Selection of classification methods and quantification of changes in extent of mangrove forests over 6 years (2019 - 2025) in Ninh Binh province

Authors

  • Ha Tri Son Center for Technology and Nature Conservation
  • Le Thao Van Vietnam National University of Forestry
  • Hoang Kim Yen Vietnam National University of Forestry
  • Bui Ngoc Quy VNU-Central Institute for Natural Resources and Environmental Studies (VNU-CRES)
  • Nguyen Hai Hoa Vietnam National University of Forestry
  • Nguyen Thi Hai Van Vietnam National University of Forestry
  • Nguyen Thuy Dung Vietnam National University of Forestry
  • Nguyen Van Thi Xuan Mai Green Technology Trading Joint Stock Company
  • Ha Tri Thien Fauna & Flora, Vietnam Programme

DOI:

https://doi.org/10.55250/Jo.vnuf.11.2.2026.072-081

Keywords:

Change detection, Google Earth Engine, Ninh Binh, Random Forest, Sentinel-2, spectral indices

Abstract

Mangrove forests in Ninh Binh Province, a key part of the Red River Delta, provide important ecological services but face significant mapping challenges due to complex land use and tidal dynamics. This study evaluates four classification methods—Traditional RGB, NDVI thresholding, CMRI thresholding, and an Integrated Approach combining spectral bands with vegetation and water indices—using Sentinel-2 imageries (2019–2025) on the Google Earth Engine platform for a coastal area of 23,871.55 ha. Classification performance was assessed using User’s Accuracy, Producer’s Accuracy, Overall Accuracy (OA), the Kappa coefficient and the F1-Score. Findings indicate that the Integrated Approach, powered by a Random Forest algorithm, is the most robust, achieving an OA of 86.27–92.74%, Kappa coefficients of 0.78–0.88 and F1-Scores for the mangrove class of 91.16–99.50%. Unlike single-index methods (NDVI), which are sensitive to tidal submergence and underestimated the mangrove extent at only 2,576.23 ha in 2025, or RGB approaches that misclassify inland crops and overestimated the mangrove extent at up to 7,026.36 ha, the Integrated Approach effectively mitigates spectral confusion and spatial noise. Spatiotemporal analysis identified a stable core mangrove area of 3,750.05 ha out of 4,387.66 ha mapped in 2019, together with approximately 1,344.97 ha gained from non-mangrove areas, reflecting local restoration efforts. These findings underscore the necessity of multi-feature fusion (incorporating CMRI and MNDWI) for accurate coastal monitoring, providing a reliable framework for the sustainable management of Ninh Binh’s mangrove ecosystems.

References

[1]. Santiago Cadena & Jonathan G. Ochoa-Gómez. (2023). Mangroves: “Superhero” Ecosystems. Frontiers for Young Minds. 10.

[2]. Arumugam Sundaramanickam, Ajith Nithin & Thangavel Balasubramanian (2021). Role of Mangroves in Pollution Abatement. Mangroves: Ecology, Biodiversity and Management. Rajesh P. Rastogi, Mahendra Phulwaria & Dharmendra K. Gupta (eds.). Springer Singapore, Singapore. 257-278.

[3]. Charles Nyanga (2020). The Role of Mangroves Forests in Decarbonizing the Atmosphere. Carbon-Based Material for Environmental Protection and Remediation. Mattia Bartoli, Marco Frediani & Luca Rosi (eds.). IntechOpen, London.

[4]. Agustina Putri Cahyaningsih, Avyda Koza Deanova, Celin Maylani Pristiawati, Yaya Ihya Ulumuddin, L. I. A. Kusumawati & Ahmad Dwi Setyawan (2022). Review: Causes and impacts of anthropogenic activities on mangrove deforestation and degradation in Indonesia. International Journal of Bonorowo Wetlands. 12(1): 12-22

[5]. Ha Tri Son, Nguyen Hai Hoa & Vu Van Truong. (2023). Mangrove Cover-Based Vegetation Indices Mapping Using Planetscope Data in Tien Yen District Quang Ninh province. Journal of Forestry Science and Technology. 15: 127-138.

[6]. J.W Rouse, R.H Haas, J.A Schell & D.W Deering. (1974). Monitoring Vegetation Systems in the Great Plains with ERTS. Third ERTS-1 Symposium NASA, NASA SP-351, Washington DC. 309-317.

[7]. Yun Du, Yihang Zhang, Feng Ling, Qunming Wang, Wenbo Li & Xiaodong Li. (2016). Water Bodies’ Mapping from Sentinel-2 Imagery with Modified Normalized Difference Water Index at 10-m Spatial Resolution Produced by Sharpening the SWIR Band. Remote Sensing. 8(4): 354

[8]. K. Gupta, A. Mukhopadhyay, S. Giri, A. Chanda, S. Datta Majumdar, S. Samanta, D. Mitra, R. N. Samal, A. K. Pattnaik & S. Hazra. (2018). An index for discrimination of mangroves from non-mangroves using LANDSAT 8 OLI imagery. MethodsX. 5: 1129-1139.

[9]. A. R. Huete. (1988). A Soil-Adjusted Vegetation Index (SAVI). Remote sensing of environment. 25(3): 295–309.

[10]. A. Huete, K. Didan, T. Miura, E.P. Rodriguez, X. Gao & L.G. Ferreira. (2002). Overview of the radiometric and biophysical performance of the MODIS vegetation indices. Remote Sensing of Environment. 83: 195–213.

[11]. H. Xu. (2008). A new index for delineating built‐up land features in satellite imagery. International Journal of Remote Sensing. 29(14): 4269-4276.

[12]. Russell G. Congalton. (1991). A Review of Assessing the Accuracy of Classifications of Remotely Sensed Data. Remote sensing of Environment. 37: 35-46.

[13]. Lee R. Dice. (1945). Measures of the Amount of Ecologic Association Between Species. Ecology. 26(3): 297-302.

Downloads

Published

15-11-2026

How to Cite

Ha Tri Son, Le Thao Van, Hoang Kim Yen, Bui Ngoc Quy, Nguyen Hai Hoa, Nguyen Thi Hai Van, … Ha Tri Thien. (2026). Selection of classification methods and quantification of changes in extent of mangrove forests over 6 years (2019 - 2025) in Ninh Binh province . Journal of Forestry Science and Technology, 11(2), 072–081. https://doi.org/10.55250/Jo.vnuf.11.2.2026.072-081

Issue

Section

Resource management & Environment

Categories

Most read articles by the same author(s)

Similar Articles

<< < 13 14 15 16 17 18 19 20 21 22 > >> 

You may also start an advanced similarity search for this article.