Dueñas-Tovar et al. (2026) Integration of spectral indices and precipitation data to assess river morphometric features in a tropical semi-humid environment

Identification

• Journal: Environmental Challenges
• Year: 2026
• Date: 2026-03-01
• Authors: Jairo Dueñas-Tovar, María Jaya-Montalvo, André Luiz Lopes de Faria, Fernando Morante-Carballo
• DOI: 10.1016/j.envc.2026.101457

Research Groups

• Facultad de Ingeniería en Ciencias de la Tierra, Escuela Superior Politécnica del Litoral (ESPOL), Guayaquil, Ecuador
• Centro de Investigación y Proyectos Aplicados a las Ciencias de la Tierra, Escuela Superior Politécnica del Litoral (ESPOL), Guayaquil, Ecuador
• Geo-Recursos y Aplicaciones (GIGA), Escuela Superior Politécnica del Litoral (ESPOL), Guayaquil, Ecuador
• Facultad de Ingeniería Mecánica y Ciencias de la Producción, Escuela Superior Politécnica del Litoral (ESPOL), Guayaquil, Ecuador
• Laboratório de Geomorfología do Quaternário, Departamento de Geografía, Universidade Federal de Viçosa, Viçosa, MG, Brazil
• Facultad de Ciencias Naturales y Matemáticas, Escuela Superior Politécnica del Litoral (ESPOL), Guayaquil, Ecuador

Short Summary

This study developed a reproducible remote sensing workflow using eight optical indices and a Random Forest algorithm to assess river channel mobility (lateral shift and sinuosity) in a data-limited tropical semi-humid environment. The workflow successfully identified episodic, reach-specific channel adjustments, with lateral shifts up to 500 meters, and revealed a short-term negative correlation between antecedent precipitation and lateral displacement.

Objective

• Develop a reproducible remote sensing workflow using eight optical indices and a Random Forest algorithm to evaluate river channel mobility on a mixed-pattern river in data-limited fluvial contexts.
• Extract multiannual river centerlines through river water masks for the quantification of year-to-year lateral shifts.
• Examine the relationship between hydrological forcing and river channel mobility using time-lag correlation analysis.
• Discuss the practical implications of centerline mobility in data-scarce fluvial contexts.

Study Configuration

• Spatial Scale: Pedro Carbo River (PRC) in the coastal region of Ecuador, spanning nearly 60 kilometers in the lower parts of the Guayas Province. The study area's altitude ranges from 10 meters to 100 meters above sea level.
• Temporal Scale: Analysis of 13 satellite images from 1985 to 2024 (approximately 40 years). Monthly average precipitation data from 1985 to 2024.

Methodology and Data

• Models used:
  • Random Forest classifier algorithm (scikit-learn package in Python) for binary classification of water bodies.
  • GRASS GIS (v.7.8.6) tools (r.thin for skeletonization, r.to.vect for vectorization).
  • QGIS Software v. 3.10.12 Prizren for geometric correction (georeferencing).
  • Pearson Correlation Analysis for time-lag correlation between precipitation and lateral shifts.
• Data sources:
  • Satellite imagery: Landsat (5, 7 ETM+, 8 OLI-TIRS) and Sentinel-2a/MSI missions.
  • Precipitation data: Monthly average precipitation data from POWER NASA (1985-2024).
  • Digital Elevation Model (DEM): 30-meter resolution DEM (Laipelt et al., 2024).
  • Geological units: BGS-CODIGEM (1993).
  • Land cover: MAGAP (2016).
  • Optical indices: NDWI, mNDWI, TC-WET, AWEI, NDVI, BSI, WRI, and NDMI.

Main Results

• The Random Forest algorithm achieved an 87% accuracy in correctly predicting water bodies, with mNDWI, AWEI, and TC-WET identified as the most important indices.
• Lateral shifts revealed reach-specific, high-magnitude adjustment events, with displacements up to approximately 500 meters, concentrated in localized river segments, particularly in upstream sections (e.g., XS 01, XS 02, XS 04, XS 06, XS 12).
• Sinuosity values showed a persistent curvilinear planform, generally ranging from 1.4 to 1.7, indicating spatial heterogeneity in curvature but minor interannual variability compared to lateral shifts.
• Time-lag correlation analysis indicated a significant negative correlation (r = -0.804, p = 0.029) between antecedent annual precipitation and lateral channel displacement at a one-year lag, suggesting a reduction in lateral shift after a year of above-average rainfall. No significant correlations were observed for other time lags, implying no long-term river memory effects.
• The results captured episodic and reach-specific channel adjustments, providing a practical basis for monitoring channel mobility and fluvial exposure hazards in data-scarce environments.

Contributions

• Developed a novel and reproducible remote sensing workflow integrating machine learning (Random Forest) with multiple spectral indices for robust water delineation and subsequent river mobility (lateral shifts) and geometrical analysis (sinuosity) in data-limited tropical semi-humid environments prone to avulsions.
• Quantified reach-specific, high-magnitude lateral shifts (up to approximately 500 meters) and characterized sinuosity patterns over a 40-year period using multi-temporal satellite imagery.
• Identified a short-term, one-year lagged negative correlation between antecedent precipitation and lateral channel displacement, offering insights into hydroclimatic influences on river dynamics in data-scarce regions.
• Provided a practical basis for monitoring channel mobility and fluvial exposure hazards, informing decision-making for land-use planning and infrastructure avoidance in unstable river sections.

Funding

• Project: “Registro de sitios de interés geológicos del Ecuador para estrategias de desarrollo sostenible” (Register of Geological Sites of Interest in Ecuador for Sustainable Development Strategies), funded by Escuela Superior Politécnica del Litoral (ESPOL) under institutional code CIPAT-004–2024.
• Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – Brasil (CAPES) - Finance Code 001.

Citation

@article{DueñasTovar2026Integration,
  author = {Dueñas-Tovar, Jairo and Jaya-Montalvo, María and Faria, André Luiz Lopes de and Morante-Carballo, Fernando},
  title = {Integration of spectral indices and precipitation data to assess river morphometric features in a tropical semi-humid environment},
  journal = {Environmental Challenges},
  year = {2026},
  doi = {10.1016/j.envc.2026.101457},
  url = {https://doi.org/10.1016/j.envc.2026.101457}
}