strategies to support sustainable local development.
Yet previous research has predominantly focused on assessing soil salinity at isolated points in time, with limited attention given to its temporal variability. The authors note that this gap is especially critical in the RRD, and they focus on Tien Hai District in Thai Binh Province.
Methods
The authors integrated remote sensing data from Sentinel-2A imagery with machine learning algorithms — CatBoost (CBR), AdaBoost (ADB), XGBoost (XGR), Decision Trees (DTR) and Gradient Boosting Regression (GBR) — to model and predict salinity patterns over time. A total of 106 soil salinity samples were collected during the rainy season of May 2023 and the dry season of March 2025, and 45 environmental and spectral variables were extracted to serve as input for model training.
Key findings
Among the models tested, CatBoost (CBR) achieved the highest performance, with an R2 value of 0.823, followed by Gradient Boosting Regression (0.802), AdaBoost (0.796), XGBoost (0.757) and Decision Trees (0.746).
The spatial mapping indicated that in May 2023, approximately 71.9% of the area was non-saline, 12.78% slightly saline and 15.3% strongly saline.
By March 2025, these shares had shifted to 63.28% non-saline, 21.62% slightly saline and 15.01% strongly saline.
Keywords:
Red River Delta
Soil Salinity
Remote Sensing
Machine Learning
Sentinel-2
Agriculture
Link:
Citation:
Nguyen, H. D., Tran, D. D., Park, E., & Ha, M. C. (2026). Data-driven mapping of soil salinity in the Red River Delta: Sentinel-2 MSI and machine learning integration. Physical Geography, 1-28.
