**1. Introduction**

Mangrove ecosystems are highly efficient blue carbon sinks, owing to their high productivity and low respiration rates, which allow them to store large amounts of biomass and organic carbon for a long time [1,2]. The capability for carbon sequestration of coastal ecosystems, including mangrove forests, has been reported to be 10–50 times higher than that of the terrestrial ecosystem [3]. Being among the most productive ecosystems, they can effectively mitigate climate change [4,5]. Therefore, the accurate estimation of the aboveground biomass (AGB) of mangrove plantations is essential for identifying the patterns of distribution in tropical and subtropical coastal zones to assess emissions from deforestation and carbon sinks from reforestation [6].

By the end of the 1990s, the total area occupied by mangrove forests in China was smaller than 15,000 ha, and had been reduced by 68.7% since its historical peak [7], owing to urban expansion, tidal flat reclamation, and deforestation for cultivation [8,9]. Since then, afforestation and reforestation projects have been implemented to conserve and restore mangroves. Mangrove plantation has been encouraged such that, by 2015, the area occupied by mangrove forests in China reached 22,419 ha [10]. The AGB of mangrove plantation should be accurately measured when monitoring, restoring, and managing wetland ecosystems, because it can help support global climate change mitigation programs, such as the Reducing Emissions from Deforestation and Forest Degradation in (REDD+), as well as the Payments for Ecosystem Services (PES) schemes [11,12].

However, field measurements on the biomass of mangroves are challenging, because they are distributed in intertidal zones that are difficult to access [13]. Thus, remotely sensed images have been widely used for this purpose [14–16]. Accordingly, regression models have been proposed by constructing relationships between the AGB and variables derived from various data sources. In general, optical and synthetic aperture radar (SAR) images are commonly used for AGB estimation studies [17,18]. The bands and vegetation indices (VIs) derived from the optical images vary according to water and chlorophyll content, and the structure of the leaf cavity of the vegetation that are correlated to the type of plant or its stages of growth. Thus, they can be used to monitor the biomass of forests [19–21]. Optical images are widely available—for example, Moderate Resolution Imaging Spectroradiometer (MODIS) [22], Landsat [19], IKONOS [21], SPOT [23], and WorldView-2 images [6]. However, these remote sensors are not capable of penetrating the surface of the canopies of forests to obtain their structure and the heights of trees needed for biomass estimation [24,25]. Since SAR images penetrate the canopy [26], they can be used to determine the structure of the canopy by emitting radiation to detect and measure branches and trunks [27]. Hence, the C-band and X-band of SAR images such as Rardarsat-2, ALOS PALSAR, and airborne SAR images are useful for monitoring the biomass of mangrove plantations [14,28–31]. However, SAR are images acquired by receiving transmitted signals contain speckle noise usually caused by the constructive or destructive interference of backscattered microwave signals that degrade image quality, and thus, may not provide accurate information concerning the target objects [32].

Optical and SAR images have considerable limitations in estimating forest biomass accurately [33]. Past studies have shown that optical and SAR images can be integrated to acquire the spectral and structural features of forest canopies to improve the accuracy of their predicted biomass [17,18,34,35]. This integration usually involves incorporating variables derived from the optical and SAR images or fusing them into new datasets (such as in wavelet transforms or principal component analysis) [18]. However, these variables exhibit saturation effects for highly dense mangrove forests that limit their availability to estimate only within a specific range of the AGB [33,36]. Hence, they can only be used to estimate AGB for areas with low biomass.

To overcome the above limitations, recent studies have focused on the vertical structure of mangrove forests (e.g., tree height) using Interferometric Synthetic Aperture Radar (InSAR), Light Detection and Ranging (LiDAR), and stereo photogrammetry of overlapping photographs, to help resolve the saturation problem [37–40]. Aerial photographs using the structure-from-motion (SfM) algorithm and the Unmanned Aerial Vehicle (UAV) platform are a low-cost option to measure the heights of trees [41]. The UAV-based digital surface models (DSM) can determine the relative tree height of mangrove forests, because they mainly grow over even terrain [42]. Hence, the integration of optical, SAR and UAV-based DSM data to represent the spectrum of the canopy, structure, and height of mangroves can improve the accuracy of the estimation of AGB for dense mangrove forests. Navarro et al. (2019) estimated mangrove AGB by combining UAV-based tree height, Sentinel-1, and Sentinel-2 images, and provided accurate estimates for young and sparse mangrove plantations [17]. However, the e ffectiveness of this approach needs to be examined further for dense and complicated mangroves.

This research aims to address the aforementioned gaps in the estimation of the AGB. We integrate images from multiple sources—GF2 optical, GF3 SAR, and fixed-wing UAV-based DSM data— to estimate the AGB of mangrove plantations. The objectives of this study are (1) to develop prediction models for the AGB using original and composite bands generated from optical, SAR, and UAV-based DSM data; (2) to evaluate the e ffectiveness of the AGB models and select the best one; (3) to determine the importance of the chosen parameters; and (4) to map the AGB to observe the spatial pattern of the biomass of a mangrove plantation in comparison with field surveys and its sequence of growth.

#### **2. Materials and Methods**

## *2.1. Study Area*

This study was conducted in the mangrove nature reserve area of Guangdong Province in Dawei Bay of Qi'ao Island, in the Pearl River estuary of China (113◦3640" E–113◦3915" E, 22◦2340" N– 22◦2738" N) [43,44]. It is a typical tropical and subtropical wetland ecosystem. The area of the reserve is 5103 ha, covering about 700 ha of mangrove forests: the largest contiguous area of artificially planted mangroves in China [6,45]. Mangrove forests are characterized by high spatial variability that represents a dynamic landscape. The mangrove forests planted in the study area were composed mainly of a fast-growing species, *Sonneratia apetala* (*S. apetala*), which was introduced from Bengal. They belong to the woody mangrove species, with features of high adaptability and seed production [46]. The heights of their trees usually increase by about 1.5 m during each of the first few years [47,48]. They were the tallest tree species in the study area, ranging from 2 m to 20 m, and the diameter at breast height (DBH) of older trees can be as large as 30 cm.

*S. apetala* has been artificially planted on the island since 1999 to control invasive species, (i.e., *Spartina alterniflora*) and reconstruct mangrove forests. *S. apetala* generally has an a fforestation specification of 1–2 m × 1.5 m with high densities. The tree ages range from 1 year to 17 years with high biomass variability and complication. Its a fforestation process runs seaward from land, which implies a gradient distribution of tree age and AGB. The mangrove plantations had extended throughout the study area, and *S. apetala* had become the dominant species, covering more than 80% of the mangrove forest.
