A DSAS-based study of central shoreline change in Jiangsu over 45 years

: A large-scale sand ridge group is distributed in the central Jiangsu coastal area, and a deposition muddy sea bank was developed in the nearshore area. Quantitative monitoring of coastline changes is of great significance for tidal beach development and protection. The shorelines of the central coast of Jiangsu within six periods (1973 – 2018) were extracted in this study, and their length changes over the years were analyzed. The Digital Shoreline Analysis System (DSAS) was employed to generate a cross section perpendicular to the baseline and calculate the linear regression rate (LRR) of the shoreline, changes in end point rate (EPR), and net shoreline movement (NSM), based on which the shoreline change features were analyzed. The DSAS results indicated that the shorelines in the study area maintained fluctu- ating growth and presented a continuous advancing trend towards the sea. From the changes in shoreline evolution distance during 1973 – 2018, the advancing shorelines in the study area accounted for over 50% of the total shorelines and presented first rising and then declining trends with the period of 2003 – 2013 taken as the time boundary. The average shoreline change rate was 207 m/year, and the periods with the highest change degrees were 1983 – 1993 and 1993 – 2013. The shoreline change tended to be stable during 2013 – 2018, and only a few estuaries and ports underwent obvious erosion and sedimentation.


Introduction
A shoreline not only marks a land-water boundary in coastal areas but also contains rich environmental information. Changes in the shoreline will have a direct impact on mudflat resources in intertidal zones and coastal environments, leading to changes in coastal resources and eco-environmental processes, as well as changes for the people who live in coastal areas (Xu et al. 2013;Wu and Hou 2016;Wang et al. 2016;Donadio et al. 2018;Daniel et al. 2019;Ding et al. 2019). Impacted by natural factors and human activities, shorelines are constantly changing (Mohanty et al. 2017). The abrupt shoreline changes have brought about contradictions and difficulties in the economy, society, ecology, and environment of coastal areas worldwide (Zhang et al. 2015), thus strengthening the research on shoreline monitoring and shoreline changes will be of critical significance.
The monitoring of coastal areas has always been a challenge for national development and environmental protection, where fundamental research on shoreline changes is considered essential (Rasuly et al. 2010;Boye et al. 2018). Information related to the location, direction, and geometric shape of shorelines is also significant for autonomous voyage, geological exploration, monitoring and modeling of coastal erosion, and inspection and management of coastal resources (Kuleli et al. 2011). It is not only difficult but also timeconsuming for traditional ground survey methods to conduct shoreline monitoring, and they are not practical for large-area monitoring. The remote sensing data analysis method can relieve these drawbacks to a great extent and has been quoted in many studies (Zhang et al. 2014;Shen et al. 2008). For instance, Darwish et al. (2017) used the remote sensing and geographic information system to detect and analyze shoreline and landform changes in the Nile Delta and found that the coastline's geomorphology greatly changed during this time period, especially at Damietta and Rosetta promontories, which were highly eroded after the construction of the Aswan High Dam. Besset et al. (2019) analyzed marginal shoreline change in the Mekong River Delta and its relationship with mangrove forests via remote sensing and statistical approaches, and they found that mangroves cannot accomplish their land-building and coastal protection roles under conditions of decreasing sediment supply and prevailing erosion. Lu and Yang (2012) used remote sensing images to continuously monitor shoreline changes in the Nantong segment located in Yancheng of Jiangsu Province over the years and concluded that the coastline of this section has rapidly advanced to the sea and the area along the beach has been increasing. Yue and Liu (2018) used RS and GIS technologies to analyze and conclude that the shoreline of the Yellow River Estuary continued to evolve from 1977 to 2017, and the land-use change appeared as a natural type instead of an artificial type.
As a plug-in of ArcGIS software, the digital shoreline analysis system (DSAS) was developed by the United States Geological Survey to analyze temporal-spatial shoreline change rates William 1994a, 1994b;Thieler et al. 2009). As an automatic computation method used to calculate the temporal-spatial shoreline change rate together with a shoreline prediction model, it has been extensively applied in both domestic (China) and foreign research (Sheik and Chandrasekar 2011;Kale et al. 2019;Ataol et al. 2019). Early research (Lu and Yang 2012;Li et al. 2018;Chen et al. 2018) regarding central shorelines in Jiangsu either emphasizes shoreline and beach evolution analysis or has a short time span with few largescale analyses of shoreline evolution rate. Therefore, remote sensing images of the coastal segment from Sheyang Estuary to Xiaoyang Estuary in Jiangsu Province in six periods over 45 years  were selected in this study to close this gap. ENVI and ArcGIS software were used to extract the shoreline, and then DSAS was used to analyze the change characteristics of the shoreline in the six periods. These are conducive to understanding the dynamic changes and expected impacts of the shoreline, so as to provide the required baseline information for the development, management, and ecological protection of the coastal zone in Jiangsu Province.

Study area
The study area is the central coastal segment from the Sheyang Estuary in the north to the Xiaoyang Estuary in the south in Jiangsu Province (Fig. 1). Located in the subtropicalwarm temperate transition zone, this area has a remarkable monsoon climate, abundant precipitation, and the development of coastal saltmarsh vegetation. The silty tidal flat of the radiant sandbar coast on this coastal segment is very broad, with a strong hydrodynamic process. Offshore hydrodynamic forcing is controlled by the tidal amphidromic system in the South Yellow Sea and progressive tidal waves in the East China Sea (Wang et al. 1998). The silty tidal flats on the radial sandbar coast of this shore section are very wide, with strong hydrodynamic processes, and the siltation and erosion of the banks are very intense, and the coastline changes rapidly (Chen et al. 2018).
The ever-increasing land demand brought by great economic development in recent decades has continuously expanded the scale and accelerated the speed of coastal reclamation in Jiangsu, and the newly built seawalls are continuously advancing towards the sea (Tang 1991). In fact, sea reclamation projects in Jiangsu usually start from the local average highwater level, so the elevation of the mean high-water level is already lower than the mean high-water spring tide. The central coastal seawall in Jiangsu is located at the seaward side and becomes a real artificial shoreline (Zhao and Li 2016). For thousands of years, Jiangsu coastal areas have experienced great changes due to the swinging of shorelines, and shoreline changes have exerted direct effects on the survival and development of people living in coastal areas (Cai and Wu 2014). Analyzing shoreline changes in Jiangsu is of great significance to the management and eco-environmental protection of coastal areas. The shorelines in the study area during the six periods from 1973 to 2018 were extracted in this study based on Landsat images, followed by a quantitative analysis of the extracted shorelines using DSAS 4.3.

Data source and processing
Based on actual requirements and data availability principles, Landsat TM remotesensing images (Table 1) in the central coastal areas of Jiangsu Province in the six periods 1973, 1983, 1993, 2003, 2013, and 2018 were selected. The cloud cover of most images was approximately 0. All images used in this study were downloaded from the GS cloud website (http://www.Gscloud.cn). Remote sensing images were analyzed and processed using ENVI 5.1 and ArcGIS 10.2. In ENVI 5.1, these images were preprocessed (correction accuracy was controlled within 0.5 pixels). The scope of this study area was drawn using ArcGIS 10.2. By referring to auxiliary information such as topographic maps of the study area, sea charts, and field survey data, the location of the shoreline boundary was comprehensively judged according to the tone, texture, spatial form, and distribution of different shoreline types on remote-sensing images as well as shoreline features and tide situation at image shooting time, and historical shoreline information. In the end, visual interpretation and automatic interpretation were combined to form linear vector layers and edit a shoreline attribute database.

Construction of feature classes of DSAS cross section
The shoreline vector layers were extracted using ArcGIS 10.2, and a specific shoreline attribute database satisfying the DSAS requirements was generated. After the satisfactory attribute database was obtained, the DSAS application program needed an arbitrary baseline to construct a cross section, and the baseline drawing methods mainly included three types: (1) start from a new feature class, (2) buffer an existing shoreline, and (3) use a preexisting baseline (Thieler et al. 2009). To properly consider the past change modes in this area, the shoreline in 1973 was selected to generate a buffer zone, and the remaining part of the buffer zone was deleted to form the baseline for the digital shoreline computation. A total of 159 equally spaced cross sections, which were perpendicular to the baseline and intersected with all shorelines, were generated through the baseline (Fig. 2).

Index selection
As required by analytical researches on shorelines in this study area, shoreline length change intensity (LCI) index, end point rate (EPR) index, net shoreline movement (NSM), and linear regression rate (LRR) index were employed in this study to calculate shoreline changes. The negative values of EPR, LRR, and NSM indicate landward erosion of the shoreline, and positive values represent seaward silting of the shoreline William 1994a, 1994b;Thieler et al. 2009). It is noteworthy that the NSM calculation gave a variable quantity relative to time, and LRR and EPR were used to calculate the change rate.

Length change intensity (LCI)
To objectively compare temporal-spatial differences in shoreline length change rates in different periods, the average change percentage of the shoreline length within a time period was used to express the shoreline change intensity (Xu et al. 2016), and the specific calculation formula is as follows: where LCI ij is the length change intensity of the shoreline in a geomorphic unit from year i to year j; L i and L j are shoreline lengths in year i and year j, respectively; negative LCI ij value means shortened shoreline and positive value represents lengthened shoreline. The greater the |LCI ij | value, the higher the shoreline change intensity.

Linear regression rate (LRR)
A linear regression rate-of-change statistic can be determined by fitting a least-squares regression line to all shoreline points for a transect, which is as follows: where L is a dependent variable denoting the spatial location of the shoreline, x is an independent variable of year, m is the slope, and b is the intercept, expressing the L change corresponding to x change per unit, namely LRR.

End point rate (EPR)
The end point rate (EPR) was calculated by dividing the distance of shoreline movement by the time elapsed between the oldest and the most recent shoreline. Different short-term (1973-1983, 1983-1993, 1993-2003, 2003-2013, and 2013-2018) change intervals were analyzed in this study to determine whether shoreline erosion and sedimentation in the study area were accelerated or decelerated between time spans.
where L 1 and L 2 are the distances from the shorelines in the most recent year and farthest year to the baseline, respectively; t 1 and t 2 are the time intervals between the most recent year and the farthest year.

Net shoreline movement (NSM)
The net shoreline movement (NSM) is the distance between the oldest and youngest shorelines for each transect and denotes distance but not rate.

Overall shoreline change features
The study results indicated that the overall coastal segment from Sheyang Estuary to Xiaoyang Estuary presented a seaward advancement trend during 1973-2018, and the shorelines were under a silting state, with an average silting rate of 207 m/year (Fig. 3). From the temporal change, the average change rate of the shorelines in the study area experienced three different phases: declining, rising, and declining. From 1973From -1983From until 1983From -1993, the average change rate of the shorelines mainly declined, that from 1983-1993 to 1993-2003 rose, and that from 1993-2003 to 2013-2018 declined by a large margin again and reached the minimum value over the years. The shoreline length in the study area presented a fluctuating rising trend: the shoreline length during 2013-2018 declined slightly, where the growth rate during 1993-2003 was the highest, and the average growth rate over the 45 years was 780 m/year. The LCI was the maximum during two periods (1993-2003 and 2003-2013), whereas those in the other three periods (1973-1983, 1983-1993, 2013-2018) were small. The composition of the proportions of advancing and receding shorelines in the entire study area presented significant staged changes. On the whole, the proportion of advancing shorelines first rose and then declined with 2003-2013 taken as the turning point, and that of receding shorelines was contrary to that of advancing shorelines. The proportions of advancing and receding shorelines were the most balanced during 2013-2018, but in other periods, the proportion of advancing shorelines was far higher than that of receding shorelines.

Shoreline length change intensity
The total shoreline length in the study area presented a rising trend increasing from 1973 to 2018 (Fig. 4a). Before 1993, the shoreline length change LCI was relatively steady, and that after 1993 was under a rising trend, where the shoreline length from 1993 to 2003 increased rapidly from 241.20 km in 1993 to 282.20 km in 2003, and the total length increased by 41 km. The growth of total shoreline length was slowed during 2003-2013, and the total length in 2013 was 293.07 km, which was an increase of 10.87 km in comparison with that in 2003. Year 2013 taken as a turning point, the total shoreline length presented a declining trend declining during 2013-2018 (from 293.07 km in 2013 to 278.79 km in 2018), and the total shoreline length was reduced by 14.28 km.
In the 45 years from 1973 to 2018, the shoreline length change intensity LCI in the study area presented an opposite change trend with 2003-2013 taken as the turning point (Fig. 4b). During 1973During -1983During and 1983During -1993, the shoreline length change intensity LCIs grew slowly, at 0.05% and 0.15%, respectively, so the shoreline length changed little, and coastal development activities in the study area were still in an initial phase in the two periods. The periods 1993The periods -2003The periods and 2003The periods -2013 were rapid growth periods of shoreline change, with intensity LCIs of 1.7% and 3.8%, respectively, and during the two periods, the coastal economy in Jiangsu was in a rapid development phase, where land use types in coastal areas went through great changes, and the expansion of urban land use, sea reclamation, and construction of salt pans and ports were the primary factors causing shoreline changes. The period 2013-2018 was a rapid declining phase of shoreline length change intensity LCI. During this period, the shoreline change intensity LCI declined from 3.8% to 0.49%, mainly because after 2013, coastal development and construction tended to be complete and slowed down in Jiangsu, and most shoreline types transited into more flat and straight artificial shorelines (Chen et al. 2018). Therefore, the shoreline length change intensity LCI was small from 2013 to 2018.

Net shoreline movement
When the NSM value was positive, the shoreline presented seaward silting, it with an advancing shoreline. However, if it was negative, the shoreline eroded towards the land, and it was a receding shoreline. It could be obtained through the analysis that the shorelines in the study area advanced towards the sea by 189.42 m every year during 1973-2018, where 99% of the shorelines were under seaward silting state, so they were advancing shorelines (Fig. 5). The maximum NSM of shorelines was 14 886.08 m, and the shoreline movement largely occurred near the Simaoyou Estuary. The maximum landward NSM was 262.24 m, and it largely happened near Sheyang Estuary. From the overall change trend, the coastal segments experiencing great NSM changes during 1973-2018 were mainly located in the north of the Simaoyou Estuary, between the south of the Simaoyou Estuary and the Wanggang Estuary, between the south of the Liangduo Estuary and the north of the Fangtang Estuary, and between the south of the Fangtang Estuary and the Beiling Estuary, but other coastal segments underwent minor changes in NSM values (Fig. 5a). The primary cause for the changes was that with socio-economic development, human demand for resources increased. Except for some estuary regions, most tidal flat wetlands in the central coastal areas of Jiangsu were reclaimed and developed into salt pans, ports, and cities, thus causing continuous seaward silting of the shorelines. Between the Xinyang Estuary and Doulong Estuary was a national-level core natural reserve of rare birds, so this shoreline changed to the minimum extent. From the proportion change of advancing and receding shorelines in different periods, the advancing shorelines in the study area accounted for over 50% of the total shorelines during 1973-2018 (Fig. 6), except that the proportion during 2013-2018 was 55%, those during 1973-1983, 1983-1993, 1993-2003, and 2003-2013 exceeded 65%, with 69%, 78%, 89%, and 94%, respectively. The proportion of advancing shorelines first rose and then declined with 2013-2013 taken as the turning point, rising from 69% (1973-1983) to 94% (2003-2013) and then declining to 55% (2013-2018). Among the receding shorelines, the proportions of receding shorelines only during 1973-1983 and 2013-2018 exceeded 30%, 31%, and 45%, respectively, and those in other periods were below 25%, whereas that during 2003-2013 was the lowest at only 6%. The proportion of receding shorelines in the study area first declined and then rose, progressively declining from 31% (1973-1983) to 6% (2003-2013), and then rose to 45% (2013-2018).

Temporal-spatial distribution of coastline change rates
If the EPR value is positive, the shoreline is in a silting state, and the negative EPR value expresses the erosion state. According to an analysis of the calculation results of cross sections, the shorelines in the study area presented significant staged and regional change trends over the 45 years from 1973 to 2018.
Regarding the spatial distribution of shoreline change rates (Fig. 8), the coastal segments with the highest shoreline change rate during 1973-1983 were mainly located in the center and south of the study area, and the situation during 1983-1993 was largely opposite to that during 1973-1983. For the period 1993-2003, except for a minority of coastal segments, the shoreline change rates of most coastal segments in the study area were high, and the erosion and silting changes in the entire study area were quite remarkable. The key coastal segment in the aspect of shoreline change during 2003-2013 migrated to the south of the study area. Except for the estuary regions during 2013-2018, the change rate of coastal segments in the whole study area was low, where the shoreline erosion and silting change tended to be steady.
From the specific location where shoreline erosion and silting change occurred (Fig. 8), the silting coastal segments were mainly located from Doulong Estuary to Wanggang Estuary, from Liangduo Estuary to Beiling Estuary, and nearby Xiaoyang Estuary during 1973-1983 (Fig. 8a), and the average silting rate was 297 m/year. The eroded coastal segment was mainly near the Sheyang Estuary, while the erosion and silting changes in other coastal segments were not obvious. During 1983During -1993 (Fig. 8b), the silting coastal segments in the study area were mainly located in ports and nearby estuaries; therefore, the main reasons for shoreline silting at the estuary were massive sediment accumulation in the river and large-scale port construction; during this period, the shoreline erosion change was not evident in the study area. During 1993-2003 (Fig. 8c), the silting coastal segments were largely located from Sheyang Estuary to Xinyang Estuary, from Doulong Estuary to Simaoyou Estuary, and from Chuandong Estuary to Liangduo Estuary, and the shoreline erosion largely took place in the northmost of Sheyang Estuary and nearby Fangtang Estuary. The main causes of shoreline silting included large-scale coastal development and accelerated tidal flat reclamation, which led to a large scope and high intensity of shoreline change. The main coastal erosion segments in the study area were located in the north of Sheyang Estuary and at Liangduo Estuary and Fangtang Estuary during 2003-2013 (Fig. 8d), while the other coastal segments mainly experienced silting changes. The shoreline erosion and silting change during 2013-2018 (Fig. 8e) reached equilibrium, and the main erosion and silting coastal segments were located at corresponding ports and estuaries, but the erosion and silting phenomenon was not obvious in other areas.

Discussion
The coastline has unique geographical, morphological, and dynamic characteristics. It is the most important geographical element for describing the boundary between land and sea (Dar et al. 2009). It is one of the 27 surface elements identified by the International Geographic Data Committee (Mujabar and Chandrasekar 2013). With global warming and rising sea levels, more than half of the world's beaches are retreating due to erosion. However, since the 20th century, the economic center of the world's coastal countries has shifted to coastal areas, and more than half of the world's population lives within 100 km of the coastline (Primavera. 2006). The coastal zone has become the most active and concentrated area of human economic activity, and increasingly saturated and crowded living and production spaces have forced some coastal countries to develop land in the form of sea reclamation, causing some coastlines to expand to the sea on a large scale against the trend of coastal erosion on rising sea level background. The speed and intensity are much greater than in the natural state, and dramatic changes in the coastline have brought about economic, social, ecological, and environmental contradictions and difficulties in coastal areas around the world (Liu and Jezek 2004). Therefore, the study of coastline dynamics is the basis of coastal environment monitoring, resource development, and management, and it is helpful to deepen the understanding of coastal environment and ecological processes and promote sustainable management and development of coastal resources and the environment.
This study uses ArcGIS 10.2 to process and calculate the quantity structure table of various landscape types in the coastal zone from 1973 to 2018 (Table 2). From the table, we can see that the quantity structure of different land types in the coastal zone of central Jiangsu has changed significantly in the past 45 years, in which the natural landscape types and areas have gradually decreased and the artificial landscape areas have been increasing, most of the natural landscape has been replaced by artificial landscape types such as breeding ponds, farmland, and buildings. From this, we can conclude that in recent years, development activities such as artificial reclamation, port development, and port city construction have become important means for expanding production and living spaces in Jiangsu coastal areas, and have promoted the rapid economic development of coastal areas in Jiangsu Province. Such developmental activities have a profound impact on the evolution of the continental coastline and is the main reason for the continuous siltation of the coastal coastline of central Jiangsu. Unreasonable development of the coastal tidal flats and offshore resources not only changed the basic form and spatial pattern of coastline of central Jiangsu, but also caused ecological problems such as resource shortage, environment deterioration, and a great impact on the ecological environment of coastal zone. Therefore, in this study, we selected six remote sensing images from 1973 to 2018 to research the status quo of the development and utilization of coastal continental coastline in central Jiangsu to provide relevant data support for the scientific guidance of coastal zone utilization and protection, so as to achieve sustainable economic development in coastal areas.

Conclusions
We used RS and GIS technology to analyze the shoreline length and shoreline change characteristics in different temporal and spatial ranges from 1973 to 2018 along the central Jiangsu coast using indices such as shoreline length change intensity, shoreline linear regression change rate, endpoint change rate, and net shoreline movement to explore shoreline evolution in the study area, comprehensively reflecting the comprehensive characteristics of length, tortuosity, temporal and spatial changes, and development and utilization of the continental coastline of the central Jiangsu coast. The research results show that due to the economic development of the coastal zone, the coastline of the study area as a whole is propelled towards the sea, and the coastal zone is in a silting state.
The length of the shoreline in the study area shows a fluctuating and rising trend, with an average growth rate of 780 m/year. For shoreline length change intensity, the growth rate of shoreline length was the highest during 1993-2003 and 2003-2013, and the shoreline length changed little in the other three periods. From the temporal change, the average shoreline change rate in the study area experienced three different phases: decliningrising-declining. From 1973From -1983From to 1983From -1993, the average shoreline change rate mainly decreased, that from 1983-1993 to 1993-2003 mainly rose, and that from 1993-2003 to 2013-2018 presented a declining trend by a large margin and reached the minimum value over these years. In terms of spatial changes of shoreline change rate, the coastal segments with the highest shoreline change rate were mainly located in the center and south of the study area during 1973-1983, while the situation during 1983-1993 was on the contrary. Most coastal segments in the study area experienced notable erosion and silting changes during 1993-2003. The emphasis of shoreline change was migrated to the south of the study area during 2003. During 2013, the shoreline change rates in the whole study area were small except for estuary regions, and shoreline erosion and silting change became steady.