*Article* **Mapping Multi-Decadal Mangrove Extent in the Northern Coast of Vietnam Using Landsat Time-Series Data on Google Earth Engine Platform**

**Thuy Thi Phuong Vu 1, Tien Dat Pham 2,\*, Neil Saintilan 2, Andrew Skidmore 2,3, Hung Viet Luu 4, Quang Hien Vu 1, Nga Nhu Le 5, Huu Quang Nguyen <sup>6</sup> and Bunkei Matsushita <sup>7</sup>**


**Abstract:** A pixel-based algorithm for multi-temporal Landsat (TM/ETM+/OLI/OLI-2) imagery between 1990 and 2022 monitored mangrove dynamics and detected their changes in the three provinces (i.e., Thai Binh, Nam Dinh and Hai Phong), which are located on the Northern coast of Vietnam, through the Google Earth Engine (GEE) cloud computing platform. Results showed that the mangrove area in the study area decreased from 2960 ha in 1990 to 2408 ha in 1995 and then significantly increased to 4435 ha in 2000 but later declined to 3502 ha in 2005. The mangrove areas experienced an increase from 4706 ha in 2010 to 10,125 ha in 2020 and reached a highest peak of 10,630 ha in 2022. In 2022, Hai Phong province had the largest area of mangrove (3934 ha), followed by Nam Dinh (3501 ha) and Thai Binh (3195 ha) provinces. The overall accuracies for 2020 and 2022 were 94.94% and 91.98%, while the Kappa coefficients were 0.90 and 0.84, respectively. The mangrove restoration programs and policies by the Vietnamese government and local governments are the key drivers of this increase in mangroves in the three provinces from 1990 to 2022. The results also demonstrated that the combination of Landsat time series images, a pixel-based algorithm, and the GEE platform has a high potential for monitoring long-term change of mangrove forests during 32 years in the tropics. Moreover, the obtained mangrove forest maps at a 30-m spatial resolution can serve as a useful and up-to-date dataset for sustainable management and conservation of these mangrove forests in the Red River Delta, Vietnam.

**Keywords:** mangrove; remote sensing; Landsat; Google Earth Engine; Red River Delta; Vietnam

#### **1. Introduction**

Mangrove forests are trees and shrubs found in tidal wetlands and located in the tropical and sub-tropical region between 30◦N and 30◦S latitude [1]. They cover only 0.1% of Earth's continental surface, yet they provide a wide range of ecosystem services, including water purification, natural hazards reduction, soil and water conservation, shoreline protection and enhanced local livelihood and are considered as natural-based solutions in

**Citation:** Vu, T.T.P.; Pham, T.D.; Saintilan, N.; Skidmore, A.; Luu, H.V.; Vu, Q.H.; Le, N.N.; Nguyen, H.Q.; Matsushita, B. Mapping Multi-Decadal Mangrove Extent in the Northern Coast of Vietnam Using Landsat Time-Series Data on Google Earth Engine Platform. *Remote Sens.* **2022**, *14*, 4664. https://doi.org/ 10.3390/rs14184664

Academic Editor: Guangsheng Chen

Received: 22 July 2022 Accepted: 16 September 2022 Published: 19 September 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

dealing with climate change impact [2]. The areas of mangrove forests have been changed significantly on a global scale due to anthropogenic disturbance (i.e., urbanization and increased agricultural production) and climate change [2]. However, recent studies pointed out that the rate of deforestation has been decreasing [3], and mangroves have expanded in some Southeast Asian nations and in Australia [4,5].

Among approximately 3260 km of the total coastal length of Vietnam, 2365 km are covered by mangroves representing 29 coastal provinces. Mangrove ecosystems, therefore, play an important role in protecting the Vietnamese coastline against flooding and erosion, providing biodiversity and livelihood for coastal communities as well as sequestering carbon, known as 'blue carbon' [6,7]. The Vietnamese mangroves are mainly distributed in the two deltas, the Red River Delta (RRD) in the north and the Mekong River Delta in the south [8,9]. However, the mangrove forest area in Vietnam has decreased dramatically over the past 70 years, falling from 408,500 ha in 1943 to 178,000 ha in 2000, and then continuously shrinking to 138,318 ha in 2016 [10–14]. Therefore, it is essential to obtain accurate information about mangrove forests in the past and current state that is useful to manage and effectively protect mangrove ecosystems across the Vietnamese coastline. However, there is no up-to-date map of mangroves in Vietnam; thus, mapping mangrove forests and detecting their dynamics are vital for sustainable conservation and management of mangrove resources.

Mapping mangroves at a large scale remains challenging due to the costs and labour intensiveness in field measurements for large areas. In recent years, remote-sensingbased techniques have become widely used for monitoring the Earth's surface including mangrove forests and has proven to be a key tool to effectively map mangrove dynamics in large areas in Southeast Asia [12,15–18]. Pixel-based and object-based approaches are the most common techniques for mapping mangroves and detecting their changes. These approaches can provide the most frequently updated data at a low cost [19,20]. For instance, the distribution of mangroves at a global scale using a multi-temporal Landsat dataset and a supervised Maximum Livelihood Classification (MLC) was reported by Giri et al. [17] with an overall accuracy ranging from 79% to 86%. In Vietnam, Nguyen-Thanh et al. [12] used the Landsat time series data and an object-based image analysis to monitor mangrove extent in the Ca Mau Peninsula, Vietnam, whereas Pham and Brabyn [13] used the SPOT imagery and a support vector machine (SVM) classifier to map mangrove dynamics in the Can Gio biosphere reserve region with overall accuracies of 77 and 83%. However, to date there is no spatial distribution map of mangroves along the RRD, and there is a complete lack of reliable updated statistical data of mangroves in the three coastal provinces of the RRD.

More recently, with the development of open-source software and cloud computing platforms such as the Google Earth Engine (GEE), the applications of remote sensing techniques in monitoring mangrove changes have become more popular. Previous studies have widely applied the GEE platform to map mangrove changes using multispectral sensors such as the Thematic Mapper (TM), the Enhanced Thematic Mapper Plus (ETM+), the Operational Land Imager (OLI), and the Operational Land Imager-2 (OLI-2) in Landsat and the Multi-Spectral Instrument (MSI) in Sentinel-2 [21–25]. However, to the best of our knowledge, the up-to-date mangrove forests maps and their change detection using time series Landsat (TM/ETM+/OLI/OLI-2) imagery between 1990 and 2022 have not been reported in Vietnam. Thus, this study aims to fill the gap in the current literature by investigating a pixel-based algorithm: (1) to map multi-decadal mangrove dynamics using Landsat time series data between 1990 and 2022 through the GEE platform, (2) to provide up-to-date statistical analysis of areas of mangrove forests in the Northern coast of Vietnam for the first time in 2022 using Landat-9 OLI-2 as an important national mangrove database, and (3) to provide a useful tool for decision makers in supporting the mangrove conservation and management in Vietnam.

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

#### *2.1. Study Area*

In the current study, three coastal provinces in the RRD were selected to test the performance of the pixel-based algorithm. They are Hai Phong province (20◦51 54.5004N, 106◦41 1.7880E), Thai Binh province (20◦27 0N, 106◦20 24.07E) and Nam Dinh province (20◦16 45.048N, 106◦12 18.533E), which are shown in Figure 1.

**Figure 1.** Study area in the northern coast of Vietnam: (**a**) map of Vietnam; (**b**) three coastal provinces in the RRD.

The RRD, consisting of nine provinces (i.e., Hai Duong, Bac Ninh, Vinh Phuc, Hung Yen, Thai Binh, Nam Dinh, Ha Nam, Ninh Binh and Quang Ninh) is the second-largest delta and is located in the northern region of Vietnam with a total area of 15,000 km2. With a population of 22 million, the RRD is the most densely populated region in Vietnam [26]. In 2019, the population density of the RRD had reached 1064 inhabitants per km<sup>2</sup> [27]. The total area of the three provinces is approximately 473,700 ha, of which Nam Dinh province is the largest with 166,800 ha, followed by Thai Binh province and Hai Phong province with an area of 153,400 ha and 152,300 ha, respectively [8].

The mangrove ecosystems in the RRD play a key role in protecting coastal habitats, supporting biodiversity, and providing coastal resources for local people. In the RRD, the Xuan Thuy National Park was listed as the first Ramsar site in Southeast Asia in 1989 to promote the sustainable conservation of wetlands [28]. The Ramsar site was defined as "the sustainable utilization of wetlands for the benefit of mankind in a way compatible with the maintenance of the natural properties of the ecosystem" [29]. There are five dominant mangrove species observed in this park being *Sonneratia caseolaris*, *Kandelia obovata*, *Aegiceras corniculatum*, *Rhizophora stylosa* and *Avicennia marina* [30]. Furthermore, this park is the habitat to 116 flora species and 106 fish and has significantly contributed to wetland biodiversity protection on the northern coastline of Vietnam [30]. A nature reserve, which is located in Thai Binh Province, is well-known as the Bird Conservation area, and there are several rare species listed in the Vietnamese Red Book [27]. There are four seasons in the RRD with a mean annual temperature of approximately 28 ◦C. The annual precipitation recorded in the last ten years is around 1800–2000 mm. In recent years, the RRD has been seriously affected by climate variability including higher temperatures, storms and flooding [28]. In particular, 2020 was recorded as the hottest year over the last 45 years and likely resulted in a dieback of mangroves [31,32].

#### *2.2. Materials*

#### Satellite Data

Multi-decadal Landsat surface reflectance (SR) data obtained through the GEE platform was used to map mangrove dynamics in the study area (Table 1). We used Collection 2, which were atmospherically corrected SR data with a single-channel algorithm developed

by the National Aeronautics and Space Administration (NASA) Jet Propulsion Laboratory (JPL). All Landsat time series Collection 2 SR data used in the current study (Table 1) were acquired using the Java scripts on the GEE.

**Table 1.** Time-series Landsat imagery used for 32 years in the study area.


Considering the seasonality changes of mangrove forests and their species, a total of 82 cloud-free Landsat scenes between 1990 and 2022 were used to map mangrove dynamics in the RRD. We applied an image normalisation technique to make all images consistent during the pre-processing process. To minimize the effects of tidal and water levels, we selected the datasets acquired in early morning when the tidal level was the lowest.

#### *2.3. Methods*

#### 2.3.1. Generation of Training and Validation Datasets for the Study Area

In this study, the training and the validation data were obtained from very high spatial resolution images in Google Earth Pro imagery (2020). A total of 2370 points were randomly selected, of which 1896 points (80%) were used for the training set and 474 points (20%) were used for the validation set (Figure 2).

As shown in Figure 2, a polygon including mangrove forest and non-mangrove was created from the spatial data and consisted of a sea dyke and river layers with a total area of 35,566 ha. A 3 km buffer generated from the high-resolution images of Google Earth Pro imagery in 2020 was used to capture the entire mangrove forests area as suggested by Thomas et al. [33] and Bunting et al. [34].

#### 2.3.2. Computation of Spectral Indices

Four indices were calculated from the SR data of Landsat (5/7/8/9) images to identify vegetation and open surface water bodies as suggested by Wang et al. [22] and Pham et al. [35]. They are the Normalized Difference Vegetation Index (NDVI) [36], the Enhanced Vegetation Index (EVI) [37], the Land Surface Water Index (LSWI) [38] and the modified Normalized Difference Water Index (mNDWI) [39]. The equations are shown below:

$$\text{NDVI} = \frac{\rho\_{nir} - \rho\_{rel}}{\rho\_{nir} + \rho\_{rel}} \tag{1}$$

$$\text{EVI} = 2.5 \times \frac{\rho\_{\text{air}} - \rho\_{\text{red}}}{\rho\_{\text{air}} + 6 \times \rho\_{\text{red}} - 7.5 \times \rho\_{\text{blue}} + 1} \tag{2}$$

$$\text{LSWI} = \frac{\rho\_{\text{air}} - \rho\_{\text{swir}}}{\rho\_{\text{air}} + \rho\_{\text{swir}}} \tag{3}$$

$$\text{mNDWI} = \frac{\rho\_{\text{green}} - \rho\_{\text{swir}}}{\rho\_{\text{green}} + \rho\_{\text{surir}}} \tag{4}$$

where *ρred*, *ρgreen*, *ρblue* and *ρswir* are the surface reflectance at red (band 3 for TM/ETM+ or band 4 for OLI and OLI-2), green (TM/ETM+ band 2 or OLI/OLI-2 band 3), blue (TM/ETM+ band 1 or OLI/OLI-2 band 2) and short-wave infrared (SWIR: TM/ETM+ band 5 or OLI/OLI-2 band 6) bands, respectively.

We proposed a framework using a pixel-based mapping algorithm to map mangrove forests and automatically detect their changes using time series Landsat images from 1990 to 2022 through the GEE platform as shown in Figure 3. We developed the Python scripts using the geemap package (https://github.com/giswqs/geemap, accessed on 16 July 2021) to map mangrove extent in the RRD.

**Figure 2.** Training and validation points for mangrove mapping during the year 2020 in the RRD, Vietnam: (**a**) Hai Phong; (**b**) Thai Binh; (**c**)

Nam Dinh.

**Figure** 

#### 2.3.3. Mangrove Mapping Algorithm

In this study, we used a pixel-based algorithm, which was developed by Wang et al. [22] for mapping coastal wetlands using time series Landsat datasets in 2018 to generate annual maps of mangrove forests between 1990 and 2022. The algorithm includes three steps for processing each pixel: (1) identifying open surface water body and green vegetation, (2) calculating annual frequency for surface water body and vegetation, and (3) generating annual maps of mangrove forest. The present study used data in 2020 and in 2022 to check and modify the thresholds provided by the original study and then used these modified thresholds to estimate the mangrove area for other years.

We used a frequency-based approach from Landsat time series to mitigate the effect of periodical tidal dynamics and bad-quality observations as suggested by Wang et al. [22]. The open surface water body and vegetation frequencies in a year were calculated using the following equations:

$$\text{WF} = \frac{\text{Nwater}}{\text{Ngood}} \tag{5}$$

$$\text{VF} = \frac{\text{Nvegetation}}{\text{Ngood}} \tag{6}$$

where

WF and VF are the frequencies of surface water body and vegetation, respectively (−1 to 1).

Nwater and Nvegetation are the numbers observations identified as water and vegetation in a year, respectively, while

Ngood is the number of observations with good quality, which was characterised as cloud and shadow-free during the observed year.

We defined the thresholds based on the training data collected from the high spatial resolution Google Earth images in 2020 to identify evergreen wetlands as follows:

$$\text{Everyreeen} = \text{VF} \ge 0.9 \text{ and } \text{WF} \le 0.2 \text{ and } \text{DEM} \le 5 \text{ m and Slope} \le 5^{\circ} \tag{7}$$

where WF and VF are the frequencies of surface water body and vegetation, respectively. These indices values are ranked between −1 and 1, while the Shuttle Radar Topography Mission (SRTM) Digital Elevation Model (DEM) data were used to generate a slope layer in the current study [22].

To identify open surface water bodies and green vegetation, a combination of mNDWI and two vegetation indices (EVI and NDVI) was employed to reduce the errors in mixed pixels of the water body and vegetation [22]. In this study, almost all pixels identified as water body in 2020 have an EVI ≤ 0.137 and mNDWI > EVI or mNDWI > NDVI.

The NDVI and the EVI are two popular indices to detect vegetation as suggested in prior studies [19,27]. Their values are defined between −1 to 1 in which the negative values indicate no vegetation, while greater positive values indicate available green vegetation cover. However, a given pixel is often mixed on vegetation, water and soil. The LSWI is an alternative useful index to identify water content in vegetation and soil, and its values are also between −1 to 1. The current study found that most vegetation pixels have EVI ≥ 0.174, NDVI ≥ 0.377 and LSWI ≥ 0. The final formulations to identify surface water body and green vegetation in 2020 are shown as below:


#### 2.3.4. Annual Maps of Mangrove Forest

As shown in Figure 3, almost all vegetation pixels have values of VF ≥ 0.5, while water pixels have WF values ranging from 0.05 to 0.85. In addition, both mangrove and non-mangrove pixels have DEM ≤ 5 m and Slope ≤ 5◦. Therefore, the criteria for mangrove mapping in 2020 and in 2022 was described as follows:

$$\text{Mangrove forest} = \text{VF} \ge 0.5 \text{ and } \text{WF} \le 0.2 \text{ and } \text{DEM} \le 5 \text{ m and Slope} \le 5^{\circ} \tag{8}$$

Non-mangrove forest = VF ≤ 0.15 and 0.05 ≤ WF ≤ 0.2 and DEM ≤ 5 m and Slope ≤ 5◦ (9)

2.3.5. Accuracy Assessment

The mangrove maps of the three provinces in 2020 and in 2022 were generated from Landsat-8 OLI and Landsat-9 OLI-2 data using the modified algorithm through the GEE cloud computing platform. In this study, the stratified random sampling approach was employed to generate the verification samples, and very high-resolution images were used to evaluate the accuracy of the classification maps in 2020 and in 2022. The size of verification points for each class (mangrove or non-mangrove) was identified by Cochran's formula (the confidence level was set to 0.95 in this study):

$$\mathbf{n}\_0 = \frac{\mathbf{Z}^2 pq}{\mathbf{e}^2} \tag{10}$$

where

n0 is the sample size,

Z is derived from the standard normal distribution,

e is the desired level of precision,

*p* is the required accuracy, and

*q* = 1 − *p*.

In this study, a total of 474 validation points (20% of the total points) were selected for evaluating the accuracy of mangrove forest mapping in 2020 and in 2022. The random sampling points include 243 mangrove samples and 231 non-mangrove samples. Then, each sample was checked with its location, which was identified from very high spatial resolution Google Earth images by visual interpretation. With the validation samples, the user's accuracy, the producer's accuracy, the overall accuracy, and the Kappa coefficient were calculated in this study [35,40].

#### 2.3.6. Analysis and Statistical Method

The mangrove distribution maps and the mangrove statistical areas in the study sites were automatically computed. In addition, the present study used QGIS 3.10.2 software to produce the spatial distribution of mangroves in the RRD of Vietnam.

#### **3. Results**

#### *3.1. Mangrove Classification and Accuracy Assessment*

As shown in Figure 4, the total area of mangrove forest was estimated as 10,125 ha and 10,630 ha in 2020 and in 2022, respectively. The largest mangrove forest area was observed in Hai Phong province (3790 ha), followed by Nam Dinh province (3325 ha) and Thai Binh province (3010 ha) in 2020.

The results in Tables 2 and 3 show that the overall accuracies obtained from the stratified random sampling points were 94.94% in 2020 and 91.98% in 2022, while the Kappa coefficients of classification maps for 2020 and 2022 were 0.90 and 0.84, respectively, indicating a good-of-fit agreement between the classification result and reference data. The Landsat-8 OLI sensor produced relatively higher accuracy for 2020 than that of the Landsat-9 OLI-2 sensors for 2022.

**Figure 4.** The estimated mangrove area in the RRD from 1990 to 2022.

**Table 2.** The confusion matrix for accuracy assessment of mangrove forest using Landsat 8-OLI in 2020.


**Table 3.** The confusion matrix for accuracy assessment of mangrove forest using Landsat-9 OLI-2 in 2022.


*3.2. Mangrove Dynamics from 1990 to 2022*

By using our defined thresholds in 2020 and the proposed framework in Figure 3, we generated mangrove maps in the three provinces (Hai Phong, Nam Dinh and Thai Binh) in the RRD between 1990 and 2015 together with mangrove maps in 2022 (See Figures A1–A3). We also estimated the areas of mangrove forests in the three provinces for other years (1990, 1995, 2000, 2005, 2010, 2015 and 2022). The mangrove distribution maps and the statistical areas in the study sites were automatically computed using the Java scrips on the GEE cloud computing platform. As shown in Figure 4**,** the mangrove forest area increased in the three provinces across the RRD over the 32 years (1990–2022). The change of mangrove area in each province and each period can be found in Table 4. Figure 5 shows the spatial distribution of mangrove in the RRD of Vietnam in 1990 (Figure 5a) and in 2022 (Figure 5b). Figures 6 and 7 represent the mangrove maps of each province in the RRD in 1990 (Figure 6)

and in 2022 (Figure 7). Mangrove forests are mainly distributed in the river mouth of the three provinces in the RRD, and they are found in front of the sea dykes (Figures 5–7).

As shown in Figure 4, the area of mangrove forests in the RRD significantly increased from 1990 to 2022. The mangrove area decreased from 2960 ha in 1990 to 2408 ha in 1995 and then significantly increased to 4435 ha in 2000. Notably, the area of mangrove forests decreased to 3502 ha in 2005. In contrast, the mangrove area experienced an increase from 4706 ha in 2010 to 8179 ha in 2015 and continued its upward trend to 10,125 ha in 2020 and reached the highest peak value of 10,630 ha in 2022.

Table 4 shows the change detection of the mangrove area in Hai Phong, Thai Binh and Nam Dinh provinces over 32 years. Overall, the mangrove area across the three provinces increased considerably since 2010. Hai Phong province had the largest area of mangrove in 2022 with 3934 ha, followed by Nam Dinh province (3591 ha). In contrast, the mangrove area in Thai Binh province was the lowest with 3195 ha.


**Table 4.** The change detection of the mangrove area in the three provinces over 32 years.

**Figure 5.** The spatial distribution map of mangrove in the RRD of Vietnam: (**a**) mangrove map in 1990; (**b**) mangrove map in 2022.

**Figure 6.** Mangrove maps in the three provinces in 1990 across the RRD, northern Vietnam: (**a**) Thai Binh; (**b**) Hai Phong; (**c**) Nam Dinh.

**Figure 7.** Mangrove maps in the three provinces in 2022 across the RRD, northern Vietnam: (**a**) Thai Binh; (**b**) Hai Phong; (**c**) Nam Dinh.

#### **4. Discussion**

#### *4.1. Uncertainty of Mangrove Mapping and Change Detection*

The overall accuracies (OA) of the mangrove maps in 2020 and in 2022 were 94.94% and 91.98% with Kappa coefficients of 0.90 and 0.84, respectively. These values indicate the successful use of the pixel-based algorithm for mapping mangrove forests and detecting change using multi-temporal Landsat datasets on the GEE cloud computing platform. The Landsat-8 OLI sensor produced better results than those obtained from Landsat-9 OLI-2 (Tables 2 and 3). It is likely due to more available multi-temporal Landat-8 datasets in 2020 with 14 time series scenes compared to only 3 cloud-free scenes available during 2022. The number of available cloud-free time series Landsat data would influence the overall accuracy and produce better results when mapping mangrove forests using the pixel-based algorithm. Future studies applying our framework and thresholds should be further tested in other mangrove regions with more available Landsat-9 OLI-2. Our results suggested satisfactory accuracies for mapping mangrove forests during 2020 and 2022. These results are relatively higher than those reported by the previous studies in Vietnam using SPOT-7 imagery with a higher spatial resolution of 6 m (OA = 92.9%) [41] and using time series Landsat data (OA ranged from 85% to 92%) [42]. Our results are similar to Hauser et al. [43] with an attempt to detect mangrove dynamics on the southern coast of Vietnam using GGE with an overall accuracy ranging from 94 to 96%. However, there is an uncertainty involved in the mangrove classification and change detection. There are several factors that could affect the accuracy of mangrove mapping in the study area during 2020. As shown in Figure A4, this study only obtained about 83% of the pixels in 2020 with seven good-quality observations. Therefore, it can be considered that the acceptable uncertainty [40,41] in mangrove area estimation probably resulted from the lower quality of available Landsat time series (TM/ETM+/OLI/OLI-2) data obtained in the current study area between 1990 and 2022. In addition, the mixed pixels of mangrove and other vegetation in the study area may also affect the accuracy of the generated mangrove maps. For example, *Casuarina* spp. sites were misclassified as mangrove forests because several *Casuarina* spp. species have a quite similar reflectance spectrum with other mangrove species observed in the RRD [30] such as *Sonneratia caseolaris***,** *Kandelia obovata***,** *Aegiceras corniculatum***,** *Rhizophora stylosa* and *Avicennia marina*. Importantly, in the RRD, many mixed small and shrub species are often observed and reported in the previous studies [30,35].

In this study, the defined thresholds were created based on the calibration data collected from the high spatial resolution Google Earth images in 2020 to automatically map and detect mangrove canopy changes across the RRD. As shown in Figure 4, the estimated mangrove area in 2015 was about 8179 ha. This number is close to the estimate as reported in the National Forest Inventory (NFI) in Vietnam during 2015 (8225 ha), fitting well with the model developed in the current study using Landsat data on the GEE.

We observed an increase in the extent of mangroves across the three provinces in the RRD located on the northern coast of Vietnam from 1990 to 2022. The trend is consistent with the forest coverage change in Vietnam, which was reported in recent studies [44,45] and is similar to those observed in other Southeast Asian countries by Goldberg et al. [3] in the southeast and northern Australia by Saintilan et al. [4]. The increase in forest coverage benefited from the efforts of the Vietnamese government in mangrove planting, restoration, and protection. The total forest area in Vietnam was slightly increased between 1990 and 2020 and includes both inland forest and mangrove forest in Vietnam [44]. Overall, the mangrove forest area increase over 32 years (1990–2022) can be automatically detected and mapped by using Landsat 5/7/8/9 time series images through the GEE platform as a result of a number of mangrove restoration projects and programs by the Vietnamese government and policy recommendations based on policy measures from many research studies [11,46].

#### *4.2. Driving Factors for Mangrove Dynamic in Three Provinces from 1990 to 2022*

As exhibited in Table 4 and Figure 4, the mangrove area changed during the period of 1990–2022. Key drivers that caused changes of mangrove forests in each period are considered and discussed as follows:

**Between 1990 and 1995**: The total mangrove area of three provinces decreased from 2960 ha in 1990 to 2408 ha in 1995. This decline was caused by the mangrove deforestation in the Hai Phong province and the Thai Binh province during the period. Specifically, the mangrove area in Hai Phong province declined from 1433 ha to 1190 ha, and in Thai Binh province it reduced from 1068 ha to 442 ha. This period witnessed the smallest mangrove area during the 32-year period. The reasons behind the decrease were the consequence of a new policy, the Reform Policy, initiated in 1986 and officially launched in 1991 [47]. During the period, natural resources, including forest resources, were exhaustively exploited for economic development.

During this period, many regions were converted to aquaculture farms, significantly destroying mangrove forests in Thai Binh and Hai Phong provinces [10]. In contrast, Nam Dinh province had a mangrove area increase of 317 ha from 1990 to 1995 thanks to strict protection and constant expansion of Xuan Thuy National Park [48].

**Period of 1995–2000**: This period witnessed an increase in the mangrove area in three provinces (Table 4). The mangrove areas of Hai Phong, Nam Dinh and Thai Binh provinces in 2000 reached 1495 ha, 1335 ha and 1605 ha, respectively. This increase was due mainly to the efforts of planting and protecting mangrove forests through various programs and projects implemented in such provinces. During this period, the Five Million Hectare Reforestation Program (661 program) was carried out between 1998 and 2010 at the national level to increase forest coverage. The percentage of forest coverage was up to 43% of the total land cover in the final year of the program. In addition, other programs and projects were also implemented. Several projects such as Red Cross of Japan, PAM5325, ACTMANG, the 661 programs (Mangrove Plantation and Disaster Risk Reduction project) were undertaken in such provinces. These projects significantly contributed to the increase of mangrove cover in the RRD [49,50].

**Period of 2000–2005:** In this period, the mangrove areas decreased from 4435 ha in 2000 to 3502 ha in 2005 (Figure 4). The main cause for mangrove loss in 2005 may probably be explained by the negative impacts of natural disasters. In 2005, an extreme typhoon event, typhoon "Damrey", hit the northern region of Vietnam [46] and damaged the mangrove forest in these areas, especially young mangrove forests. This typhoon was also reported by Hong, Avtar and Fujii [9] as the amongst the strongest tropical cyclones impacting the coastal zones of Vietnam during the last 30 years.

**Period 2005–2020:** This period witnessed a continuous increase in the area of mangroves in such provinces sustained for 15 years. As shown in Table 4, the total area of mangrove reached 10,125 ha in 2020. This number was three-times higher than that in 2005. The mangrove restoration received priority attention and investment by the Vietnamese government in this period and enhancement of community-based mangrove management [51,52]. In addition to the 661 Program implemented from 1998 to 2010, many other projects and programs funded by the Vietnamese government and other organizations were implemented in the whole country [52], especially in the Red River Delta [53]. Further sustainable mangrove conservation and management across the Vietnamese coastline should be carefully considered in protecting existed mangrove forests and restoring degraded mangroves as well as planting new ones to enhance not only the mangrove area but also quality and biodiversity in the context of climate change issues.

**Period 2020–2022:** This short period was characterised by an increased upward trend in mangrove area in the RRD. The Vietnamese government continued to support mangrove conservation and management schemes in dealing with climate change impact.

#### **5. Conclusions**

Mangrove forests play an important role in mitigating climate change impacts and are able to sequester blue carbon for their protection and restoration. Mapping mangrove extent at a large scale remains challenging due to cloud coverage and spatial limitations of single satellite sensors. This study developed a framework using the pixel-based algorithm applied to Landsat TM/ETM+/OLI/OLI-2 time series data on the Google Earth Engine cloud computing platform to automatically map and quantify mangrove forest changes in the three provinces of Hai Phong, Nam Dinh and Thai Binh across the RRD over 32 years.

The results showed that the mangrove area has increased considerably in the RRD over 32 years in response to the mangrove restoration programs and policies by the Vietnamese government and local governments. The mangrove areas were 2960 ha, 2408 ha, 4435 ha, 3502 ha, 4706 ha, 8179 ha, 10,125 ha and 10,630 ha in 1990, 1995, 2000, 2005, 2010, 2015, 2020 and 2022, respectively.

The overall accuracies of the Landsat-8 OLI and the Landsat-9 OLI-2 image processing for 2020 and 2022 were 94.94% and 91.98%, respectively, while the Kappa coefficients were 0.90 and 0.84, indicating promising results for mapping mangrove forest cover in the tropics using the GEE platform associated with free open-source code. It could be said that the pixel-based algorithm and Landsat time series images on the GEE cloud computing are suitable for long-term monitoring of mangrove change in tropical regions. The Landsat family has shown the potential use in mapping mangrove dynamics in the tropics and should be further used worldwide.

**Author Contributions:** Conceptualization, T.T.P.V., T.D.P. and B.M.; methodology, T.T.P.V., T.D.P., H.V.L. and Q.H.V.; software, T.T.P.V. and H.Q.N.; validation, T.T.P.V., Q.H.V. and T.D.P.; formal analysis, T.T.P.V. and H.Q.N.; investigation, T.T.P.V.; resources, T.T.P.V. and N.N.L.; data curation, T.T.P.V.; writing—original draft preparation, T.T.P.V. and T.D.P.; writing—review and editing, T.D.P., N.S., N.N.L. and A.S.; visualization, T.D.P. and H.V.L.; supervision, B.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** The authors are thankful to the JICE (Japanese International Cooperation Center) for financial support to the first author studying a master course at the University of Tsukuba, Japan for 2 years. In addition, we also would like to thank the Forest Inventory and Planning Institute (FIPI), Vietnam for providing useful data for the current study. T.D.P. is supported by a Macquarie University Research Fellowship (Grant No. MQRF0001124-2021).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Appendix A**

**Figure A1.** Mangrove distribution maps in Hai Phong between 1995 and 2020.

**Figure A2.** Mangrove distribution maps in Thai Binh between 1995 and 2020.

**Figure A3.** Mangrove distribution maps in Nam Dinh between 1995 and 2020.

**Figure A4.** The number of pixels with good-quality observations in 2020.

#### **References**


### *Review* **A Review of Spectral Indices for Mangrove Remote Sensing**

**Thuong V. Tran 1,2,\*, Ruth Reef <sup>1</sup> and Xuan Zhu <sup>1</sup>**


**Abstract:** Mangrove ecosystems provide critical goods and ecosystem services to coastal communities and contribute to climate change mitigation. Over four decades, remote sensing has proved its usefulness in monitoring mangrove ecosystems on a broad scale, over time, and at a lower cost than field observation. The increasing use of spectral indices has led to an expansion of the geographical context of mangrove studies from local-scale studies to intercontinental and global analyses over the past 20 years. In remote sensing, numerous spectral indices derived from multiple spectral bands of remotely sensed data have been developed and used for multiple studies on mangroves. In this paper, we review the range of spectral indices produced and utilised in mangrove remote sensing between 1996 and 2021. Our findings reveal that spectral indices have been used for a variety of mangrove aspects but excluded identification of mangrove species. The included aspects are mangrove extent, distribution, mangrove above ground parameters (e.g., carbon density, biomass, canopy height, and estimations of LAI), and changes to the aforementioned aspects over time. Normalised Difference Vegetation Index (NDVI) was found to be the most widely applied index in mangroves, used in 82% of the studies reviewed, followed by the Enhanced Vegetation Index (EVI) used in 28% of the studies. Development and application of potential indices for mangrove cover characterisation has increased (currently 6 indices are published), but NDVI remains the most popular index for mangrove remote sensing. Ultimately, we identify the limitations and gaps of current studies and suggest some future directions under the topic of spectral index application in connection to time series imagery and the fusion of optical sensors for mangrove studies in the digital era.

**Keywords:** vegetation index; mangrove index; mangrove forest; mangrove above ground; biomass; carbon sink; bibliometric analysis

#### **1. Introduction**

Mangrove is a term which corresponds to intertidal ecosystems or lignified plant communities that grow in coastal environments between 40◦S and 30◦N throughout the world (Figure 1). The mangrove boundary is extended to the south of Japan (30.4◦N) and Bermuda (32.4◦N); to the south of New Zealand (38.05◦S), Australia (38.85◦S), and the east coast of South Africa (32.98◦S) [1,2]. Mangrove distribution is restricted generally to areas where the mean temperature ranges 20–35◦C, annual rainfall is between 1500–2500 mm, and there is a substantial riverine input of freshwater discharge [2]. Actually, the number of frozen days in the year may play on mangrove presence at high latitudes [3]. Decreases in the frequency of extreme cold occurrences could lead to considerable increases in mangrove cover near the current poleward limits of mangrove forests. The global mangrove distribution is classified into two groups, including the Indo-West Pacific (IWP) and the Atlantic East Pacific (AEP). Mangroves initially developed on the Tethys Sea's coastlines in the late Cretaceous-early Tertiary period [1,2,4]. Three million years ago, modern mangrove taxa emerged on the eastern borders of Tethys, diversified into present-day IWP regions, and subsequently spread into AEP regions [5,6]. The richness in the distribution of mangrove species reduces from the IWP to AEP. Globally, there are approximately 77 mangrove species, but about 54 species in 20 genera from 16 families constitute the group of "true

**Citation:** Tran, T.V.; Reef, R.; Zhu, X. A Review of Spectral Indices for Mangrove Remote Sensing. *Remote Sens.* **2022**, *14*, 4868. https:// doi.org/10.3390/rs14194868

Academic Editor: Chandra Giri

Received: 12 August 2022 Accepted: 26 September 2022 Published: 29 September 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

mangroves" occurring only in mangrove habitats. Among these 77 species, 65 species in 32 genera and 24 families are recorded from IWP, while only 15 species in 10 genera and 8 families are in AEP [1,2]. The most commonly found genera in both IWP and AEP are *Rhizophora* and *Avicennia* [2].

The biophysical variables of mangroves (i.e., leaf area, basal area, tree height, percent canopy closure, diameter at breast height, carbon stock, and biomass) mostly depend on climatic conditions, while sea level rise has an influence on the structure and spatial distribution of mangroves [7,8]. Temperature, precipitation, and storminess explain 74% of the global trends in the maximum values of canopy height and above-ground biomass [7]. Globally, 75% of mangroves are distributed in tropical regions. The largest cover and highest mangrove diversity are found in Asia (39%), followed by Africa (21%, but mostly on the eastern side), North and Central America (15%), South America (12.6%) and Oceania (Australia, Papua New Guinea, New Zealand, south Pacific islands) (12.4%) [9,10]. The mangroves in the equatorial regions have the maximum biomass and tree canopies can reach an average height of 30–40 m. The highest mangrove forests are found in Gabon, an equatorial African nation, where heights reach 62.8 m. A quarter of the estimated 11.7 Pg C global mangrove carbon stock, which includes soil, above- and below-ground biomass, is stored in Indonesia [7,8]. These biophysical parameters gradually decrease with increasing latitude due to varied temperature and environmental conditions.

**Figure 1.** (**a**) The distribution of mangroves (in pale green) in the world, (**b**) in the Caribbean, and (**c**) in South and Southeast Asia (Source: adapted from Global Mangrove Watch, [11,12]).

Mangrove ecosystems produce valuable goods and services, including regulation (e.g., coastal protection, water filtration), supply (e.g., fisheries, aquaculture, timber, fuel, honey, construction materials, medicine), culture (e.g., tourism), and support (e.g., nursery habitats, climate mitigation) [13,14]. In some areas, mangroves have been proposed to provide a natural barrier to coastal erosion process, defending inland areas home to 120 million people from natural hazards (e.g., typhoon, cyclones, tsunamis) [7,15]. Mangrove

restoration for coastal defence is expected to be up to five times more cost-effective than "grey infrastructure" such as breakwaters [16]. Mangroves also regulate water quality, and it is estimated that 2–5 ha of mangroves can treat the effluent from 1 ha of some aquaculture practices [17]. The carbon storage potential of mangroves is 3–5 times higher than that of tropical upland forests due to strong carbon storage in the soil [18,19]. Mangroves are also a valued source of timber, fuel, and tourism. There are over 2000 mangrove related attractions globally, such as boat tours, boardwalks, kayaking, and fishing [20]. Together, the economic value from mangrove ecosystem services has been estimated to exceed 800 billion per year [15,21]. However, approximately 35% of the global mangrove forests has been lost over the past 50 years due to both anthropogenic activities and physical stressors [22,23]. While restoration of mangrove forests has been increasing over the past 40 years, the net reduction in mangrove cover area and species richness is still a high 1–2% per year [22,24].

Traditionally, monitoring mangrove ecosystems used field observation and survey methods [25–27]. However, these approaches are difficult to monitor and measure mangroves in situ due to their dense understory and geographical location in intertidal zones [21,25]. Additionally, field observation and survey methods are labour-intensive, costly, and frequently limited in extent. Many surveys are qualitative and difficult to reproduce or revisit over time. Remote sensing (RS) has overcome the drawbacks of traditional field surveys and is has been continuously improving in terms of spatial resolution, revisit time and user costs over the past four decades [25,28]. RS is acknowledged in this context as the science and technology of acquisition of information about Earth's surface materials from a distance, typically from aircrafts or satellites [29]. The two types of remote sensing we refer to are (i) optical and (ii) radar sensors, which are classified according to the energy source of the signal used to identify the object. The remotely sensed data, acquired from these sensors, allows us to gather accurate information about the geographical distribution of mangrove ecosystems and biophysical properties at the pixel level [13,27].

In remote sensing, mangroves can be identified based on the textural and spectral properties of the canopy and leaves [13,30]. Their structural appearance, which can be either homogenous across the forest or heterogeneous, is affected by factors including species composition, growth form, density, and stand height. Almost all mangrove species can be discriminated within the visible and near infrared (NIR) region because of scattering in the spongy mesophyll cells in plants [31,32]. Using structural information extracted from several remotely sensed products regional and global estimation can be made of mangrove height, canopy, species succession, biomass, and carbon stocks [26]. The highest spectral reflectance of mangroves was observed in the NIR region for both Landsat 8 and Sentinel-2A surface reflectance sources (Figure 2). With the Sentinel-2A in particular, mangrove reflectance was observed to rise rapidly at the red-edge. Therefore, mangrove ecosystems can be observed using indices computed from spectral bands in the visible and NIR regions of optical remote sensing.

**Figure 2.** Spectral signatures of mangroves (*Rhizophora*) in Can Gio Mangrove Biosphere Reserve, Vietnam, derived from median values of (**a**) Landsat 8 and (**b**) Sentinel-2A surface reflectance in February 2021 (Source: obtained from Google Earth Engine).

The importance of remote sensing in mangrove studies has been recognised in many review studies [25,26,28,30,33–35]. These publications serve as a good starting point for researchers who are interested in mangrove remote sensing. However, the application of different spectral indices in mangroves has not been reviewed extensively in most of these studies. For instance, Green et al. (1998) [33] considered the significance of remote sensing for mangrove mapping from 1972 to 1996. While the study is the first paper that mentioned applying NDVI to mangrove classification, it only focused on NDVI even though at the time there were over 40 vegetation indices that could have been relevant to mangrove ecosystems. During the 1998–2018 period, most review papers highlighted remote sensing as a technique or approach for mangrove studies, while remote sensing has been defined as the science of acquiring information from distance [29,36,37]. Recently, Wang et al. (2019) [25] revealed common gaps in previous publications (i.e., research topics, key milestones, and mangrove driving forces) in mangrove remote sensing and investigated the importance of remote sensing for mangrove studies from 1956 to 2018. However, Wang et al. (2019) did not clearly state what kinds of spectral indices are specifically applied to mangrove remote sensing.

The present study intends to address the aforementioned knowledge gaps, by answering the following research question: what spectral indices have been applied and have been proven effective for mangrove remote sensing? Our objectives are to (i) examine and categorise spectral indices used in publications related to mangrove remote sensing; (ii) assess their applications in the study of mangrove ecosystems; and (iii) propose future directions for the application of additional spectral indices in mangrove remote sensing.

#### **2. Search Strategy and Data Analysis**

Mangrove scholars used various qualitative and quantitative approaches to understand and organise earlier findings of mangrove studies. Among these, a quantitative analysis of academic literature, defined as bibliometrics, was investigated as a potential tool to introduce a systematic, transparent, and reproducible review process [38–41]. Compared to other literature review techniques, bibliometric analysis of the published literature is effective to identify research gaps and direct future avenues of research [41]. For bibliographic citations, Web of Science (WoS) and the Scopus platforms are the most extensive databases which are widely used to obtain metadata for bibliometric analysis [42,43]. Scopus was launched in 2004, but WoS launched in 1997 and is considered the earliest international bibliographic database [42,44]. WoS comprises four citation databases with more than 10,000 journals [45]. Journals indexed in the WoS must meet 28 criteria (i.e., 24 quality criteria and 4 impact criteria) [44,46]. The fulfilment of 28 criteria contributes to enhancing academic quality and minimising the influence of multiple predatory journals. The journal listed in the WoS database primarily provides impact factor (a ratio between citations and citable items published the previous year) and h-index (an index based on a list of publications ranked in descending order by Times Cited count) [46]. Therefore, journals with high impact factor or h-index are cited more often than journals with lower impact factor or h-index.

Various keywords were entered in the searching process associated with global mangrove ecosystems based on spectral indices application (Table 1). We used Thomson Data Analyzer (TDA) integrated in the Web of Science (WoS) Core Collection to retrieve annual publications and their citations [47]. The keyword search resulted in 293 papers published between 1992 and 2021. We then reviewed the abstract and content of the 293 papers and removed from our study the papers that did not relate to the application of spectral indices in mangroves. This left 195 publications (including 90 journal papers, 14 conference papers and one book chapter) for our review.


**Table 1.** The predefined keywords of the searching process.

Full records (i.e., author, title, source, abstract) and cited references of the search results were downloaded in the BibTex format in several batches, each comprising no more than 500 data entries. For further processing, all of the obtained result files were zipped together and imported into the R-statistical and VOSviewer software packages. The bibliometric analysis was carried out with the help of Bibliometrix package in R [38]. The annual cited times of the gathered articles were calculated using TDA's citation report tool, while the top journal sources for publication and citation were determined using Bibliometrix package. Finally, using VOSViewer software [48], we performed co-word analysis to visualise density networks of author keywords for trend analysis in mangrove remote sensing. The co-word method is a technique to analyse the co-occurrences of key words and identify relationships and interactions between the topics researched and emerging research trends [48]. Details regarding the theory and practical function of the co-word approach utilising VOSViewer software may be found at [45,48]. However, the record of the online bibliographic database prior to 1990 may be incomplete [42,44,49] because the internet-based Web of Science was firstly launched in 1997. Therefore, several publications from print only sources may be missing. We thus supplemented our database with metadata from outcomes of Green et al. (1998) [33] and Bannari et al. (1995) [50] to increase the number of studies applying spectral indices in mangrove studies prior to 1996.

Publications and annual citations in this field of research, retrieved from citation report analytics of the Web of Science Core Collection, significantly increased between 1996 and 2021 (Figure 3). The number of publications dropped in 2021, which may be related to COVID-19 pandemic because most field trips were delayed or cancelled and this impacted field data collection for validation and other known impacts on academic workloads [51]. Most of the publications (21) were published in the journal of Remote Sensing (IF: 4.848, Open access), others were mainly from the International Journal of Applied Earth Observation and Geoinformation (8, IF: 5.993, Open access), International Journal of Remote Sensing (8, IF: 3.362), Estuarine, Coastal and Shelf Science (7, IF: 2.904), Modeling Earth Systems and Environment (6, IF: N/A) and ISPRS Journal of Photogrammetry and Remote Sensing (5, IF: 10.565). The journals Ecological Indicators, Regional Studies in Marine Science, and Wetlands had four papers each. The remaining journals had published less than three papers involving mangrove remote sensing since 1996. In-text citations of work that applied mangrove remote sensing indices were found in Remote Sensing of Environment (881, h-index: 281), Remote sensing (445, h-index: 124), International Journal of Remote Sensing (437, h-index: 174), Estuarine, Coastal and Shelf Science (249, h-index: 134), Aquatic Botany (170, h-index: 94). The publications most cited were those in higher ranked (higher IF) journals, regardless of the open access status of the journal.

**Figure 3.** Change in 195 peer-reviewed publications and their citations per year between 1996 and 2021 (based on the citation report tool in TDA using mentioned keywords).

#### **3. Overview of Spectral Indices Used in Mangrove Remote Sensing**

A spectral index is an equation that combines pixel values from two or more spectral bands in a multispectral image using various algorithms, mainly focused on band ratio or feature scaling (e.g., normalised or standardised algorithms) [52,53]. Spectral indices are calculated to highlight pixels in an image that not only show the relative abundance of a land cover of interest, but also emphasise an ecosystem function [52–54]. They show better sensitivity than individual spectral bands for spectral signature detection. Throughout the mission of Earth surface observation, spectral indices have significantly contributed a more thorough understanding of environments and ecosystems across space and time [28]. The geographical extent of mangrove studies using spectral indices has also seen a significant change over time (Table 2). Most studies (92%) were carried out at the national level, while a few publications implemented research on intercontinental/global scale during the 2000–2015 period and this trend is increasing (16% in 2021 vs. 8% in 2016). The study areas were mainly in India (12.8%), Mainland China (12.3%), Indonesia (9.7%), the US (9.7%), and Mexico (8.2%).

**Table 2.** The mangrove study area of spectral indices application, obtained from our searched publications.


Spectral indices can be categorised as either satellite or airborne system indices based on the platforms used for data acquisition. Depending on the spectral bands of passive satellite remote sensing, spectral indices may be further grouped into indices with (i) simple ratio, (ii) visible and near-infrared (VNIR) bands, (iii) visible and red edge bands, (iv) visible and mid-infrared bands, and (v) visible and shortwave infrared (SWIR) bands. In addition, following the applications of spectral indices in mangrove remote sensing, we separate spectral vegetation indices and spectral mangrove-specific indices. The key differences between the two types of indices are their applications and the spectral bands used in the indices. Vegetation indices are spectral indices computed using spectral bands in the visible, red edge, and near-infrared regions [55]. They have been widely applied to mangrove ecosystems previously [13,30]. However, as mangroves have common spectral characteristics as other vegetation, separating mangroves from other types of vegetation using a single vegetation index is challenging [56]. To address this issue, some mangrovespecific spectral indices have been proposed for separating mangroves from terrestrial vegetation [56–60]. These spectral mangrove indices include spectral bands from VNIR to SWIR regions. In this review, based on publications that used spectral indices for mangrove ecosystems and the approaches they applied, we classified indices into four categories: (i) visible and near-infrared bands; (ii) visible and red edge bands; (iii) visible bands of airborne systems; and (iv) spectral mangrove specific- indices.

#### *3.1. Spectral Indices with Visible and Near-Infrared Bands (VNIR)*

Indices measured from spectral bands in the VNIR regions are acknowledged as vegetation indices [50], except Normalized Difference Water Index (NDWI) [61]. The history of spectral vegetation indices development is associated with the Landsat mission in 1972. Particularly, Pearson and Miller (1972) developed the first vegetation indices, i.e., Ratio Vegetation Index (RVI) and Vegetation Index Number (VIN), to estimate and monitor vegetative cover [50,62]. Following Pearson and Miller (1972), Rouse et al. (1973) introduced the Normalized Difference Vegetation Index (NDVI) that is now widely applied for land cover and environmental studies [63,64]. Over the last four decades, more than 40 vegetation indices have been developed, but 28 of these were used in mangrove investigations (Table 3). Two categories of vegetation indices can be separated: ones that include only spectral bands and others that include spectral bands that are adjusted by non-spectral factors [e.g., soil adjustment factor (*L*), soil line factors (*a*, *b*), and coefficients of atmosphere resistance (*c*<sup>1</sup> and *c*2)]. The first category (e.g., RVI, VIN, NDVI–Table 3) is based on linear combinations (difference or sum) of spectral bands or raw band ratios without considering environmental interactions. The second group (e.g., SAVI, TSAVI, EVI–Table 3) is based on the knowledge of physical phenomena which explains interactions between electromagnetic radiation, the atmosphere, the vegetative cover, and the soil background.

In the first group without adjustment factors, NDVI is the first index to show the highest correlation with field measured mangrove canopy cover (*r* = 0.91), higher than PVI, GVI, and RVI [65]. NDVI (No.3–Table 3) is a simple indicator that is acquired in red (visible) and near-infrared (NIR) regions, based on a normalised algorithm. The NIR band in the NDVI equation is useful for vegetation detection because healthy vegetation (which contains chlorophyll) reflects more NIR compared to other wavelengths [63]. The normalised algorithm mitigates the atmospheric effects and the impacts of sensor calibration degradation in the red and NIR bands [64,66]. Mathematically, NDVI also forms the basis for other indices. For example, a bijective relationship between NDVI and VIN is demonstrated by Equation (1). However, the NDVI values are affected by soil background when the green leaf area is small or the majority of the scene is soil [67,68]. Therefore, a number of vegetation indices in the second category have been developed for taking into account environmental conditions. For example, to adjust soil background, soil factors (e.g., *L*, *a*, and *b*) were included in SAVI and TSAVI (No.8–9, Table 3). Both indices SAVI and TSAVI are equal to NDVI, if the value of the soil adjustment factor in SAVI (No. 8, Table 3) is zero (*L* = 0) or the slope (*a*) and ordinate (*b*) at the origin of bare soil line parameters in TSAVI (No.9, Table 3) are one and zero (*a* = 1 and *b* = 0), respectively.

$$NDVI = \frac{NIR - R}{NIR + R} = \frac{NIR/R - 1}{NIR/R + 1} = \frac{VN - 1}{NIR + 1} \tag{1}$$

Some studies show that the indices that include adjustment factors are less sensitive to atmospheric and soil background effects than the first-generation spectral indices [50,64,65]. Enhanced Vegetation Index (EVI, No.4–Table 3) [69], for example, which corrects for some atmospheric conditions and canopy background noise, and provides a better estimation of mangrove biophysical properties in high density forests than does NDVI [70]. However, the blue band requirement may be a disadvantage of EVI, which cannot be generated from optical sensors that do not have a blue band, such as the Advanced Very High-Resolution Radiometer (AVHRR) and the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER). Alternatively, the EVI-2 (No.28–Table 3) with two spectral bands in the red and NIR regions was proposed, achieving similar performance to EVI, which can be produced by almost all optical sensors [71]. However, the EVI-2 is sensitive to the impact of the bidirectional reflectance distribution function (BRDF). Therefore, each index has advantages and disadvantages in relation to vegetation characteristics with overlapping spectral features due to background signals from soil and confounding factors (e.g., sensor and calibration effects, quality assurance and quality control, BRDF, and atmospheric and topographic effects). Hence, there are no perfect vegetation indices for all aspects of mangrove studies under all conditions.

**Table 3.** The 28 indices in the spectral wavelength 400–1000 nm used in mangrove research (R: Red; G: Green; B: Blue; and NIR: Near Infrared). For whole algorithms, *L* is a soil adjustment factor that addresses nonlinear model, differential NIR and red radiant transfer through a canopy [72]; *a* and *b* are soil line factors [73]; *c*<sup>1</sup> and *c*<sup>2</sup> are coefficients of atmosphere resistance term which use blue band to correct for aerosol influences in red band [69].



#### **Table 3.** *Cont.*

#### *3.2. Spectral Indices in Visible and Red-Edge Bands*

The red edge is a band in the red-NIR transition zone that indicates the transition between red visible absorption by chlorophyll and NIR scattering due to leaf internal structure [25]. This transition zone is used in various vegetation indices, the most important of which is the normalised difference between red visible (0.6 nm) and NIR (0.8 nm) reflectance. Spectral indices obtained from the red edge region (Table 4) were used in several investigations for mangrove chlorophyll and biophysical parameters [81,119], mangrove density and carbon analyses [85,247], and mangrove biomass [207]. These investigations found strong correlation (*r* > 0.95) between ground truth data and red edge vegetation indices. Normally, red edge bands are available from hyperspectral datasets that are difficult for large-scale collection. Alternatively, since 2015, three red-edge bands have been incorporated in the Sentinel-2A sensor, which could be combined with other visible bands to provide essential information about mangrove ecosystems at a spatial resolution of 10 m. Therefore, the use of red-edge bands promises the potential to benefit mangrove ecosystems associated with mangrove health monitoring in the future.

**Table 4.** The used indices with VNIR and red-edge bands (R: Red; B: Blue; and NIR: Near Infrared). With three Red-edge bands of Sentinel-2A, each formula can compute three sub-equations.


#### *3.3. Spectral Indices with Visible Bands of Airborne Systems*

Traditionally, aerial photographs (AP) were widely used for mangrove mapping and assessment [94]. Nowadays, with the support of Unmanned Aerial Vehicles (UAVs) or drones, the application of aerial photographs has become more convenient. Some spectral indices for mangrove classification have been proposed using visible bands of aerial photographs that can achieve an overall accuracy of over 95% in mangrove cover mapping [157,254] (Table 5). Notably, integration of Red-edge, NIR, and shortwave infrared (SWIR) bands in the optical sensors of UAVs promises an advantage for monitoring mangrove ecosystems at a spatial resolution of centimetres. This drone-based multispectral remote sensing can be as the future for mangrove remote sensing. However, these applications are only suitable for a small area and can be employed in a short time due to cost and energy limitation. More importantly, the use of UAVs for field data collection is regulated by the government in most countries throughout the world.

**Table 5.** The used UAV spectral indices for mangrove ecosystems.


#### *3.4. Mangrove-Specific Spectral Indices*

Mangroves are found in coastal wetlands where they are regularly submerged by tides [55,60,263]. As a result, fluctuations in tide levels, and thus the presence of water, results in tide-dependent variation in spectral signatures for mangrove forests, leading to inaccurate mapping results, particularly in locations with large tidal ranges [57,135]. Recently, several studies have developed specific spectral indices that can adapt to changes in tide conditions and applied them in order to separate mangroves from non-mangroves [57,58,60] and from other land cover types [55]. These are the mangrove-specific indices. During the 2013–2021 period, six mangrove-specific indices were proposed to improve the accuracy of single image remotely sensed data during high tide (Table 6).

Zhang and Tian (2013) [57] proposed a mangrove recognition index (MRI–Equation (2)) for mangrove detection using multi-temporal Landsat TM images that is insensitive to the stage of the tide. Winarso et al. (2014) [56] further developed this method to create a mangrove discrimination index (MDI–Equation (3)) for estimate mangrove density from Landsat 8 images. Kumar et al. (2017) [59] proposed two new vegetation indices (Normalised Difference Wetland Vegetation Index and Shortwave Infrared Absorption Index) and combined them with two previously published indices (Normalised Difference Infrared Index and Atmospherically Corrected Vegetation Index) to integrate the Mangrove Probability Vegetation Index (MPVI–Equation (4)) for mangrove classification. Gupta et al. (2018) [58] published a Combined Mangrove Recognition Index (CMRI–Equation (5)) to distinguish mangrove from non-mangrove with an accuracy of more than 60%. Jia et al. (2019) [60] generated the Mangrove Forest Index (MFI–Equation (6)) using spectral bands from Sentinel-2 data to detect submerged mangrove forests at high tide.

**Table 6.** The proposed mangrove indices for mangrove classification (*EQN: equation number*). In the Equation (2), *GVI* and *WI* are green vegetation index and wetness index at low (*L*) tide and high (*H*) tide, respectively. In Equation (4), *n* is the total number of bands in the image, *Ri* is the reflectance value at band *i* for a pixel of the reflectance image, and *ri* is the reflectance value at band *i* for candidate spectrum of mangrove forest. In Equation (6), the *ρλ* is the reflectance of the band centre of *λ*, and *i* ranged from 1 to 4; *λ*1, *λ*2, *λ*3, *λ*<sup>4</sup> represent the centre wavelengths at 705, 740, 783 and 865 nm, respectively. *λ<sup>i</sup>* is the baseline reflectance in *λi*. ρ665 and ρ2190 are the reflectance of band 4 (centred at 665 nm) and 12 (centred at 2190 nm), respectively. In Equations (3), (5), and (7), G, R, NIR, and SWIR are green, red, near-infrared, and shortwave infrared bands, respectively.


These proposed mangrove indices involve the signature of mangroves in the context of tidal fluctuation, which is sensitive to greenness and wetness patterns. Thus, MRI and MPVI are sensitive to tidal extent and period, and cannot be used in site comparisons where sites differ in hydrology. Additionally, the number of spectral bands required for the MFI calculation is only available using Sentinel-2 or hyperspectral sensors. Baloloy et al. (2020) [55] recently analysed the shortcomings of earlier integrated mangrove forest indices (i.e., MRI, CMRI, MPVI, NDI, and MFI) and developed a new index: the mangrove vegetation index (MVI–Equation (7)), for enhancing the accuracy of mangrove forest extent mapping. MVI is a single index that classifies mangroves, terrestrial vegetation (forest and non-forest), bare soil, built-up areas, water, and clouds using reflectance data from Sentinel-2A and Landsat-8 in the NIR, Green, and SWIR bands. MVI validation was initially used at an intercontinental scale and demonstrated an accuracy of more than 80% for the entire set of geographical research locations. The high index accuracy of MVI can provide a possibility for global mangrove studies, although MVI is limited by biophysical and environmental parameters due to its reliance on SWIR. Using the SWIR spectrum, in particular, has been a problem for sensor systems constructed with solely visible and NIR wavelengths (e.g., Landsat-1,4 and Planetscope). Additionally, SWIR reflectance value is frequently mixed with built-up land noise, water bodies, and vegetation surrounding [61,264,265]. Neri et al. (2021) [266] investigated misclassification of mangroves from other land cover types in aquaculture zones, irrigated croplands, and palm tree sites when applying MVI due to spectral similarity between mangroves and vegetation in these areas. Notably, the significant drawback of MVI is that it does not have a specific optimal threshold for mangroves, which differs from ranges of vegetation and other mangrove-specific indices for mangrove separation.

In summary, there are several newly improved spectral indices for mangrove classification, but none of these can completely reduce the impact of environmental factors (e.g., tidal influences, land cover mixture).

#### **4. Evaluation of Spectral Indices Applications in Mangrove Remote Sensing**

As mentioned in Section 3, a variety of spectral indices have been used for (i) mapping mangrove extents and distributions; (ii) measuring above-ground properties of mangroves; and (iii) detecting mangrove changes. The co-word map of keywords that resulted from the

co-word analysis of the literature on spectral indices application for mangroves is presented in Figure 4. The mangrove and forest were core words of the network, and NDVI and remote sensing were prominent terms in a field study. This showed that NDVI has a strong relationship with aspects of mangrove remote sensing in terms of spectral index application. Also, the result predicted that NDVI was the most common index applied to mangrove ecosystems. In total, 82% of our reviewed publications used NDVI and 28% used EVI.

**Figure 4.** Co-word keywords map of spectral indices' application in mangrove research. Colours indicated the density of keywords or terms, ranging from blue (lowest density) to red (highest density). The larger the number of items in the vicinity of a point and the higher the weights of the neighbouring items, the closer the point is to being red.

#### *4.1. Mangrove Extent and Distribution*

Mapping mangroves is important for recording the present mangrove area and species distribution, provides information to support management decisions around potential threats of degradation due to uncontrolled development, and enables modelling change in relation to driving factors. Mangrove mapping can be binary, which classifies an image into two classes: mangroves and non-mangroves, or into a number of land covers based on a particular land over classification scheme with mangrove as one of the land cover types. Binary mapping normally uses single spectral indices (e.g., NDVI, EVI, SAVI, or mangrove indices) to highlight the pixels of mangrove vegetation and separate them from non-mangrove pixels [58]. Mapping can also include continuous classifiers such as height and canopy area.

Image classification for mangrove mapping can be supervised or unsupervised, pixelbased or object-based. Almost all previous research applied composited bands of satellite images based on classification approaches such as unsupervised algorithms (e.g., ISODATA or K-means) and parametric supervised algorithms (e.g., maximum likelihood) to detect mangrove extents without spectral indices application [25,26,267]. The overall accuracy of post-classification from these approaches ranges from 42% to 68% for most optical sensors and methodologies to measure accuracy [26]. Since spectral indices were used

for mangrove ecosystems under pixel-based or object-based categories based on machine learning algorithms (e.g., artificial neural network, support vector machine, and random forest), an improvement of post-classification overall accuracy has been achieved to more than 80% [26]. More recent studies have used mangrove classification using spectral indices based on random forest, one of the machine learning algorithms, and offer the highest post-classification overall accuracy (>92%) [27,196,268].

From using single spectral indices for mangrove separation, some studies found that in dense mangrove areas, the NDVI or EVI value threshold for discriminating mangroves was 0.3 and above [104,269]. Le et al. (2020) applied NDVI, derived from Sentinel-2, to detect mangrove cover in the Can Gio Mangrove Biosphere Reserve, Vietnam [183]. This study considered that the NDVI mangrove value as NDVI > 0.3 with an overall classification accuracy of 83%. In addition to NDVI/EVI, six mangrove spectral indices (Table 6) have been developed and applied for mangrove separation with an overall classification accuracy of 80% above [55].

Vegetation, soil, and water are the three principal factors that contribute to the pixel composition of remotely sensed data in mangroves (Figure 5). In addition, seasonal and diurnal intertidal interactions influence the surface appearance [13,263]. These factors have a significant impact on the spectral characterisation of picture components. Therefore, depending on the effects of natural surroundings and mangrove density, the NDVI/EVI thresholds can be adjusted. In addition to physical influences, mangroves and other types of vegetation may generate similar signals from the vegetation index [56]. As a result, utilising a single NDVI or EVI threshold to distinguish mangroves from other types of vegetation may lead to an inaccurate outcome. Therefore, several soil and water spectral indices (Table 7) have been used concurrent with vegetation indices to improve mangrove detection [150,198].

**Figure 5.** Mangroves in Victoria State, Australia (Source: photo taken by author).


**Table 7.** The indices in the Visible, Near Infrared (NIR), and Shortwave Infrared (SWIR) bands.

The reflected signals from the SWIR regions capture information on radiation absorption by water, cellulose and lignin, and a variety of other biological elements. Nevertheless, collecting satellite imagery at SWIR wavelengths has distinct advantages, such as better atmospheric penetration and better contrasts among different vegetation types. The utilisation of the SWIR band in conjunction with the visible and NIR bands, in particular, aids in enhancing the presence of water in plant leaves [264] or urban characteristics [265]. For example, NDWI and NDBI threshold values more than 0 visualise water bodies and impervious surface, respectively [61,264,265]. Besides, several studies used elevation data and tasselled cap transformation to further improve the classification accuracy [55,57,134]. Consequently, an improved performance (≥90% of overall accuracy) was investigated when apply multiple spectral indices for detecting the mangrove cover [27,267].

Our search did not reveal a spectral index that has been applied for mangrove species separation. Previous studies applied spectral bands to separate mangrove species based on maximum likelihood classification or machine learning algorithms (e.g., random forest and support vector machine) because each mangrove specie reflects a particular wavelength of the spectrum [25–27,164,273]. These studies revealed that spectral reflectance properties of some mangrove species are similar, making a challenge for identification. Hirata et al. (2014) [273] proved that the spectral reflectance properties for *A. alba* and *S. alba* were clearly distinct in three of four VNIR bands (i.e., Green, Red, and NIR), whereas those for the *Rhizophora* and *Bruguiera* species were similar in most spectral bands. A similarity of spectral bands among mangrove species leads to a uniformity/resemblance as using spectral indices because spectral index is computed from spectral bands ratio. It shows that almost all of mangrove species also have the same threshold value in spectral index. For instance, red mangroves (*Rhizophora*), black mangroves (*Avicennia*), and white mangroves (*Laguncularia racemosa*) may have particular spectral reflectance in a single spectral band, but the signals of three species in NDVI/EVI are normally more than 0.3.

#### *4.2. Above-Ground Properties of Mangroves Estimation*

The term "above-ground mangroves properties" in our study refers to the estimation in aspects of mangrove ecosystems above ground such as leaf area index (LAI), biomass, carbon, vertical structure, and mangrove health. Understanding these variables is beneficial in detecting the interaction of vegetation, the stability of that interaction, and the change in mangrove population [7,263,274]. Historically, the majority of research on these parameters' estimation employed ground-based approaches that were time-consuming, costly, and distribute sparsely across space, making regional mangrove monitoring challenging [25]. Alternatively, a number of papers predicted mangrove above-ground properties using vegetation indicators [177,209,220]. These research applied regression analyses to establish empirical relationships between remotely sensed vegetation indices and measured-above ground mangrove (AGM) data (e.g., leaf area index, height canopy, carbon sink, and biomass) [207,209,275,276]. These studies considered that indices derived from satellite data successfully modelled and estimated the mangrove above ground features. However, above ground mangrove ecosystems today have not been compared with the patterns of 30–50 years ago.

In these studies, NDVI and EVI demonstrate the most explanatory curvilinear relationships with AGM [25,65,93]. In fact, there is a saturation issue with NDVI that is mainly due to the red band, the energy in which is strongly absorbed by pigments. When a leaf contains a certain number of pigments, the reflectance remains low and practically constant with more pigment (e.g., increased leaf area) [70,214]. As a result, where forests are high in biomass, NDVI struggles to differentiate moderately high plant cover from very high plant cover [153,214]. Since 2012, most studies confirmed that EVI indicated a higher correlation coefficient with mangrove in field measurements than NDVI to significant extents [59,153,169]. Meanwhile, a few publications found NDVI to be the best mangrove predictor, relative to the performance of EVI and other vegetation indices [80,144,160]. The scale of the analysis also affects the suitability of the index used, with a few studies finding that at a higher resolution (smaller scale) NDVI is the preferred index [84]. In fact, each ecosystem has its unique characteristics, and each index is a separate indication for green vegetation. The best vegetation index to use for AGM estimation varies and thus its selection requires substantial field measurements to validate the results.

#### *4.3. Mangrove Changes*

Understanding of variations in mangrove patterns is critical in providing fundamental source for proposing appropriate strategies in mangrove ecosystem management and serving as a reference for broader worldwide applications. The changes in mangrove ecosystems are acknowledged as a result of natural influences and human activities. To investigate mangrove changes, two methodologies are usually applied (i) bi-temporal analysis and (ii) long-term monitoring. Bi-temporal analysis uses two images per 5 or 10 years to assess the changes in mangrove cover. A bi-temporal technique is common and easy to apply, and it calculates the differences of mangrove cover at two times in the context of land use and land cover change. However, there are some limitations in regard to environmental factors (e.g., tide variations, terrain, and atmospheric conditions) if images are only obtained on a single day of the year. This is because we can utilise the method outlined in Section 4.1 to retrieve information about land cover for each year before employing an intersection of two scenes. Besides, the significance of physical factors (e.g., erosion, typhon) as a source of mangrove loss and the trend in mangrove cover may be underestimated.

In contrast, long-term monitoring normally applies time series of spectral indices to understand mangrove dynamics through space and time. Most of studies applied NDVI or EVI and linear regression algorithms to analyse spatiotemporal change and anticipate trends in mangrove distribution at a local scale [172,183,214,277]. These analyses concluded that the loss of mangrove ecosystems is mainly caused by conversion in land use and land cover, compared to natural factors. The aquaculture ponds and impervious surface expansion in the coastal area are a threat to mangrove ecosystems [28]. Globally, Hansen et al. (2013) [277] first used NDVI and ordinary least squares slope of the regression to examine forest loss and gain from 2000 to 2012. The study revealed that a decreased trend in mangrove cover occurred in Asian and Caribbean countries. However, the global analysis overlooked driving factors (e.g., land use and land cover transformation and physical hazards) because policies for land use and land cover changes are different among nations in the world.

Overall, the application of spectral indices for mangrove remote sensing provides several advantages in relation to mapping spatial distribution, above-ground mangrove properties, and mangrove changes. However, examples of knowledge gaps from previous studies should be included (i) visualising mangrove changes from the past to the present; (ii) identifying the driving elements impacting mangroves; and (iii) evaluating effects of environmental factors on satellite images.

#### **5. Discussion and Future Directions**

#### *5.1. The Potential Indices for Mangrove Remote Sensing*

Over the past 50 years, the importance of spectral indices in mangrove remote sensing has been recognised, but several knowledge gaps still exist in relation to the best index selection for mangrove characterisation. A perfect spectral index for an ensemble of mangrove biophysical parameters is yet to be developed. Our study explored that NDVI accounted for the highest proportion (82%) of the applied spectral indices for mangrove ecosystems, followed by EVI (28%). These normalised algorithms mitigate the atmospheric effects and the impacts of sensor calibration degradation in the red and NIR bands [64,66]. The widespread adoption of remote sensing has resulted in the creation of low-cost image data that may be used to broaden NDVI applications. Hence, NDVI will continue to be a dominant vegetation index used for mangrove remote sensing. However, this does not mean that NDVI is always effective because of its limitations in relation to soil background and vegetation density.

Each index has its own advantages and disadvantages and can be affected by the impacts of the soil background and confounding factors such as sensor and calibration effects, bidirectional reflectance distribution function, atmospheric and topographic effects, or other local environmental conditions (e.g., tide). Hence, for future applications, instead of constructing or discovering a prospective mangrove index, we should examine local conditions and the factors influencing the effectiveness of spectral indices before deciding on the use of them for analysis. Additionally, the number of spectral bands available on optical sensors influenced the indices chosen for mangrove remote sensing. For instance, the four-band version of the Planetscope instrument with no SWIR band is only able to produce spectral indices in the visible and NIR regions. In the case that one index cannot meet the needs of mangrove assessment or other purposes, another index should be applied.

#### *5.2. Long-Term Mangrove Monitoring with Time Series-Based Approaches in Relation to Driving Factors*

Monitoring mangrove dynamics normally includes seasonal and annual changes that require a series of historical and regular imagery. In fact, there are many factors, including tide conditions, atmospheric factors, or missing or mis-registered data, that can cause errors in image acquisitions. Therefore, using single-date images to calculate spectral index has shown significant limitations on a large scale because environmental conditions vary from day to day and across sites. Alternatively, generating optical images using averaging is less susceptible to high resolution noise and are thus capable of characterising both long-term and abrupt mangrove changes. For example, using annual mean/median spectral indices that are derived from daily/5-days/8-days/16-days timeseries data enables us to reduce the environmental factors' influence on the image of interest. The study of multiple remotely sensed data has been widely employed in phenological investigations of mangrove ecosystems [121,134,153].

In addition to data, time series analysis provides pieces of information on the timing of mangrove change, as well as improving the quality and accuracy of information being derived using remotely sensed data [167,172,174,217,219]. Also, the time series analysis of spectral indices data evaluates trends and predicts the persistence of mangrove trends under spatial regression application. A variety of time series analysis techniques have been produced [(e.g., National Forest Trend [278], Recurrent Neural Network [279] to analyse and monitor spatiotemporal changes in mangrove ecosystems [191,221,280]. The digital number (DN) value of each pixel from time series images gives more sensitivity than single composited spectral band so that it can easily compare with natural factors (e.g., rainfall, temperature, and ocean dynamics) to certain significance of physical influences. This method holds significant promise for studying the long-term dynamics of environmental variations, and it can monitor future mangrove regeneration.

#### *5.3. Fusion of Images from Multiple Sensors*

In the process of the Earth's surface observation and particularly in mangrove remote sensing, selecting a potential optical sensor to calculate spectral index is crucial for assessment with high accuracy. However, there are several factors that can influence the choice of optical remote sensing platforms, such as the purpose of the research, the data availability, the national context, the budget constraints, the scale of the study, and the location of the study area. For instance, to explore spatiotemporal changes in annual mangrove patterns, a long-term time series analysis from Landsat imagery should be preferred because the data is available from 1972 to date. Additionally, to understand mangrove quality or seasonal changes, a variety of MODerate Resolution Imaging Spectroradiometer (MODIS) products with a high temporal resolution (1 day) may be the best choice. Recently, several studies fused multi-sensor images to have more information about mangrove ecosystems. For example, Kanniah et al. (2021) [281] used three optical sensors (i.e., Landsat, MODIS, and Sentinel-2A) to study mangrove fragmentation and health conditions. Guo et al. (2021) [209] used UAV and WorldView-2 datasets to validate the Sentinel-2 imagery for LAI estimation. Besides, several publications fused passive and active sensor images to understand mangrove structure or biomass. Pham et al. (2020) [89] combined optical bands (Sentinel-2A) with active sensors (i.e., Sentinel-1 and ALOS-2 PALSAR-2) to calculate some vegetation indices for mangrove above-ground biomass. These studies concluded that fusing multiple remote sensing sources helps to provide a large amount of information about mangrove ecosystems, compared to single sensor applications.

Fusion of multiple sensors can enhance the accuracy of the data. Integration of NDVI from Advanced Very High-Resolution Radiometer (AVHRR–launched in 1979) and MODIS (launched in the 2000s) enabled a long-term dataset from 1979 to date. AVHRR NDVI composites at 1 km spatial resolution [92,93] was used for mangrove monitoring prior to the 2000s. However, the AVHRR satellite system has degraded in orbit to the point that it is advised that NDVI MODIS products should be used for longer periods in the future [282]. Additionally, combination of ASTER (launched 1999) and Landsat 4,5, 8, and 9 is an alternative approach for line correction of Landsat 7, allowing us to obtain a set of data at 30 m spatial resolution from 1988 to the present. Notably, using multiple sensors enables improving the re-visit days of satellite data, which is better for smoothing data [25,27]. For example, when Landsat 8 and Landsat 9 are combined, the re-visit days are reduced from 16 to 8 days. Hence, fusion of multiple sensors (i.e., passive, and active sensors) is a recommendable approach to compute the spectral index for mangrove studies in the future.

#### **6. Conclusions**

Land use and land cover transformation in relation to natural hazards are the primary factors threatening mangroves in the future. Spectral indices have been applied to mangroves and demonstrated their effectiveness in various studies over 50 years. Each spectral index has its own strength and limitation in mapping mangrove distributions and measuring their above-ground biophysical properties in various environments. Therefore, to select a potential index, we should understand the interaction between the local conditions and mangrove ecosystems. NDVI is the most popular index that can be applied for mangrove ecosystems, followed by EVI, although both are sensitive to environmental conditions.

Long-term mangrove monitoring is crucial for identifying the trend in mangrove pattern changes in connection to driving variables. Using time series analysis of spectral indexes helps to reduce the effect of external influences. Using multiple sensors enables obtaining a set of databases for long-term monitoring associated with natural hazards and human activities. Nowadays, accessing big data has become easier with the help of technology and digital cloud platforms (e.g., Google Earth Engine). These technological advancements will shift mangrove studies from a local to a global scale and imply the necessity to learn programming skills.

In the context of the digital era, mangrove scholars should apply the advantages of cloud computing platforms for spectral index computation. These approaches assist image processing quickly and enable analysis of mangrove ecosystems at global scale. However, approaching these techniques requests users to have some knowledge about information technology and quite understand about coding. Therefore, developing a tool or application on cloud storage for mangrove monitoring based on vegetation index should be taken into account for new users or scholars who do not have good information technology skills. In addition to tools, a guideline for algorithm selection (e.g., machine learning, deep learning) should be developed to save time for spectral index computation.

**Author Contributions:** Conceptualization, methodology, formal analysis, investigation, writing original draft preparation: T.V.T.; writing—review and editing: T.V.T., R.R. and X.Z.; supervision: R.R. and X.Z. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was funded by an Australian Research Council Discovery Award DP180103444 to R.R.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** Acknowledgement is given to Monash University for supporting this research. The authors would like to acknowledge the valuable comments of anonymous reviewers and editors that assisted with the finalisation of this manuscript.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **Acronyms**



#### **References**


MDPI St. Alban-Anlage 66 4052 Basel Switzerland www.mdpi.com

*Remote Sensing* Editorial Office E-mail: remotesensing@mdpi.com www.mdpi.com/journal/remotesensing

Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Academic Open Access Publishing

mdpi.com ISBN 978-3-0365-8887-2