Estimation of Shoreline Changes of Can Thanh Townlet and Long Hoa Commune in Hochiminh City

Can Gio is considered as the “green lung” of Ho Chi Minh City with the mangrove ecosystem with many investment projects as well as livelihood models associated with the strength of mangrove forest. Currently, in Long Hoa commune and Can Thanh town, the “Can Gio sea encroachment project” is being implemented with the goal of creating an urban area of sea encroachment tourism, resort and business tourism with events, conferences and seminars. In the process of construction, infrastructure expansion, service development and dense population concentration, it is inevitable that the shoreline change processes will be affected, especially within these two coastal areas. This study presents a method of using the integrated DSAS tool in ArcGIS together with Sentinel 2A image data to assess the shoreline change of Long Hoa commune and Can Thanh town from 2015 to 2022. By extraction shoreline through the MNDWI index, the results show that the process of accretion and erosion were intertwined, except for the sections of the shore that have been reinforced with sea dykes. It is evident that in the land at Dong Tranh Cape, southwest of Long Hoa commune and northeast of Can Thanh townlet, the process of erosion is taking place, while in the southwest, the process of accretion is predominant. The results of this study can be considered as a reference for coastal development in Can Gio sea area in the present and in the future.


Introduction
Can Gio district is located about 50 km southeast of the city center and is the only district of Ho Chi Minh City bordering the sea, with a coastline stretching 23 km.The district has a total area of more than 71,300 hectares, out of which 70% are mangroves and rivers.Can Gio mangrove forest, in addition to the interwoven system of canals, supports extremely rich ecosystems with high biodiversity with many endemic species of animals and plants.Development plan for Can Gio district by 2030 assumes that Can Gio will become a green, smart, environmentally friendly coastal city with a focus on a high-quality resort eco-tourism.The sea reclamation project in Can Gio was started in 2007 with an area of 600 ha.In 2019, it was adjusted and expanded to an area of 2,870 ha.This project has an artificial lake of about 757 ha.Due to the process of socio-economic development, as well as the expansion of many construction works towards the coast, the natural factors, especially alluvial deposition and sea currents have been changed.Researching shoreline position and its variability is of great importance for scientists, engineers and managers of riparian and 1247 (2023) 012007 IOP Publishing doi:10.1088/1755-1315/1247/1/012007 2 coastal areas.Knowledge and assessment of the changes in shoreline position have proved crucial in the overall understanding of dynamics in coastal areas and the morphodynamic processes driving the change.The shoreline is the natural interface between the land and water.Its position is identified in the immediate time and it can change remarkably over time [1].Shoreline geometry and position are perhaps the most basic indicators with which to evaluate changes in coastal regions [2].There are 2 types of shoreline change, the first is due to the accumulation process: the phenomenon of the coastline gradually encroaching towards the sea.The second is due to the erosion process: the coastline is gradually encroaching towards the mainland [3].Many studies used remote sensing data to classify water surface from multi-temporal satellite images to identify and evaluate shoreline changes.The remote sensing data, including Landsat satellite images and ultra-high resolution (SAR) images are used for supervised classification and combined with texture analysis to assess shoreline changes [4].Other study used Landsat MSS, TM, ETM image data and the ration band to extract the coastline, using the LRR method in the DSAS tool to analyse shoreline changes for northern part of the Coromandel coast [4,5].Abdul Basir Mahmoodzada (2019) [6] used Landsat images and DSAS tools to investigate quantitatively the trend of riverbank erosion, by detecting changes in the coastline and displacement of the Amu River during 14 years in Afghanistan.In Vietnam, there are many studies using water-related indicators (NDWI, MNDWI, etc.) and DSAS tools to analyse shoreline changes.For instance, shoreline changes were analysed in the Tien and Hau river areas in the period 1989 -2014 using the MNDWI index to classify shorelines based on multi-dimensional Landsat images.DSAS tool was used to calculate riverbank change rate [7,8], while the rate of accretion and coastal erosion in Ca Mau was determined with DSAS tools and Landsat images [9,10].Other studies focused on identification of the eroded coastlines in Thua Thien Hue with the DSAS 4.3 tool [11], and the threat to the natural stability of the Tam Giang -Cau Hai lagoon system due to the coastal erosion and accretion at the Thuan An and Tu Hien estuaries [12].The focus of this article was the application of DSAS tools to analyse changes in shoreline positions on the basis of water surface and land extraction by MNDWI index.In this research, the shorelines were extracted from Sentinel 2A satellite images (with high resolution) from 2015 to 2022.The imagery was digitized and analysed using software programs ArcGIS and DSAS.The objective of this study was to assess the shoreline change in the period 2015-2022 for the coast in Long Hoa commune and Can Thanh townlet in order to provide information for the management and planning of future projects.

Study Area
Based on administrative data that was downloaded free from web https://www.diva-gis.org/,we focused on Can Thanh townlet and Long Hoa commune.The satellite image that we used in mapping of the study area map was collected from Google Earth in 2020.

Shoreline extraction by using MNDWI index
The interface between land and sea, defined as the shoreline, changes variably due to multiple factors, which may be morphological, climatological or geological in nature [13] and also is the standard waterline that moves in and out with the ebb and flow of the tide.In this study, shoreline means the contiguous part between land and water, and it was the basis for separating the shoreline feature.MNDVI index was used for shoreline extraction for this study area.
For shoreline extraction, the MNDWI method was used to identify the land and water features.According to Umit Duru [14], the MNDWI index was calculated according to the formula: MNDWI = (Green -MIR)/(Green + MIR) For Sentinel 2A images, the MIR band corresponds to band 11 (SWIR).Because the Green band has a resolution of 10 m and the SWIR band has a resolution of 20 m, the MNDWI index was calculated at a resolution of 20 m.Therefore, to synchronize the spectral resolution, the SWIR band was converted to a resolution of 10 m.To perform this step, using the Resample command in ArcGIS software (in the Raster Processing tool box).The steps were shown in Figure 2. Band 4 was selected as the input parameter, as the basis for converting Band 11 with a resolution of 10 m.After extracting the shoreline using the MNDWI index, the Reclassify tool in ArcGIS was used to classify the threshold into 2 classes -land and water surface, land was an object with value 0 (black) and water had value 1 (white).Using binary operations, from the resulting image of MNDWI, pixels with a value greater than 1 on the ratio image would be assigned the value 1, pixels with a value less than 1 assigned the value 0. Thus, water would be assigned the value 1, the land would be assigned the value 0. 5 In this study, land had the threshold from [-1] to [-0,4] pixel value and the pixel value of water was from above [-0,4] to 1.Then, we used convert raster to polygon command in ArcGIS to digitize the shoreline.Due to the limitation of actual survey data, the classification results were based on Google Earth for fixing the shorelines in the period 2015-2022.Besides, the correction of riverbank results with tidal level was ignored because the images were selected in the dry season, so the water level difference is not much.
The shoreline by time study in this research: The calculation and analysis of the shoreline were carried out as follows: -determine the baseline and the calculated shorelines, -create transect lines perpendicular to the shore (transect).The baseline is considered as a starting point for all the transects opened by the DSAS program.DSAS creates transect lines that are perpendicular to the reference line, i.e., baseline, at a user-specified distance interval.The distance between the baseline and each point of the intersection of the shoreline with the DSAS transect can be used to calculate the distance measurements such as shoreline change envelope and net shoreline movement [15].The DSAS tool calculates shoreline change rates based on measured differences between shoreline locations relative to specific time points [15].Transects are constructed perpendicular to the baseline, it is necessary to have at least two shorelines from 2 different time points to calculate the rate of change using DSAS.
The shoreline was carefully digitized and exported to shapefile format, and the extracted waterfront was used as input to the DSAS model to calculate the rate of change.The baseline is constructed inland and parallel to the general trend of all shorelines.To assess the trend of shoreline change, an onshore baseline was created at a distance of approximately 30 m behind the riverbank.Based on the baseline and shoreline, the landslide and accretion modes are calculated for the transects with a distance of 5 m apart.

Shoreline analysis and interpretation
The DSAS approach calculates shoreline rates of change based on the measured differences between the shoreline positions associated with specific time periods.In this study, the statistical calculation methods used include: -Shoreline Change Envelope (SCE): used to calculates the displacement of 2 shorelines (in meters).SCE is the maximum shoreline change distance considering all available shoreline positions and gives their distances, without reference to specific dates of shorelines.SCE = maximum distance (m) between all shorelines (excluding shoreline dates) -EPR (End Point Rate): is the method that requires the least amount of input data.EPR is calculated by dividing the distance of the shoreline over time between the oldest and newest shorelines in a given data set [16].If the EPR value is below 0, it means that the area has been eroded and if the value is positive, it means that the area has been accreted.EPR = Volatility Distance / Volatility Time

Result
After performing all the steps of Sentinel-2A image processing and digitizing to separate the waterline on ArcMap, we performed a dynamic analysis of shoreline in Long Hoa commune and Can Thanh townlet on the DSAS tool then we obtained the following results: Erosion took place in the northern part of Can Thanh, where alluvial zones were narrowed to build wharf structures and residential areas.The main erosion rate in this period was from 0 to 1m/year.There are 248 transects showing erosion trends and locations shown in Figure 6.Between 2015 and 2018, the shoreline of Can Thanh moved from 5m to 19.9 and was mainly distributed within this threshold (in Table 4).The shoreline sections displacement due to the impact of the dyke and sea embankment system.Besides, the displacement of the shoreline towards the sea was due to the impact of the encroachment project at Can Gio.The results of shoreline analysed within 3 years (from 2015 to 2018) increasing to 19.9m, when compared on Google Earth satellite image, we could see that the areas with shoreline displacement > 15m, concentrated in sections with embankments and mangroves.This embankment made the accretion process stronger.In 990 transects EPR values were below 0 m/year, which means that most of the townlet's shoreline during this period was subject to erosion; in most cases the retreat was up to 1 m/year.However, the distribution of these transects was not continuous (they are not adjacent) The areas with most erosion were in the coastal residential areas, where erosion rates exceeded 2m/year (over 426 transects).The accretion was most evident in mangrove areas.The accretion rate per year during this period was mainly from 0 to 1m/year, occurring in 128 transects.The place with the highest accretion rate (>2 m/year) was in the southwest of the townlet, the area that could be seen most clearly on satellite images was the expansion of alluvial flats towards the sea.

The period 2020 -2022
In the period of 2020 -2022, the accretion process was dominant and occurred in over 50% of the total of transects (787 transects), located mainly along Duyen Hai road, where the encroachment project was built.Here, the highest accretion rate was 2 m/year.However, in the process of expanding construction towards the sea, some sections of the shoreline with mangroves have been narrowed, causing the shoreline to be eroded by > 2m/year.The obtained data (Table 8 and Figure 10) shows that between 2020 and 2022 in the majority of sections (435 transects) the shoreline shifted by 0 m to 3.7 m; mainly in the area of the sea encroachment project.The shoreline displacements exceeding 7m were occurred in sections with the alluvial decline of the mangrove forest or the expansion of the residential area and constructions towards the sea.9 presents calculations of the rate of shoreline change in Long Hoa commune in the period 2015 -2018 obtained by applying DSAS tool.The shoreline of Long Hoa commune was mainly subject to accretion that was recorded in 1140 transects).However, the intensity of the process was moderate, accretion occurred in the eastern part of this commune, spanning 394 transects, with a sedimentation rate of 0 to 1m/year.Erosion was greatest along the coast of the April 30 beach, with landslide rates exceeding 2m/year, locally reaching 6.7 m/year.The shoreline at Dong Tranh Cape displayed average sedimentation rates (1-2m/year).In the remaining part of the shoreline, the erosion process was weak, with a rate of 1-2m/year, occurring in 420 transects, distributed in the mangrove forest along Duyen Hai street.

The period 2018 -2020
The transects of Long Hoa commune were mostly subject to accretion; the eroded sections were located mainly in mangrove area and the part of the beach 30/4.Most of the shoreline of Long Hoa commune was accreted, with more than 800 transects with accretion rates > 2m/year, located in Can Gio beach area, 30/4 beach and some beaches that tend to expand to the sea.The erosion process took place along only 40% of the shoreline, with a landslide rate of 0 -1m/year occurring on 338 transects.Eroded shoreline is located mostly along Duyen Hai street.The banks with erosion exceeding > 2m/year and locally up to 10.8 m/year were concentrated in the hamlets of Phan Duc and Long Thanh.

The period 2020 -2022
Similar to the previous periods, the erosion process occurred mainly at Dong Tranh Cape and some alluvial sections that were hit by waves (due to the impact of sea dykes).The erosion rate exceeded 2 m/year in 378 transects.
The accretion process dominated in this period: more than 70% sections were subject to accretion as shown in the DSAS calculation.In most cases this occurred at the shoreline of the construction site of the encroachment project.In general, the shorelines were being eroded and accreted at different rates.In Can Thanh, the accretion process was more dominant, with accretion rate reaching > 1m/year, there were also many sections with shore speed > 2m/year, mainly concentrated in the shoreline of the project of sea encroaching.Besides, the shoreline change at Long Hoa also had some notable things: The process of accretion and erosion occurred alternately but only at an average speed (1 -2m/year).After 2018, the accretion process was dominant, with an accretion rate > 2m/year.The areas with fluctuations were concentrated at the beach on April 30, the section from the mangrove forest in the west to the cape of Dong Tranh.
The results of this study can be used as a reference to serve the sustainable management of the shoreline and propose measures to limit erosion; from there, making policy with a strategy for coastal development.

Figure 1 .
Figure 1.The location of Long Hoa commune and Can Thanh townlet

Figure 2 .Figure 3 .
Figure 2. Band 11 from Sentinel 2A images was increased from 20m to 10m by using Resample If the spatial resolution of band SWIR is increased from 20 m to 10 m, the MNDWI with the spatial resolution of 10 m was calculated: MNDWI10m = (Green -MIR10m)/(Green + MIR10m)

Figure 4 .
Figure 4.The shorelines were extracted from MNDWI index 4.Using DSAS tool for determining the shoreline changes 4.1.Calculating the baseline and the shoreline Having obtained the shoreline for each year to calculate the rates of landslide accretion in this study we used DSAS tool version 5.0 (Digital Shoreline Analysis System) integrated in ArcGIS 10.7.1.DSAS supports the ESRI ArcGIS software to calculate the shoreline change rate over time by creating transects (cutting lines) orthogonal to the established distances, thereby calculating the shoreline change rate combined with statistics presented in the attribute table.The calculation and analysis of the shoreline were carried out as follows:-determine the baseline and the calculated shorelines, -create transect lines perpendicular to the shore (transect).The baseline is considered as a starting point for all the transects opened by the DSAS program.DSAS creates transect lines that are perpendicular to the reference line, i.e., baseline, at a user-specified distance interval.The distance between the baseline and each point of the intersection of the shoreline with the DSAS transect can be used to calculate the distance measurements such as shoreline change envelope and net shoreline movement[15].The DSAS tool calculates shoreline change rates based on measured differences between shoreline locations relative to specific time points[15].Transects are constructed perpendicular to the baseline, it is necessary to have at least two shorelines from 2 different time points to calculate the rate of change using DSAS.

5. 1 .
Shoreline change in Can Thanh townlet 5.1.1.The period 2015 -2018 Table 3 and figure 5 in 2015-2018 the shoreline of Can Thanh was affected mainly by accretion, distributed in the western part of the area (with mangrove forests and alluvium).There were 998 transects showing the accretion trend in which the sedimentation rate exceeds 2 m/year.

Figure 6 .
Figure 6.Comparison the shoreline in 2015 and 2018 with the displacement above 19.9m

Figure 7 .
Figure 7. SCE map of Can Thanh townlet in the period 2015 -20185.1.2.The period 2018 -2020Based on the figure7, the locations with the most obvious shoreline change are: the eastern cape of Can Thanh, the coastal campsite on April 30 beach, and the part from the beach on Tac Xuat street to the end of the west bank of Can Thanh townlet.These locations were mainly concentrated along mangrove forests and coastal alluvial flats, with some locations that have encroached on the sea due to construction projects (in comparison to the previous period (2015 -2018).For all 1536 transects, End Point Rate (EPR) and total Shoreline Change Evelop (SCE) were calculated by DSAS in ArcGIS; the results of the shoreline change calculations in the period of 2018 -2022 were as follows:Table5.Statistical calculation EPR of Can Thanh townlet during 2018 -2020

Figure 8 . 6 .Figure 9 .
Figure 8. EPR map of Can Thanh townlet in the period 2018 -2020 The shoreline in 2020 and 2022 in Can Thanh townlet mainly shifted from 0 to 3.1 m, over 709 transects, corresponding to the coastal areas where erosion or accretion only reached 0 to 1 m/year.And most of these shorelines were reinforced with sea dykes.Parts of the shoreline where erosion or accretion occurred at a high rate, the shoreline could shift from 12.3 m to the maximum value of 36.8 m, occurring on 198 transects.Table 6. Statistical calculation SCE of Can Thanh townlet during 2018 -2020 Length of shoreline change between 2 times (meters) Total number of transects 0 < SCE ≤ 3.1 709 3.1 < SCE ≤ 6.2 66 6.2 < SCE ≤ 9.3 133 9.3 < SCE ≤ 12.3 185 12.3 < SCE ≤ 36.8 192

Figure 10 .
Figure 10.EPR map of Can Thanh townlet in the period 2020 -2022The obtained data (Table8and Figure10) shows that between 2020 and 2022 in the majority of sections (435 transects) the shoreline shifted by 0 m to 3.7 m; mainly in the area of the sea encroachment project.The shoreline displacements exceeding 7m were occurred in sections with the alluvial decline of the mangrove forest or the expansion of the residential area and constructions towards the sea.Table 8.Statistical calculation SCE of Can Thanh townlet during 2020 -2022

Figure 11 .
Figure 11.SCE map of Can Thanh townlet in the period 2020 -2022 5.2.Shoreline change in Long Hoa commune 5.2.1.The period 2015 -2018 Table9presents calculations of the rate of shoreline change in Long Hoa commune in the period 2015 -2018 obtained by applying DSAS tool.The shoreline of Long Hoa commune was mainly subject to accretion that was recorded in 1140 transects).However, the intensity of the process was moderate, accretion occurred in the eastern part of this commune, spanning 394 transects, with a sedimentation rate of 0 to 1m/year.Erosion was greatest along the coast of the April 30 beach, with landslide rates exceeding 2m/year, locally reaching 6.7 m/year.The shoreline at Dong Tranh Cape displayed average sedimentation rates (1-2m/year).In the remaining part of the shoreline, the erosion process was weak, with a rate of 1-2m/year, occurring in 420 transects, distributed in the mangrove forest along Duyen Hai street.Table 9. Statistical calculation EPR of Long Hoa commune during 2015 -2018

Figure 12 .
Figure 12.EPR map of Long Hoa commune in the period 2015 -2018 Most of the shoreline of Long Hoa during this period shifted < 14.7m (in 903 transects).The sections of shoreline that were eroded or accreted at high rates were those with shoreline displacement from 22 m to 80.6 m.

Figure 13 .
Figure 13.SCE map of Long Hoa commune in the period 2015 -2018

Figure 14 .
Figure 14.EPR map of Long Hoa commune in the period 2018 -2020

Figure 15 .
Figure 15.SCE map of Long Hoa commune in the period 2018 -2020

Figure 16 .Figure 17 .
Figure 16.EPR map of Long Hoa commune in the period 2020 -2022 Although both erosion and accretion rates exceeded 2 m/year, most of the shoreline between 2020 and 2022 had moderate displacement (from 7.9 m to 15.7 m).The places with large and even the maximum displacement (107.1 m) were located e.g. from the end of Nguyen Van Manh street to the west along the town or in some alluvial areas of mangrove forests affected by sea currents and sea dykes.Table 14.Statistical calculation SCE of Long Hoa commune during 2018 -2020 Length of shoreline change between 2 times (meters) Total number of transects 0 < SCE ≤ 7.9 398 7.9 < SCE ≤ 15.7 528 15.7 < SCE ≤ 23.6 311 23.6 < SCE ≤ 31.4 220 31.4 < SCE ≤ 107.1 264

Table 2 .
The information of the satellite images

Table 9 .
Statistical calculation EPR of Long Hoa commune during 2015 -2018

Table 10 .
Statistical calculation SCE of Long Hoa commune during 2015 -2018

Table 11 .
Statistical calculation EPR of Long Hoa commune during 2018 -2020

Table 12 .
Statistical calculation SCE of Long Hoa commune during 2018 -2020

Table 13 .
Statistical calculation EPR of Long Hoa commune during 2020 -2022