Coastal morphodynamic analysis in Buleleng Regency, Bali—Indonesia

Sediment as erosion product can affect shorelines, making sediment transport a key process to consider in coastal and shoreline management. Field surveys and secondary data can identify where suspended matters are distributed and deposited to analyze sediment uniformity factors: beach morphology and materials. This research set out to determine the Buleleng Regency's morphodynamic aspects based on the coastal landscape's physical characteristics and the processes acting upon each sediment cell. Field observations were conducted at five stations, from Tukad Gerokgak to Tukad Saba estuary. Jaelani's spectral transformation has been applied to analyze Total Suspended Solids using Sentinel 2A imagery. The laboratory test results of grain-size samples were processed on GRADISTAT, then the depositional environment and sediment transport direction were determined from average grain size, standard deviation, skewness, and kurtosis. Shoreline change, an indicator of coastal morphodynamics, was mapped from Landsat images in 2000, 2008, and 2019 using the Digital Shoreline Analysis System. Statistical analysis on GRADISTAT provided details on depositional environment and sediment transport and deposition based on grain-size distribution. Results indicate poorly sorted medium grain size: gravel (stone) to coarse sand, making up the sediment population from Tukad Gerokgak to Tukad Saba. Generally, sediment is deposited toward coarse, even very coarse, grain on a strongly sloping beach, and there is a high likeliness of sediment accretion. Identified morphodynamic characteristics suggest that the coastal landscape needs structural mitigation to overcome the accelerating impact of human activities and physical processes.


Introduction
Coastal areas form at the interface of three major natural systems on earth: atmosphere, ocean, and land surface, which continually change in response to human and natural forces in the form of both physical and non-physical processes, such as storms, currents, erosion, 1 3 and sedimentation (Weill and Tessier 2016;Mutaqin 2017;Fan et al. 2018;Arjasakusuma et al. 2021). Developing countries, including Indonesia, have reported the most cases of severe natural disaster impacts. Indonesia is particularly vulnerable to sea-level rise, which conduces to erosion and tidal floods (Marfai et al. 2008a, b;. Flood vulnerability is estimated to worsen in the next 30 years, especially in the coastal urban (Ward et al. 2013).
Changes in a coastal region are closely related to morphology, beach material, and acting process (Bird 2007). A coastal landscape is described through morphodynamic aspects that produce sediment characteristics, beach geometry, and shoreline change. Shoreline change is among the most dynamic processes occurring in it and results from longshore drift, extreme waves, or anthropogenic factors (Bagli and Soille 2003;Mutaqin 2017;Arjasakusuma et al. 2021). Physical processes and human activities have increasingly put pressure on coastal regions, and the latter always leave specific features that differ between regions depending on the scale of modification (Lentz and Hapke 2019). Coastal management that considers morphodynamic conditions due to the strong influence of geological control along the world's coasts is very important as a preparation to face changes and uncertainties in the future (Gallop et al. 2020). One of the morphodynamics is controlled by the sediment grain size (Trenhaile 2016) and understanding sediment transport is essential for effective management (Hooke et al. 1996). Coastal dynamics, including variations in shoreline changes (erosion or accretion) distribution of headlands and bays, are usually used to plan coastal management strategies (Montreuil and Bullard 2012;Mutaqin 2017;Arjasakusuma et al. 2021;Septiangga and Mutaqin 2021). The sediment cell concept is used as a planning unit for coastal management planning (Cooper and Pontee 2006;Montreuil and Bullard 2012;Gallop et al. 2020), especially when it comes to dealing with shoreline changes and their management plans, including shoreline restoration and mitigation efforts (Simon et al. 2016;Mushkin et al. 2016;Ramesh et al. 2021;Smith et al. 2021).
Abrasion and accretion are apparent in the northern coast of Bali Island in Indonesia, notably the Buleleng Regency. Erosion has impacted about 54.83 km or 45% of the regency's shoreline (Heliani et al. 2014). Sea-level rise predictions suggest that about 7.4% of its coast is at risk of inundation, especially Gerokgak and Seririt (Heliani et al. 2014)two densely populated districts (237 and 652 people/km 2 , respectively) with rapid physical development (Badan Pusat Statistik 2019). In one year (January-December 2019), there were 1,084,168 tourist arrivals, with 29% being international visitors (Badan Pusat Statistik 2020). Buleleng has grown as a sea transportation node that further generates movements from and to its cargo port in Celukan Bawang, a regional seaport in Sangsit, and a small-sized traditional port in Pengametan Sumberkima.
The shoreline can be divided regionally into several cells based on the sediment budget or net sediment transport (Syaefudin 2008). Sediment cell identification is carried out in several ways, such as through geomorphic features, littoral drift indicators, shoreline structures, field surveys, satellite imagery, or statistical models (Herman and Zhang 2015). Field surveys to observe beach profiles include beach type, slope, width, coastal constituent materials, erosion, abrasion, coastal plants, spit direction, and coastal structures (Syaefudin 2008). If field observations are not enough, sediment samples can be taken for laboratory analysis or collect hydro-oceanographic data (waves, currents, winds, bathymetry) for modeling the direction of sediment movement. The sediment cell approach to shoreline management is an effort to restore the coast naturally (soft engineering strategies). So that in addition to knowing the boundary of sediment movement patterns on the coast, it is also necessary to know the main cause of damage to the coastal environment. Van Rijn (2010) recommends handling coastal erosion by restoring the average sediment balance in the sediment cells. An understanding of the cycle of erosion, deposition, sources of sediment and waste, and transport routes is needed. Hooke et al. (1996) analyzed the sediment transport process, form input, flow, to sediment output, as a component in coastal management. Sediment movement is influenced by factors originating from land and sea, so according to Ouillon (2018), the geomorphological units in the sediment transport study are divided into 6, namely river basins, estuaries, or deltas, littoral-estuary transition, littoral zone, continental shelf, and deep sea. This study examines sediment transport using geomorphological units in the littoral zone (which can be reached from remote sensing images), including river estuaries.
Coastal area management requires an understanding of the system, including dynamics, interactions, environmental conditions, system sensitivity, and physical processes that shape coastal morphology . Although many sediment transport studies have adopted various approaches, the sediment cell has been reported to cover more effective administrative boundaries for shoreline management (Cooper et al. 2002;Collins and Balson 2007). The sediment cell concept approach in coastal management has been widely used. However, without integration with coastal processes and geomorphological responses, it is less comprehensive. So it is important to uncover other factors such as the size of the sediment grain, the role of the estuary, the relationship to the underlying geology, and changes in the shoreline. To improve the management of Buleleng coastal areas, it is necessary to consider the boundaries of sediment cells on the coast from the Tukad Gerokgak to the Saba estuary. New development of coastal morphodynamics assessment in situ with grain size integrated with techniques for extracting information with remote sensors, as well as the measurement of coastal characteristics, can be used for coastal morphodynamic analysis with better data at lower costs. This research has been designed to describe a coastal landscape based on the morphodynamic aspects developing in the Buleleng Regency parts.

Methods
The research area is the northern coast of Buleleng Regency, Bali, Indonesia, which administratively covers three districts: Gerokgak, Seririt, and Banjar ( Fig. 1). Gerokgak has the longest beach on the island, 76.89 km (Badan Pusat Statistik 2019). Per the 2018 data of the BPS-Statistics Indonesia, the coastline observed is 157.05 km in length and shows signs of abrasion at varying degrees.

Data and tools
The research collected primary data and records from several relevant agencies. Shoreline data were digitized from Landsat images captured on July 17, 2000, July 31, 2008, and April 1, 2019, and downloaded from the USGS website: https:// earth explo rer. usgs. gov. SPOT-7 Pansharpen was obtained from the Indonesian Institute of Aeronautics and Space (LAPAN), while the level-2A Sentinel products recorded on July 20, 2018, and March 27, 2017, were downloaded from https:// scihub. coper nicus. eu. Landsat is one of the most widely used satellite images for coastal dynamics studies (as a data source) because it is freely available and has long historical data series since 1984 (Pardo-Pascual et al. 2018). Similar studies on shorelines are Duru (2017) Wicaksono and Winastuti (2020). Landsat's main limitation is that the maximum spatial resolution it offers is only up to 30 m, restricting the detection of shoreline changes occurring in a smaller coverage (Pardo-Pascual et al. 2018). The three-year data sets used are Landsat imagery to determine long-term shoreline variations (for 20 years). Changes in the shoreline form the basis for determining sediment samples. There are three types of imagery (different scales) used, namely Landsat (for shoreline), Sentinel-2A (for TSS models), and SPOT 6 (for landcover). The three of them have different roles in obtaining information so that they are processed separately, although there are constraints on accuracy (Marfai et al. 2008a, b). Tide predictions were acquired from the Indonesia Geospatial Information Agency (BIG), and data on coastal buildings were from the Center for Coastal Research and Development, the Indonesian Ministry of Public Works and Housing. Bathymetry data were obtained from BIG through a hydrographic survey using the single beam echo sounding (SBES) method in 2015 and had been corrected for tides. Other supporting data like wave conditions were measured and documented from wave appearances in the field.
The research mapped shoreline changes from 20-year long data (2000,2009, and 2019) on the Digital Shoreline Analysis System (DSAS) and conducted observations and measurements in the field for morphodynamics study. DSAS facilitates shoreline mapping and calculation of rates of change (Thieler et al. 2009). Many research works have utilized this application for the same purpose, e.g., Duru (2017), Mutaqin (2017), Dewi (2019), Yulianto et al. (2019), and Wicaksono and Winastuti (2020). Satellite images recorded in the eastern monsoon were selected because they have little to no cloud cover. Likewise, the fieldwork was conducted in the same season because the wave conditions are less extreme than those in the western counterparts.

Data processing
a. Digital image processing for shoreline detection used multitemporal Landsat images with a 30 m resolution and geometric, radiometric, and atmospheric corrections.The Modified Normalized Difference Water Index (MNDWI) transformation with a threshold Xu (2006), Rokni et al. (2014), and Wicaksono and Wicaksono (2019) have justified the application of this method for shoreline identification and suggest that MNDWI gives good results when used for distinguishing between sea and land features in both built-up and open land (sand). b. Shoreline change was calculated using an additional plug-in on ArcGIS, i.e., DSAS.
DSAS contains statistical features useful for this purpose, namely Shoreline Change Envelope (SCE), Net Shoreline Movement (NSM), and End Point Rate (EPR). The formula of each statistical feature is presented below. SCE = the largest difference in the distance across all shorelines (in meters) NSM = the difference in distance between the oldest and most recent shorelines (in meters) EPR = the difference in distance between the oldest and most recent shorelines (in meters) divided by the length of time between the two (in years). c. The Jaelani algorithm is a spectral transformation index developed by Jaelani et al. (2015) to enable Sentinel 2A imagery in building models associated with Total Suspended Solids (TSS), as shown in Eq. (2).
d. Sediment cells were delineated based on TSS analysis, differences in beach geometry, landcover, and the presence of barrier structures. Landcover was identified from SPOT-6 Imagery. e. Sediment samples for grain size analysis were tested in the laboratory, and the results were processed on GRADISTAT. Depositional environment and sediment transport direction were determined using several statistical parameters: the average grain size, standard deviation, skewness, and kurtosis. Their mathematical expressions are written in Table 1.

Fieldwork
Field observations were carried out from May 30 until June 2, 2019, at five stations, starting from the first station at the Tukad Gerokgak estuary to the last in the Tukad Saba estuary area (2) log TSS (mg∕L) = 1.5212 × (log (green)∕ log (red)) − 0.3698 Table 1 Grain size parameter equations. Source: Blott (2001) α sediment samples f frequency (in percent) m midpoint of each class interval (in metrics, mm, or phi units)

Mean
Sorting Skewness Kurtosis Fig. 2). The distribution of field samples takes into account the initial sediment cell boundaries (coastal structures, landcover, TSS, and beach geometry) as well as the dynamics of the largest shoreline for 20 years. Fieldwork was carried out under the same tidal conditions when the image is recorded. This stage observed and validated data processing results from the first stage and collected relevant primary data. Data collected during fieldwork include sediment, beach morphology (length, width, slope), and hydro-oceanography aspects (wave condition, wind speed, and wind direction). Beach slope was measured by using Abney level (Fig. 3). Wave conditions were visually observed; moreover, wind direction and speed were measured using a compass and anemometer. The research stages are shown in Fig. 4.

Sediment cells
Geological distribution and control have a major impact in shaping coastal morphodynamics (Klein and Menezes 2001;Benedet et al. 2004;Jackson and Cooper 2009;Short 2010;Scott et al. 2011;Loureiro et al. 2012;Gallop et al. 2020). Understanding the geological conditions of a coastal system is key in predicting the effects of the system's human-environment interaction and temporal changes (Park et al. 2009). Based on the Indonesian geological map, the stratigraphic rock layers found in the Buleleng  Regency are composed of breccia, lava, tuff, and lahar scattered almost in the entire regency. Suspected faults in the Gerokgak District area consist of two large faults that lie parallel to the west and east and are parts of the Pulaki Volcano Rock formation, which is made up of breccia and lava. There are two suspected horizontal faults in the western tip of Bali Island (the Prapat Agung Formation is predominantly covered by limestone, and the Palasari Formation consists of sandstone, conglomerate, and reef limestone). Also, there are two other faults around the Tejakula District, precisely between the Buyan Bratan and Batur Purba Formations (tuff and lahar deposit), and there is a rock stratification structure consisting of tuff and lava from the ancient volcanic rock formation, Buyan Bratan. Most of Buleleng is a hilly area stretching in the south and low-lying land (coast) in the north. Among the hills are several mountains that are no longer active. The type and nature of beach morphodynamics are sensitive to headland spacing, shape, wave obliquity, indentation ratio, nearshore morphology, and substrate control (Klein and Menezes 2001). The coast's shape is categorized as a bay, with tides, waves, and river discharges being the dominant acting processes. Grain-size distribution is strongly influenced by the type and availability of sediment source material and the processes involved (Folk and Sanders 1978;Mutaqin et al. 2021). Rock formation requires complex geological processes that have lasted for thousands to millions of years. It is estimated that seabed sediment type and distribution from Tukad Gerokgak to Tukad Saba represent geological complexity in parts of Buleleng and its surroundings. Major rivers carry erosion products to the sea, accumulating seabed sediment in Buleleng waters.
A coastal system consists of a number of units related to and associated with many sediment movement processes occurring at different temporal and spatial scales. Sediment cells are seasonal and annual water mass circulations, with wind or currents being the forces responsible for cell formation. They can be interpreted as accumulated nutrients or limited to sediments without nutrient content (Marfai et al. 2018). The sediment cell concept originates in a balance between sediment transport and wave energy-coarse sediment interaction nearshore that transports or deposits sediment at certain limits. Here, cells are associated with sand or gravel movements along the coast or nearshore in one cell that does not significantly affect adjacent cells (Motyka and Brampton 1993).
Sediment motion is not directly apparent on satellite images, but turbidity can be clearly traced. Turbidity level can describe the direction and distribution of suspended solids and identify deposition location, allowing the coast's uniformity limit to be observable (Khakhim et al. 2005). It also enables the zoning of suspended sediment movements. Coastal physical conditions can be recognized from sediment uniformity, particularly its appearance on a satellite image. TSS is sensitive to land input through river flows and displacement caused by sediment resuspension after erosion (Tarigan and Edward 2003). Sediment loads that enter the sea from river estuaries spread depending on river flow discharge, sediment load volume, current, wave, and tide. Upstream river flows carry sediment to the river mouth, and the rest is transported to the sea. There are differences in suspended sediment levels during floods and ebbs. Floods mean that tidal processes dominate the bay, contributing to higher suspended sediment concentration than ebbs (Alongi 1997). During the flood phase, river flows and inland tidal currents meet in the estuary, which accumulates and deposits TSS originating in the land and the sea at this water body.
Five sediment cells were identified from Tukad Gerokgak to Tukad Saba. Their boundaries are a combination of delineated TSS zones (modeling output is shown in Fig. 5a), beach geometry, and barrier structures. Despite the different physical processes, they are divided into interdependent cells (Dinas Kelautan dan Perikanan 2004). This research determined these boundaries by considering buildings' presence ( Fig. 5b) as barriers to sediment transport because of systemic linkages that form new dynamic interaction and interconnection between social characteristics and ecological system (Biggs et al. 2015).
Coastal slope and seabed sediment distribution illustrate shoreline stability. Time scale and land area, the amount of external energy, and beach material resistance determine coastal stability against shoreline change (Diposaptono 2004). Coastal slopes are linked to sediment type and distribution that cause abrasion and accretion on the beach. The coastal slope (Table 2) in sediment cells 1, 2, and 4 were strongly sloping (14-19%) with a coarse grain texture, and in sample 3, it was moderately sloping (8-12%) with a coarse grain texture. Sediment with coarse fractions is typical of a strongly sloping terrain because as the coastal slope increases, more of this sediment will be transported. However, slopes dominated by grains with moderate fractions, such as sediment cell 5, indicate coastal abrasion (sediment transport) that causes sand particles to accumulate near high tide lines and anthropogenic factors that result in the loss of sediment particles with coarse fractions.
Tukad Gerokgak (sediment cell 1), Tukad Banyuraras (sediment cell 4), and Tukad Saba (sediment cell 5) flow throughout the year with varying discharge. Small discharge in the dry season diminishes river flow's ability to equal the sedimentation rate generated by tides, sea waves, and currents, thus creating a sand-based dam in the river mouth or sand barriers jutting into the sea. The curved coastline in sediment cell 1 has an 11.97 m-wide beach, undulating morphology, and black sand and bomb/lapilli on the surface. Furthermore, flash floods frequently hit Tukad Gerokgak; Fig. 6a, b show mounts of wastes left on the riverbanks-evidence of flash floods that often accumulate litter on their path. In general, the northern coast of Bali has sloping terrain. However, the estuary area tends to be level, which is thought to be the result of sedimentation of mud materials carried by several rivers emptying into the western part of Buleleng, e.g., Gerokgak, Banyuraras, and Saba in the east. Because of intensive sedimentation, the original river's cross-section is not visible. Figure 6c, d shows pictures of the Tukad Gerokgak estuary.
The Tukad Gerokgak estuary has coarse grains and a narrowing river mouth (see Fig. 6c, d). Because of its nearly level coastal slope, many people utilize it for residential purposes and as a location to harvest shellfish during low tides and moor their fishing boats. The same condition is also found in the southeastern part of sediment cell 3 (Nusantara Beach). Barriers like sand deposits play an essential role in forming shoreline geometry and are used extensively for human activities (Aagaard et al. 2004). Barriers can protect physical features of regional development, such as settlements, ports, and other various land use activities. However, they are among the most dynamic coastal elements and the most vulnerable to sea-level rise in addition to having regulatory ecosystem services. Here, the development is oriented towards settlements and shrimp ponds instead of the tourism industry. Based on oral sources, history saw that development began with shipping activities in North Bali dating back to the seventeenth century, which is attributed to the Buginese people (Makassar) who migrated to and settled in the Bugis Buleleng Village, Penyabangan, especially in Celukan Bawang Village, and Sumberkima Village (Astiti 2018).
Sedimentation analysis using the statistical approaches on GRADISTAT provided information on the depositional environment and an overview of sediment transport and depositional processes based on grain-size distribution. The frequency distribution of grain size is closely related to the processes, seafloor sediment type, dynamics, and energy in a depositional environment (Carranza-Edwards et al. 2005). The lower the sorting value, the better the process for sorting sediment samples (Warrier et al. 2016). Sediment sorting in a coastal environment is generally good (Folk 1980), but the sediment analysis results showed that the particles making up the sediment population from Tukad Gerokgak to Tukad Saba had poorly sorted, moderate grain size: mud, fine sand, gravel (rock), and coarse sand. Rocky surfaces were found at almost all observation points, meaning that the sediment population is composed of coarse to moderate fractions, with varying deviations between the mid and mean values of this population-which indicates poorly sorted sediment particles. Sediment distribution also varies: highly concentrated at certain points but evenly distributed as a whole. Sediment is deposited toward coarse, even very coarse, grain.
The statistical calculation results of sample 1 are shown in Table 3. The sample location had a gritty soil texture, i.e., 88% sand was categorized as slightly gravelly sand. This dominant medium-size sand was poorly sorted (328.4 m, with an average diameter of 1.607). The sorting classification showed non-uniformity because of the currents' energy that moved and deposited material on the shoreline. The flow distribution can separate between fine and coarse particles with skewness of −0.185 m (fine skewed), while kurtosis showed very leptokurtic sediment.
In general, the sediment in cell 2 was very coarse sand with poor sorting, possibly because currents prevent water flows from properly separating particles. Besides, the sample location was far from sources of alluvial sediment. However, the sediment had symmetrical skewness, just like sample 3. Sample 3 is part of the bay closest to a steam-electric power station in Celukan Bawang. It is called Nusantara Beach, which is currently not managed. However, at present, there are concerns about the condition of the coast being threatened by its sustainability and vulnerable to the physical aquatic environment in its surroundings. Threats to the coast include shoreline changes caused by high abrasion and damage to coral reefs due to intensive industrial practices and coal shipping.
Sediment characteristics that reflect current geological complexity are beginning to change. Coastal problems and management should factor in social aspects. Humans are an essential key in land use development and, thus, a driving factor in socio-ecological   (Crossland 2005;Mutaqin 2020). Sediment in sample 3 was extremely poorly sorted because it was mixed with coral reefs on the seabed sediments. The bay's deeper areas had finer sediment, whereas the closer areas to the sea (mouth of the bay) had coarser grains than other areas. This finding suggests that the sediment comes from the sea and is then transported and finally deposited at the observation points. A high presence of marine biota shells and dead marine organisms categorizes the sediment as biogenic. On the other hand, some sediments nearshore had terrigenous deposits in the form of fine-sized rocks, clay minerals, and plant remains, indicating influence from the land. During low tides, the location of sediment cell 3 turns into a seagrass bed with sea cucumbers in it; its seaward end is up to 1 km from the low tide line in the morning (Fig. 7). Sediment cell 4 is located in the Tukad Banyuraras estuary. As observed from the remote sensing images, it is highly dynamic and separated into two smaller estuaries by a sand bar (Fig. 8). Land utilization, including resort buildings, is believed to induce changes in future beach morphology. Sediment cell 5 is at the Tukad Saba estuary abutted by densely populated slum settlements. Its location had an irregular shoreline and dominant black sand material, with a beach extending up to 19.1 m in width. Waste accumulating in the estuary narrows its size and is thereby considered the factor of the estuary's dynamics; this is in contrast to sample 4, where natural processes are responsible. This narrowing began with piles of garbage washed up by strong winds and flash floods in 2010 and then in February 2019. The kurtosis patterns of samples 3 and 4 were classified as very platykurtic, meaning that deeper bathymetry has a more platykurtic grain-size distribution curve and poorer sediment sorting (Folk and Ward 1957;Cadigan 1961). Sample 5 had mesokurtic kurtosis, which shows that coarser sediment (coarse silt) is in line with its texture, i.e., muddy sandy gravel. In sample 5 (very coarse silt), the skewness value approaches the very fine skewed region, while in sample 4 (sandy gravel), it forms a very coarse skewed curve.
Geological conditions have an effect in shaping coastal morphology, whether headlands or bays, and the influence on the thickness of the material where the material is deep/thick allows the tendency for least interaction with bedrock (Gallop et al. 2020). The slope of the slope affects the size of the sediment grains, where the larger the grain size, the steeper it is (Gallop et al. 2020). In line with coarse materials such as gravel can be deposited in slope gradient > 5°, it is also suggested that only gravel and other coarse materials are coarse sand on slopes between 2° and 5° (Trenhaile 2004). The dominant material tends to come from volcanoes and Alluvium (Qa) deposits in the form of scale, gravel, sand, silt. The Fig. 7 Tide variation on Nusantara Beach at 09:00 in the morning (a) and 04:00 in the afternoon (b) beach material is covered by coarse material such as gravel to rocks and silt. The megascopic material in sediment cell 1 has black physical properties, is very fine-coarse in size in the sand, mostly gravelly, rounded-angled grain shape, and poor sorting by the main composition of gravel. Sediment cell 2 consists of a material with grayish physical properties, coarse size, some coral reefs, and poor sorting sediment cell 3 has physical material, very fine light white sand-very poorly sorted material, with a biogenic mixture that is limestone, consisting of foraminifera, mollusk shells, and coral reefs. Sediment cell 4 shows physically black material with coarse size with very poor sorting, while sediment cell 5 material is physically black and coarse in size.

Shorelines
Shoreline change is a natural dynamic phenomenon that is controlled by beach shape, sediment characteristics, climate change, and anthropogenic effects (Duru 2017). Sediment transport, sea-level changes, and geomorphological characteristics work together to erode the shoreline and shift its position (Dewi 2019). Based on satellite image analysis, the shoreline observed is 21.02 km in length, stretching from the coordinates 256,825.77 to 273,052.485 mE and from 9,095,248.503 to 9,095,855.771 mN. Each of its segments was affected by abrasion at varying degrees, and only a few show signs of accretion. These indicate intensive abrasion due to cross-shore transport, where most of the lost sand is deposited into the deeper parts of the sea. Therefore, it is implausible that sand returns to its original location due to deep bathymetry at a fairly close horizontal distance. Longshore transport was also found due to severe abrasion in several places after coastal protection structures were added to adjacent regions. Landsat imagery shows shoreline variation in 20 years (2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014)(2015)(2016)(2017)(2018)(2019). Within this period, the shoreline in parts of the Buleleng Regency extended as a result of accretion. After comparing the length of shorelines on multitemporal images, it is apparent that the shoreline grew longer in 20 years, with an increase of 0.875 km. Morphodynamics is not the sole cause of this increase because of the role of anthropodynamic factors, e.g., constructions of ports, docks, and other coastal defense structures. Because the shoreline changes position from time to time, its monitoring needs to factor in spatial and temporal aspects (Dewi 2019). Moreover, it is a morphological feature that is often used to understand how coastal systems work and how mid-to-long-term processes can affect their development (Pardo-Pascual et al. 2018). Figure 9a shows a significant shift in some parts of the shoreline, and Fig. 9b presents shoreline changes in the entire research area from 2000 until 2019. Figure 9b shows the most significant distance between shorelines in 2000, 2008, and 2019. Most of the 119 transects were accreted from 0 to 170.6 m, and a significant change (> 8 m) was found almost on the entire shoreline, including bays and headlands. Meanwhile, from 2018 until 2019, the shoreline changed by 0-35.1 m, mostly in the range of 0-4.3 m. Significant changes (> 6.4 m) were seen on headlands and estuaries and only in the bay near the steam-electric power station Celukan Bawang.
One transect was set at a length of 350 m and a distance of 200 m to neighboring transects; for the entire length of the shoreline observed, this setting creates 119 transects. On the graph (Fig. 10), positive values indicate accretion, while negative ones show abrasion. Abrasion susceptibility is presented in an EPR graph, with different colors marking different degrees of resultant damages: dark green for mild damage, light green for moderate damage, green-yellow for heavy damage, and orange for very heavy damage, and red for extremely heavy damage.
The largest shoreline change envelopes (SCE) of up to 160 m are on transects 8, 30, 31, 55, 103, and 104 (see Fig. 10) located in bays and river estuaries. Significant shoreline shifts occurred in these locations. As depicted in the NSM graph (shoreline changes in 2000-2019), more shoreline transects extended seaward, with a maximum rate of change (EPR) reaching 9 m/year. As seen on Landsat images, the abrasion susceptibility in 20 years of observation was mainly linked to mild coastal damages due to accretion.

Bathymetry
Waters around Buleleng had varying depths, from 0 to 236.36 m. The eastern coast was more profound and had a substantially wider depth gradient than the western counterpart (Fig. 11). The latter had a more gradual and wider change of depth, while the former had a narrower interval of change, except in the southeast, where changes in water depth were regular and gradual and had a wide interval of distance from one water depth to the next.
Morphologically, sample 3 was a very sloping beach with the smallest area. This variation indicates weaker pressure compared to other locations. The results also showed that part of the beach close to the high tide line (0-20 m) had a large slope gradient, while the rest had smaller ones. These slopes are different from the topography of the island as a whole. Shoreline dynamics are mainly influenced by tidal currents, although waves create 1 3 heavy pressure in certain seasons. Level beach slopes mean that the pressures acting upon them come from currents influenced by tidal waves. Sediment distribution correlates with depth: the more profound the bathymetry, the finer the sediment (Putra and Nugroho 2017).
Correlation of beach slope with sediment type and distribution (see Tables 2, 3) revealed that higher gradient slopes allowed sediment transport with coarse fractions. In general, tide-dominated coasts have a tidal range of above 2 m. The west season, peaking in February, triggers high waves that cause erosion and damage public facilities and residential buildings. Coastal defense structures built to reduce wave energy are engineering conservation made of stones and concrete. This structural mitigation measure is presented in Fig. 12.
The above description indicates that the coastal landscape pattern in the Buleleng Regency is mainly covered by settlements and industrial buildings growing along the shoreline. However, these components cannot develop optimally, unlike in urban areas where the landscape pattern grows following elevations, slopes, and the developments of roads, city centers, airports, and ports (Han et al. 2009). Here, the rapid development of resorts and hotels (see Fig. 8c) leads to extensive land-use conversion, bringing about environmental degradation due to water pollution and seawater intrusion (Gössling 2001;SOPAC-UNEP 2005;Marfai et al. 2020), exacerbated by the impact of groundwater exploitation, landuse change, trade sector, and shipping activities (Marfai 2014). Cultural and natural tourism sectors are the regency's development priority (Astiti 2018). A collaborative approach in enforcing policies that combine institutions and natural sciences can start with studies of changes in coastal areas to realize systematic environmental management (Mazé et al. 2017). Along the shoreline observed, different parties are responsible for shaping the coastal management in their localities: fisher communities in sediment cells 1, 2, and 5, port corporations, and tourism awareness communities (Pokdarwis) in sediment cell 3, and local people in sediment cell 4. In the context of coastal management that integrates  physical with social processes, institutions play a necessary part in empowering communities, conserving resources, and managing land utilization, including structural mitigation measures constructed according to coastal landscape characteristics and morphodynamics.
Coastal management requires a detailed understanding of the area's anthropogenic characteristics to have a complete picture of the coastal system and mechanisms of change to predict future coastal trends and plan appropriate management strategies. Coastal management requires analysis of sediment cells (Cooper and Pontee 2006;Montreuil and Bullard 2012;Gallop et al. 2020), analysis of sediment characteristics through grain size (Trenhaile 2016), and understanding sediment transport (Hooke et al. 1996), coastal morphodynamic (Dias et al. 2015), as well as the analysis of shoreline changes (Montreuil and Bullard 2012). Therefore, a management planning process that considers shoreline management considers the natural processes of the coast and the interactions that occur in the coastal system at various temporal and spatial scales.

Conclusion
The coastal landscape, particularly the morphodynamic aspects, of the Buleleng Regency shows that the sediment population from Tukad Gerokgak to Tukad Saba is composed of poorly sorted, medium-size grains and that, overall, sediment is deposited toward coarse to very coarse grain. Its strongly sloping morphology allows accretion to dominate along the entire shoreline. As identified from Sentinel-2A imagery, for 20 years (2000-2019), shoreline change is mainly caused by accretion (an increase of up to 0.875 km) attributed to morphodynamic and anthropodynamic factors. Landsat image analysis indicates that abrasion is categorically light because accretion dominates the processes acting upon the shoreline.
Furthermore, as identified from Sentinel 2A images, the abrasion susceptibility is associated with moderate damage level because massive abrasion occurs in a few locations. The results of the morphological analysis in this study are useful as input for policymakers in Buleleng Regency to determine the locations that need to be strengthened in order to mitigate strong erosion effects, either using structural or non-structural mitigation. Based on our research results, the beach material comprises sand and gravel; therefore, large waves can quickly erode the shoreline, warranting the need for structural mitigation than nonstructural. However, it is also possible to carry out the two mitigation methods together.
This research shows the advantage of the combination of cell sediment and morphological analysis to provide the dynamic environment of the coastal areas to support coastal management projects by coastal managers. This research used multi-years scales data to solve the problem of data availability (especially in a developing country, where data availability on a large scale not available in many cases), however still considering the quality of the data and reliability of the data source. For further study, it is recommended to include an accuracy assessment so that the resulting model is more reliable.