Vegetation Indices for Spatio-Temporal Analysis of the Quality of Can Gio Mangrove for Biodiversity and Conservation

Mangrove forest has an important role in protection against coastal erosion and carbon dioxide sequestration for climate change mitigation. The natural increase in vegetation quality can explain the increase in its health and biodiversity. This study aimed to examine the health of Can Gio Mangrove vegetation to determine its potential for preventing coastal erosion and maintaining biosphere’s conservation function. Algorithms in GIS and remote sensing were used along with the time-series temporal Landsat images in the 2010-2020 period to investigate the changes of vegetation indices: Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), Leaf Area Index (LAI), and Normalized Difference Water Index (NDWI). The results showed positive changes of vegetation values between dry season and rain season, and large distribution of their good values in the 2010-2020 period. Decrease in the vegetation area and stability of surface waters area indicate the expansion of urbanized area. These results may support modification of the policy to control the urbanization, to maintain the mangrove biosphere reserve, and to enhance coastline stabilization.


Introduction
Mangrove forest located in the coastal area is one of the most productive ecosystems of the world.Connecting land and marine environments, it protects coastlines, enriches eco-diversity and sediments, and interacts with natural stressors.It has important roles as a freshwater resource (riverine inputs of freshwater and nutrients), in reducing soil salinity, and as a coastal environmental defence [1][2][3].In terms of climate change mitigation, mangrove forests act as carbon sinks in the world carbon cycle.To the communities, it provides fisheries and commercial products [1,3].As they constitute a transitional zone linking terrestrial and marine environments, mangroves regularly suffer deforestation resulting from overexploitation (water pollution, overused fisheries, and human encroachment), biodiversity loss and ecosystem change (related to species abundance, ecosystem structure and compositions changes, and freshwater regimes alternation) [3,4].It also faces vulnerabilities connected to climate change and natural impacts (such as hurricanes, coastal erosion, and soil salinity, etc.) [3].Monitoring the mangrove interactions between coastal land and sea, its sedimentation processes and morphology can help develop long-term policies for mangrove health, coastal protection, local inhabitants and heritage conservation, as well as climate change mitigation.
Tracking the mangrove morphological processes and damages can be conducted through temporal and spatial analyses and measurements.These can serve as grounds for construction of a database for management and monitoring, as well as for data inventory for development [2].There are many vegetable indices that can be used to examine mangrove health, its morphology, its density and distribution [5].The well-known indices used by many researchers are normalized difference vegetation indices (NDVI) and enhanced vegetation index (EVI) [6][7][8].The leaf area index (LAI) is another important index to analyse vegetation structure and density [9][10][11].Other indices such as PTC (percentage tree cover), SAVI (soil adjusted vegetation index), MSAVI (modified soil adjusted vegetation index), OSAVI (optimized soil adjusted vegetation index), TSAVI (transformed soil adjusted vegetation index), FAPAR (fraction of absorbed photosynthetically active radiation), etc., can well support analyses of vegetation quality and health [2,7,10,12].In the mangrove forest management, the Normalized Difference Water Index (NDWI) can be used for monitoring water bodies and water content in vegetation canopies [7,13].There are direct and indirect methods to monitor vegetation.Vegetation structure and health can be directly measured through sample sites in the forest [14].This method is effective and produces precise results.However, it is time-consuming and costly if it is to be applied in an analysis of a large area.For large areas, such as mangrove forests, spatial observation using remote sensing methods based on the satellite images from the multispectral scanners becomes more convenient and efficient [10].Remote sensing technology is well developed and advanced, with images in various resolution that can be freely accessed [15,16].It provides advantageous time-series of multispectral data that allow for a rapid monitoring and effective evaluation of large areas such as forests.The optical and digital images are useful to obtain information on vegetation and surface waters in relation to mangrove distribution, coastline protection and conservation, and to monitor human activities [6,10,16].This study used Landsat images, available from the USGS, to determine indices of NDVI, EVI, LAI, and NDWI and monitor vegetation distribution and water body changes of Can Gio Mangrove of Ho Chi Minh City of Vietnam in the 2010-2020 period.

Description of study area
Figure 1: Can Gio Mangrove study area Can Gio Mangrove, located in the south of Ho Chi Minh City, is Vietnam's first world biosphere reserve recognized in 2000 by the United Nations Educational, Scientific, and Cultural Organization (UNESCO).The mangrove covers more than seventy-five percent of Can Gio territory.The mangrove biosphere reserve was recognized as a multi-purpose conservation model with a harmony between nature and humans based on biodiversity and cultural diversity.The mangroves are located in a low elevation and flat topography: most of the mangrove area is in coastal plains.The terrain elevation changes little: from 0.2 and 0.5 meter to 1.3 and 1.5 meter, with the lowest elevation in the centre.The Can Gio Mangrove has a sub-equatorial monsoon climate, that is hot and warm with two distinct rainy and sunny seasons.Can Gio is typified by an interlaced network of rivers and canals.The water comes from the sea by the two main funnel-shaped gates of Dong Tranh Bay and Ganh Rai Bay.The second water source is from the Saigon and the Dong Nai rivers.The area is under the irregular semi-diurnal tides and Can Gio has the highest tide in Vietnam amounting to 4 to 4.2 m.Can Gio Mangrove Biosphere Reserve maintains the following functions: (1) it plays a vital ecological role by limiting erosion, maintaining soil fertility, regulating climate, and limiting air and water pollution; (2) supports the study of natural ecosystems, genetics, maintenance and creation of sustainable livelihoods for people and environmental protection; (3) facilitates sharing of experiences and knowledge exchange for sustainable development; and (4) contributes to local socio-economic development.Mangrove has three wildlife conservation areas: Bird Sanctuary (with 602.5 ha), Bat Conservation Area (with 123.2 ha), and Monkey Sanctuary (with 482.2 ha).These locations play a significant role in terms of conserving genetic resources, conserving biodiversity and serving scientific research.

Data acquisition
In 1972 the first Landsat was launched on the Earth's orbit.In recent decades, satellite images are freely accessed and widely used in many scientific research fields [16].The high-resolution satellite images from multispectral scanners, such as Landsat 5, 8 and 9, contain consistent information for spatial analysis and for thematic mapping [17].The Landsat 5 TM and 8 OLI images were acquired to analyze the NDVI, EVI, LAI and NDWI in Can Gio Mangrove.To well understand the mangrove distribution and development, the study compared the indices in dry season and rain season of each year to investigate the changes of the forest.The season indices were then compared to the annual indices.The Landsat 5 TM images were collected for the year 2010, and the Landsat 8 OLI images were for the years 2015 and 2020, since the Landsat 8 launched in 2013.The Landsat satellites issues every month two images.These images were freely acquired from the USGS website to calculate the average annual and season indices of the years 2010, 2015, and 2020.Most of the chosen images of the study area had land cloud cover less than 20 percent, and they were described in the Table 1 [2,7,18].NDVI is commonly used for vegetation segregation, plant growth, and vegetation cover estimation [9].This index ranges from −1.0 to 1.0.The positive and negative values (above and below 0, respectively) indicate non-vegetated features (bare land, water, snow or ice, or clouds).The low green values are from 0 to 0.2.The moderate values are from 0.2 to 0.3 (shrub and grassland), and the high green values range from 0.4 to 0.9 [19].The NDVI values in the 2010-2020 of Can Gio Mangrove were calculated with the formula: However, NDVI cannot identify some effects on vegetation (such as dust, atmospheric aerosols and soil background) to obtain precise calculation [20,21].Therefore, EVI was developed to overcome the NDVI problems with temporal and spatial vegetation variations and clarify topographic differences [9,20].In the urban area, EVI is as a proxy for analysing vegetation variations due to urbanization.It also has high sensitivity to biomass areas [22].EVI range value is similar to NDVI and is calculated with the formula: 3.3.LAI calculation LAI, which is an approximate of nitrogen contents and leaf chlorophyll, is one of the vegetation indices to estimate the largest amount of plant canopy in the vegetation area and geometrically estimate oneside area of leaf photosynthetic tissue per ground area [11,23].In the case of changes in microclimate, LAI can estimate the mass and energy, and the vegetation changes in canopy structures.It can also explain the interchanges between the atmosphere and leaf canopy, and the changes of environment [24].By using Landsat images, LAI can be estimated through the following formula: LAI = (3.618x EVI) -0.118 (3) In Asrar et al. (1992) study [25], LAI equals to 1 when canopy has lower clump leaf area in the greater ground cover.When canopy has higher clump leaf area in the lower ground cover, LAI equals to 10. Scurlock [26] showed the LAI values ranging from 2 to 4 represent crops, from 3 to 5 represent vegetation, and from 6 to 8 are for the forest.

NDWI calculation
The NDWI of McFeeters [27] is proposed to recognize the content changes in open surface water and it has association with the wetlands.It is calculated from the near infrared band and the green bands.Gao [28] proposed the NDWI to monitor the changes in the leaves' water content.And it is computed by using the two satellite bands of near infrared (NIR) and the shortwave infrared bands (SWIR1).

NDWI =
The different reflectance between NIR and SWIR1 bands is to count the vegetation's water content, since the wavelengths of SWIR1 can absorb water.Therefore, NDWI is a useful indicator to monitor the changes of mangrove forest, even the small disturbance or tiny cuts.This index can seize changes by the shortwave infrared bands [13].The area of Can Gio Mangrove forest in 2010 was 33,050 ha, in 2015 its area increased to 33,752.46 ha and by 2020 to 34,562.03ha thanks to the efforts of the afforestation projects of HCMC [29].
Mangroves have dense trees and interlaced roots, which contribute to reducing water flows, creating sediment to be deposited in coastal estuaries.This contributes to enlarging of the land area, and reduces the amplitude of waves and storms passing over the land.In 1953 -2010 period, the total increasing area of alluvial land was 316 ha, the average accretion area was 5.5 hectare.At present, newly formed accretionary land areas in Can Gio Mangrove contribute to increasing the area of the land suitable for afforestation programs to adapt to climate change.

Distributions of NDVI and EVI in 2010-2020 period
Can Gio has 32,446.44hectares of forest land, according to the Decision No. 3901/QD-UBND dated July 20, 2017 of the People's Committee of Ho Chi Minh City [30].In 2010-2019, the new afforestation area has increased 1,916.23 ha (5.9% of the total forest area), including the activities of zoning and promotion of afforestation and reforestation.
Compared with the record of the first periodical assessment since 2010, Can Gio Mangrove Forest Biosphere Reserve conserve high diversity of species, ecosystems as well as genetic diversity.Plant and animal species generated, compared to the past ten years.According to Can Gio People Committee in 2019 [31], Can Gio Mangrove Biosphere Reserve has 68 families, 212 genera, and 318 species.

LAI estimation
Leaf area index describe the dense level of leaves in the unit land area.Compare to NDVIs and EVIs above, LAIs show the dense distribution of leaves in most of Can Gio mangrove in 2010-2020 period (Figure 5).However, the dense of leaves is thinner in 2015 and even thinner in 2020 in both dry season and rain seasons.The urbanization brings higher demand of forest-based livelihoods that affect the mangrove distribution [32,33].

Figure 2 :
Figure 2: The change of mangrove forest in 2010 -2021The area of Can Gio Mangrove forest in 2010 was 33,050 ha, in 2015 its area increased to 33,752.46 ha and by 2020 to 34,562.03ha thanks to the efforts of the afforestation projects of HCMC[29].Mangroves have dense trees and interlaced roots, which contribute to reducing water flows, creating sediment to be deposited in coastal estuaries.This contributes to enlarging of the land area, and reduces the amplitude of waves and storms passing over the land.In 1953 -2010 period, the total increasing area of alluvial land was 316 ha, the average accretion area was 5.5 hectare.At present, newly formed accretionary land areas in Can Gio Mangrove contribute to increasing the area of the land suitable for afforestation programs to adapt to climate change.

Figure 5 :
Figure 5: LAI in dry and rain season in 2010-2020 The LAI values below 1.5 in dry season have larger distribution than in rain season.The LAI values above 1.5 in rain season have positive changes in larger distribution than in dry season which can indicate the mangrove's health with high ability of forest recovery.4.4.Dynamic changes of mangrove water body Can Gio Mangrove has diverse water system with very dense network of canals and rivers.NDWI indicates various values of open surface water.It estimates as well the absorbed water in the leaf canopies.Figure 6 shows the dynamic changes of surface waters of Can Gio Mangrove between dry season and rain season in the 2010-2020 period.The low values of NDWIs in dry season distribute in larger area in the mangrove.However, they have smaller area in rain season.The higher values of NDWIs (from 0.4 to 0.8) show a positive increase in the rain season, in comparison with the same these values for the dry season.

Figure 6 :
Figure 6: NDWI in dry and rain season in 2010-2020 4.5.Dynamic changes of mangrove vegetation The distributions of NDVIs, EVIs, LAIs, and NDWIs in Figure 7 to Figure 10 indicates the healthy forest of Can Gio Mangrove.Most of the vegetation indices have large distribution of high value.NDVIs have a dynamic change between dry season and rain season, however, EVIs are more stable in two seasons.In rain season, LAIs have positive values than in the dry season, and evenly expand in most area of the forest.

Figure 7 :
Figure 7: Distribution of NDVI values (in hectare) in a whole year of 2010, 2015, and 2020

Table 1 :
below: The information of the satellite images NDVI and EVI calculation NDVI and EVI are the most popular vegetation indices and widely used in many researches in recent decades 4 3.2.

Table 2 :
The yearly value ranges of vegetation indices of Can Gio in 2010-2020 The indices differencing is to examine the changes of indices in area.The indices were classified in groups of values to compute and compare the changed areas in the dry season and rain season in the years of 2010 to 2020.It can be used to monitor the coastline stabilization of Can Gio mangrove.Positive changes of Can Gio mangrove During the war, Can Gio Mangrove was heavily destroyed in most of the area.It had periods of reforestation since 1975.The total natural area of current mangrove forest is 32,451.02ha (of the total 34,672.79ha).In 2008, the People's Committee of Ho Chi Minh City issued the plans for Can Gio Mangrove Forest Biosphere Reserve with an area of 75,740 hectares, divided into three areas: (1) core area, 4,720 hectares; (2) buffer zone, 37,340 hectares; and (3) transition zone, 29,310 hectares.