Study on the Coastline Evolution in Sopot (2008–2018) Based on Landsat Satellite Imagery

The coastline is the boundary between the water surface in a reservoir or watercourse and the land, which is characterised by high instability and functional diversity. For these reasons, research on coastal monitoring has been conducted for several decades. Currently, satellite images performed with synthetic aperture radars (SARs) are used to determine its course and variability together with high-resolution multispectral imagery from satellites such as IKONOS, QuickBird, and WorldView, or moderate-resolution multispectral images from Landsat satellites. This paper analysed the coastline variability in Sopot (2008–2018) based on Landsat satellite imagery. Furthermore, based on multispectral images obtained, it was determined how the beach surface in Sopot changed. Research has shown that the coastline keeps moving away from the land every year. This was particularly noticeable between 2008 and 2018 when the coastline moved on average 19.1 m towards the Baltic Sea. Moreover, it was observed that the area of the sandy beach in Sopot increased by 14 170.6 m2, which translates into an increase of 24.7% compared to 2008. The probable cause of the continuous coastline shift towards the sea and the increase of the beach surface is the oceanographic phenomenon called tombolo, which occurred in this area as a result of the construction of a yacht marina near the coast.


Introduction
The coastline is a dynamically changing boundary between land and water [1], characterized by instability and functional diversity depending on the region [2]. This boundary is important for the ecological and economic policies of coastal states because areas located in the coastal zone are rich in natural resources. This results in about 50% of the world's population being currently settled in zones located within 100 km from the coastline [3]. Therefore, coastal monitoring research is currently being conducted in areas such as deltas and estuaries [4,5], wetlands [6], bays [7] and other geographical forms along the coast [8].
Coastline formation is the result of many factors (both anthropogenic and natural), which include, for example, sea erosion, water level rise [9], sediment transport [10], earthquakes [11], sea currents, tides, waving, flooding of coastal areas [12], rise in ocean temperature and acidity levels [13], sea water

Materials and Methods-Vectorization of Satellite Images Taking into Account Hydrological Data
ArcGIS software was used for data processing. Satellite images from Copernicus and Landsat were used for the analyses and constituted a source of data on the coastline course. For these images to be cartometric, they had to be given in a specific coordinate system. Therefore, taking into account the scale of these images, the first stage of works was undertaken, i.e., georeferencing [47] (Figure 1). This process was based on the determination of control (reference) points for both the vector layer (pier model), drawn on the basis of a topographic map from the Geoportal website and a satellite imagery. Once these were determined, raster calibration was conducted, i.e., transformation and recording information about the given spatial reference. For research purposes, the pier model and satellite image were fitted into the Universal Transverse Mercator (UTM) system for publishing nautical charts [48]. to be cartometric, they had to be given in a specific coordinate system. Therefore, taking into account the scale of these images, the first stage of works was undertaken, i.e., georeferencing [47] (Figure 1). This process was based on the determination of control (reference) points for both the vector layer (pier model), drawn on the basis of a topographic map from the Geoportal website and a satellite imagery. Once these were determined, raster calibration was conducted, i.e., transformation and recording information about the given spatial reference. For research purposes, the pier model and satellite image were fitted into the Universal Transverse Mercator (UTM) system for publishing nautical charts [48]. Prior to the analyses, a reference line was established against which the coastline shifts from 2010-2018 were calculated against the 2008 shoreline. It was assumed to be 800 m long and to cover both the left and right sides of the pier in Sopot. Subsequently, the coastlines were drawn based on satellite images from the Google Earth Pro platform to start the vectorization process [47] (Figure 2).
The waving range and sea levels can cause a coastline shift. Taking into account that the waving range is very small along the Polish Baltic coast and the waves are short and steep [49], it was decided not to take it for the analysis of the shoreline variability in Sopot. However, it was decided to analyse the differences in sea levels ( Table 1). The data were obtained from the Institute of Meteorology and Water Management-National Research Institute (IMGW-PIB) gauging station in Gdynia, which is nearest to the place of measurements [50]. Based on the sea-level data from the gauging station in Gdynia, it can be seen that in selected years the averages and medians are similar. The difference between the minimum and maximum Prior to the analyses, a reference line was established against which the coastline shifts from 2010-2018 were calculated against the 2008 shoreline. It was assumed to be 800 m long and to cover both the left and right sides of the pier in Sopot. Subsequently, the coastlines were drawn based on satellite images from the Google Earth Pro platform to start the vectorization process [47] (Figure 2).
The waving range and sea levels can cause a coastline shift. Taking into account that the waving range is very small along the Polish Baltic coast and the waves are short and steep [49], it was decided not to take it for the analysis of the shoreline variability in Sopot. However, it was decided to analyse the differences in sea levels ( Table 1). The data were obtained from the Institute of Meteorology and Water Management-National Research Institute (IMGW-PIB) gauging station in Gdynia, which is nearest to the place of measurements [50]. Based on the sea-level data from the gauging station in Gdynia, it can be seen that in selected years the averages and medians are similar. The difference between the minimum and maximum sea-level value in a given month and year is less than 1 m. Therefore, it is not necessary to relate the coastlines to the mean level of the Baltic Sea for the Kronstadt mareograph. sea-level value in a given month and year is less than 1 m. Therefore, it is not necessary to relate the coastlines to the mean level of the Baltic Sea for the Kronstadt mareograph.

Results
This chapter analyses changes in the beach surface and the coastline variability in Sopot in the years 2008-2018. As a measure of change assessment, an increase or decrease of the beach area was assumed compared to 2008 (m 2 ) and the standard deviation of the distance between the 2008 shoreline and the land-water boundary determined between 2010-2018 (m).

Results
This chapter analyses changes in the beach surface and the coastline variability in Sopot in the years 2008-2018. As a measure of change assessment, an increase or decrease of the beach area was assumed compared to 2008 (m 2 ) and the standard deviation of the distance between the 2008 shoreline and the land-water boundary determined between 2010-2018 (m).

Analysis of Beach Surface Changes in Sopot
After completion of the vectorization for all the coastlines and assessment of factors that may affect their location, the beach surface changes in Sopot were analysed. To do so, it was necessary to determine the boundaries of the area between the beginning of the beach and the shoreline. Then they had to be converted to polygons. Next, with the Calculate Geometry tool, available in ArcGIS software, information about the beach surface in specific years was obtained (Figure 3).

Analysis of Beach Surface Changes in Sopot
After completion of the vectorization for all the coastlines and assessment of factors that may affect their location, the beach surface changes in Sopot were analysed. To do so, it was necessary to determine the boundaries of the area between the beginning of the beach and the shoreline. Then they had to be converted to polygons. Next, with the Calculate Geometry tool, available in ArcGIS software, information about the beach surface in specific years was obtained ( Figure 3).

Analysis of Beach Surface Changes in Sopot
After completion of the vectorization for all the coastlines and assessment of factors that may affect their location, the beach surface changes in Sopot were analysed. To do so, it was necessary to determine the boundaries of the area between the beginning of the beach and the shoreline. Then they had to be converted to polygons. Next, with the Calculate Geometry tool, available in ArcGIS software, information about the beach surface in specific years was obtained ( Table 2). Based on the results obtained, we can see that the land area adjacent to the waterbody near the pier in Sopot had the smallest surface (57 415.2 m 2 ) in May 2008, i.e., before marina construction started. The beach area was the largest (71 585.8 m 2 ) in May 2018, i.e., 7 years after the marina was commissioned. Note that over the 10 years the beach increased by 14 170.6 m 2 , which is a 24.7% growth compared to 2008. Moreover, from Table 2 it follows that the beach surface in Sopot kept growing every year, except for 2012. Waving, visible in the satellite image of May 2012, which worsened the coastline visibility, is the probable cause ( Figure 3). The next factor which contributed to the coastline visibility on satellite image in 2014 are longshore, rip and undertow currents [51].

Analysis of Coastline Variability in Sopot
Next, coastline variability in the years 2008-2018 was assessed. The distance between the landwater boundary from 2008 and remaining shorelines determined for the years 2010-2018 was taken as a measure of the assessment of changes. To determine them, the Digital Shoreline Analysis System (DSAS) extension of the ArcGIS software was used, which allowed the statistics of changes in the land-water boundary to be calculated based on time series [52]. The calculation procedure starts with defining a reference line in the form: where: X RL , Y RL -rectangular coordinates PL-UTM of the points that determine the reference line, b-slope of the reference line, a-x-intercept of the reference line. The distance from the reference line to the coastline was then calculated. For this purpose, straight lines were drawn perpendicular to the reference line, which can be described by the formula (Figure 5a): where: where:   The Formula (2) does not provide the numerical value of parameter a i because it depends on the distance between successive perpendicular lines. It was assumed for this study that the distance will be 1 m.
The distances between the reference line and the coastline (d i ) were calculated from the coordinates of these lines intersecting with the perpendicular line drawn to the reference line (Figure 5b): where: X RL i , Y RL i -rectangular coordinates PL-UTM of the reference line intersection points with the i-th line perpendicular to it, X C i , Y C i -rectangular coordinates PL-UTM of the coastline intersection points with the i-th line perpendicular to the reference line.
After calculating the distance between the reference line and the coastline from 2008-2018, the spatial and temporal variability of the land-water boundary at the pier in Sopot was determined. For this purpose, distances were calculated between the 2008 shoreline and the land-water boundary determined for 2010-2018 (∆d i ) using the following formula ( Figure 5c): where: where: ∆d-arithmetic mean of the distances between the 2008 shoreline and the land-water boundary determined for 2010-2018, n-the number of lines perpendicular to the reference line. From Figure 6 it follows that the coastline keeps moving away from the land every year, with the exception of the years 2012, 2014 and 2017. Note that the standard deviations of the distance between the 2008 shoreline and the land-water boundary between 2011 and 2012 are very similar. This may be due to the fact that the marina construction, which is the main reason for the formation of tombolo in Sopot, was commissioned in mid-2011. Thus, within a year of this event, there were no significant changes in the coastline shape. The value of the σ ∆d in 2014 is smaller by 1.6 m than for the previous year. It is caused by wave motion [53,54] visible on the satellite image from the Google Earth Pro platform (Figure 4). Strong waves that occur transverse transport to the shore contribute to the sediment movement towards the sea [55]. Another factor that can affect the shoreline shape are undertow currents [51]. Moreover, it should be emphasised that the standard deviations of the distance between the 2008 coastline and the land-water boundary between 2017 and 2018 are almost identical, 19.2 and 19.1 m respectively. This may have been due to silting works carried out in the area to extract about 6 000 m 3 of sand from the beach around the pier in 2017 [56].

Discussion
The paper discusses the analysis of coastline variability in Sopot (2008-2018) based on Landsat satellite imagery. Apart from determining the spatial and temporal shoreline variability, the beach surface changes in Sopot were also analysed.
The conducted research indicates that the most effective and optimal method of determining the coastline course is currently an analysis of satellite images taken using SAR and multispectral imagery from satellites such as IKONOS, QuickBird, WorldView or Landsat. It allows accurate (less than 1 m), fast and large-scale determination of the land-water boundary. In addition, this method enables the spatial and temporal shoreline variability (from 10 to 30 years) to be determined thanks to extensive databases which contain satellite images, even decades old. However, it should be noted that historical satellite imagery are characterized by low resolution and may miss hydrological data, e.g., sea levels or waving, which are necessary to precisely determine the land-water boundary [57].
The presented results are clearly indicative of a Sopot coastline shift. This was particularly noticeable between 2008 and 2018 when the shoreline moved on average 20 m towards the Baltic Sea. As the coastline moves away from the land, a continuous increase in the beach surface in Sopot was observed. Based on conducted analyses, the sandy area increased by 14,170.6 m 2 , which is 24.7% more than in 2008. This is very apparent on the right side of the pier, where a strip of land called tombolo [43][44][45][46] has formed between the shoreline and the yacht marina. The tombolo phenomenon forms as a result of the interaction between the hydrotechnical structure and hydrodynamic processes [58]. One of the elements that changes its properties when encountering an obstacle (e.g., breakwater) is waving. The wave coming to the edge bends around the obstacle, causing change in the direction of wave propagation [59]. Refraction is also of key importance for the development of the tombolo phenomenon. It consists in the fact that during the wave refraction most of the energy transported by it dissipates, while the remaining part causes the formation of currents. In the coastal zone, the strongest of them is the longshore current. All the aforementioned factors cause sediment transport. Sediments transported by the longshore current accumulate on the up-current side of the building, while their erosion occurs on the down-current side of the building [55]. This phenomenon causes many negative effects on both the aquatic and human environments, including navigational hazards, the blooming of cyanobacteria and other bacteria, and moving sand away from other places, e.g., from the Orłowo Cliff. Therefore, it is crucial to monitor coastline variability in this waterbody to prevent its shifting.