Annual Mangrove Maps for Pacific Island Countries from Landsat-8, 2013–2024

Wenting Cao developed an automated framework for annual mangrove mapping across all Pacific Island Countries from 2013 to 2024 using Landsat-8 imagery. The dataset reveals a net mangrove increase of 53,153 hectares (9.0%) with substantial local turnover, driven by landward migration and losses from cyclones and development. The framework achieved high classification accuracy (Overall Accuracy 94.5–96.0%) and is designed for cloud-prone, data-scarce regions.

Use Cases

• Analyze mangrove gain and loss dynamics based on the reported 110,523 ha gains and 57,376 ha losses.
• Model mangrove migration under sea-level rise based on the described landward expansion patterns.
• Assess the impact of tropical cyclones on coastal vegetation based on the cited loss drivers.
• Evaluate classification methods for cloud-affected imagery based on the adaptive temporal refinement framework.
• Study spatial heterogeneity in mangrove distribution based on the described differences between western high islands and eastern low-lying islands.

Strengths

• Covers a 12-year time series from 2013 to 2024 across all Pacific Island Countries.
• Achieved high classification accuracy with reported Overall Accuracy of 94.5–96.0% and Kappa of 0.89–0.92.
• Framework specifically addresses persistent cloud contamination and limited ground data, common challenges in the region.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data is packaged in a RAR archive, which requires specific extraction tools.

Provenance

Source

Wenting Cao via figshare.

Collection Method

Automated adaptive framework using Landsat-8 time series imagery with pixel-wise temporal refinement and grid-based adaptive sampling.

Time Range

2013–2024

Freshness

Last updated 2026-04-23 14:46:50; freshness should be verified.

Geography

All Pacific Island Countries