Mangrove biomass monitoring has become critical today, as it aids in measuring the amount of carbon that will be contained in these forests in the future. Remote sensing is possibly the solution for estimating AGB on a large scale. It allows biomass assessments over large areas, providing a spatially comprehensive measure of forest biomass variance in a relatively short time and at a low cost. Because of the penetration capability of radar remote sensors, the high-frequency L-band and P-band can be used to measure aboveground biomass effectively. The study emphasizes the importance of Synthetic Aperture Radar imaging and the correlation between backscatter reflectance of remotely sensed data and live biomass over a given area. Despite many studies on the use of machine learning algorithms for biomass estimation, no single model fits all problems. Furthermore, finding the best machine learning model is critical for advancing global expertise in the field of mangrove forest management. The hybrid models such as extreme gradient boosting regression (XGBR) and extreme Gradient Boosting (XBoost) are also used along with regression models linear regression, multilayer perceptron (MLP), and machine learning approaches like support vector Machine (SVM), random and random forest (RF) is used in mangrove studies. In comparison to the allometric equations based on ground truth data, remote sensing methods are coming up as an effective tool for AGB estimation of mangroves. This vision survey focuses on SAR radar remote sensing-based studies in mangrove forest biomass estimation and the limitation of SAR imaging in AGB estimation.
Read full abstract