Remote sensing-based mangrove blue carbon assessment in the Asia-Pacific: A systematic review

Background

Mangrove forest carbon stocks, both above- and below-ground, offer opportunities for blue carbon sequestration as a potentially scalable, cost-effective nature-based climate change solution. Accurate monitoring of mangrove forests and their carbon storage capacity is important for informing practitioners and policymakers interested in blue carbon conservation and crediting. Remote sensing and modeling can help to estimate above-ground mangrove carbon stocks through robust, low-cost, high-accuracy, non-invasive data collection. This review describes trends in remote sensing and modeling approaches, along with attributes and variation in blue carbon estimates.

Goals and Methods

The researchers conducted a literature search to identify articles on remote sensing and modeling of mangrove forest blue carbon. They selected 105 articles from 1990 to June 2023 to review and analyze trends in remote sensing and modeling technology and algorithms.

Conclusions and Takeaways

This review found that the most commonly used multispectral sensors were Sentinel-2 MSI and Landsat 8 OLI, with the most commonly used synthetic-aperture radar (SAR) sensor was ALOS-2 PALSAR-2. There was a shift from parametric modeling (i.e., least square linear regression, non-linear regression) to non-parametric machine learning models for genetic algorithms, the latter of which better captures stand complexity and accurate blue carbon measurements. The most common machine learning models used with remote sensing data are Support Vector Machine (SVM), Random Forest Regression (RFR), and Extreme Gradient Boost Regression (XGBoost or XGBR). The authors recommend using a combination of remote sensing data, field-based data, and machine learning models to improve estimates of above-ground carbon.