Quantifying Total Shrimp Pond Coverage Using Remote Sensing and a Machine Learning Algorithm: A Case Study in the Ecuadorian Coast

Remote sensing and machine learning have emerged as powerful tools for monitoring rapid land‐use changes in coastal ecosystems. However, accurately quantifying the spatial extent and temporal dynamics of shrimp aquaculture remains challenging, particularly in regions with complex coastal landscapes such as the Ecuadorian coast. This study aims to develop and apply a reproducible computational workflow to automatically detect and map shrimp ponds using multispectral Sentinel‐2 imagery and machine learning techniques. The proposed approach integrates spectral, structural, and neighborhood‐based features within a random forest classification framework to distinguish shrimp ponds from other land and water classes and to delineate individual pond boundaries. The analysis, conducted for the period 2019–2024, achieved an average class accuracy of 95.68% on the held‐out test subset. In a unified feature‐ablation and temporal‐holdout experiment, the complete representation achieved 96.11% accuracy and a Cohen's kappa of 0.838 on the unseen 2024 mosaic. Hessian edge and neighborhood features improved performance independently and produced the strongest result when combined. Results indicate that mapped shrimp‐pond‐water surface area increased by 34.4% relative to the 2019 baseline, reaching approximately 164,000 ha by 2024. Spatial analysis reveals that 67% of shrimp ponds are concentrated in Guayas Province, which also exhibited the highest expansion rates. Despite this spatial growth, trends suggest that shrimp production has continued to increase even as the rate of pond expansion has slowed in recent years. These findings demonstrate the effectiveness and scalability of the proposed methodology for monitoring aquaculture dynamics at regional scales. The generated geospatial dataset provides valuable insights for ecological informatics applications and supports evidence‐based decision‐making aimed at balancing aquaculture development with coastal ecosystem conservation.

Authors

Institutions

Publication Details

Journal
New Zealand Journal of Marine and Freshwater Research
Published
2026-11-13
DOI
https://doi.org/10.1002/nzm2.70091
Primary Topic
Remote-Sensing Image Classification
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Quantifying Total Shrimp Pond Coverage Using Remote Sensing and a Machine Learning Algorithm: A Case Study in the Ecuadorian Coast

Daniel Ochoa, Stanislaus Sonnenholzner, Romina Cifuentes, Jorge Vulgarin
4 links
New Zealand Journal of Marine and Freshwater Research
Remote-Sensing Image Classification
article

Quantifying Total Shrimp Pond Coverage Using Remote Sensing and a Machine Learning Algorithm: A Case Study in the Ecuadorian Coast

Daniel Ochoa, Stanislaus Sonnenholzner, Romina Cifuentes, Jorge Vulgarin
article en

Abstract

Remote sensing and machine learning have emerged as powerful tools for monitoring rapid land‐use changes in coastal ecosystems. However, accurately quantifying the spatial extent and temporal dynamics of shrimp aquaculture remains challenging, particularly in regions with complex coastal landscapes such as the Ecuadorian coast. This study aims to develop and apply a reproducible computational workflow to automatically detect and map shrimp ponds using multispectral Sentinel‐2 imagery and machine learning techniques. The proposed approach integrates spectral, structural, and neighborhood‐based features within a random forest classification framework to distinguish shrimp ponds from other land and water classes and to delineate individual pond boundaries. The analysis, conducted for the period 2019–2024, achieved an average class accuracy of 95.68% on the held‐out test subset. In a unified feature‐ablation and temporal‐holdout experiment, the complete representation achieved 96.11% accuracy and a Cohen's kappa of 0.838 on the unseen 2024 mosaic. Hessian edge and neighborhood features improved performance independently and produced the strongest result when combined. Results indicate that mapped shrimp‐pond‐water surface area increased by 34.4% relative to the 2019 baseline, reaching approximately 164,000 ha by 2024. Spatial analysis reveals that 67% of shrimp ponds are concentrated in Guayas Province, which also exhibited the highest expansion rates. Despite this spatial growth, trends suggest that shrimp production has continued to increase even as the rate of pond expansion has slowed in recent years. These findings demonstrate the effectiveness and scalability of the proposed methodology for monitoring aquaculture dynamics at regional scales. The generated geospatial dataset provides valuable insights for ecological informatics applications and supports evidence‐based decision‐making aimed at balancing aquaculture development with coastal ecosystem conservation.

New Zealand Journal of Marine and Freshwater ResearchVol. 60(4)
Escuela Superior Politecnica del Litoral (EC)
Escuela Superior Politécnica del Litoral, European Commission
Life below water, Zero hunger
Remote-Sensing Image Classification
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon