Multi-Decadal Shoreline Dynamics and Pathways for Sustainable Coastal Management in Ujung Pangkah, Indonesia

Coastlines are dynamic land–sea interfaces shaped by waves, tides, currents, sediment supply, and human activities. These processes generate spatially variable patterns of erosion and accretion. In deltaic settings, shoreline dynamics are often more complex because fluvial sediment inputs, monsoon-driven circulation, and low coastal gradients interact over time [1], [2]. Ujung Pangkah, located at the mouth of the Bengawan Solo River in Indonesia, represents this type of dynamic deltaic coast. Previous studies have reported high variability in currents, suspended sediment transport, and shoreline responses along the Gresik–Madura Strait corridor [3], [4]. These changes have direct implications for mangrove condition, fisheries production, aquaculture viability, and coastal community resilience. Therefore, reliable long-term shoreline monitoring is needed to support sustainable coastal planning, conservation, and livelihood protection [5], [6].

Satellite-based Earth observation provides a practical basis for detecting shoreline change over long time periods. Multi-epoch satellite imagery can be combined with water–land separation indices, such as the Normalized Difference Water Index (NDWI) and Modified Normalized Difference Water Index (MNDWI), to extract shoreline positions consistently. When these shoreline datasets are analyzed using the Digital Shoreline Analysis System (DSAS), they can provide standardized measurements of shoreline displacement and change rates across different coastal segments [7]. However, shoreline metrics alone are insufficient for management interpretation. They need to be evaluated together with hydro-morphological conditions, including tides, waves, currents, and coastal slope, to identify possible drivers and intervention priorities [8].

Despite the environmental and socio-economic importance of Ujung Pangkah, a unified multi-decadal baseline of shoreline change remains limited. Existing information has not fully integrated long-term satellite-derived shoreline positions, DSAS-based change metrics, hydro-oceanographic interpretation, and land-use context into a single management-oriented assessment. This gap limits the ability to distinguish persistent erosion hotspots from areas where accretion has been sustained. It also constrains the translation of shoreline-change evidence into segment-specific coastal management strategies.

This study addresses this gap by quantifying multi-decadal shoreline dynamics in Ujung Pangkah from 1973 to 2021. The analysis aims to: (i) establish a long-term shoreline baseline across the satellite era; (ii) identify spatial patterns of accretion and erosion using DSAS metrics; (iii) interpret these patterns in relation to hydro-oceanographic conditions and land-use change; and (iv) translate the results into management pathways for coastal protection, mangrove rehabilitation, and sustainable coastal planning. Two hypotheses are tested. First, shoreline change in Ujung Pangkah is organized into accretionary and erosional cells rather than forming a uniform coast-wide trend. Second, accreting sectors are associated with conditions that favor sediment retention and mangrove stabilization, whereas persistent erosion is linked to stronger hydrodynamic exposure and reduced sediment trapping [5], [9].

This study argues that integrating multi-decadal satellite-derived shorelines, DSAS metrics, hydro-oceanographic context, and land-use information can reveal whether shoreline change in Ujung Pangkah forms a uniform trend or a segment-structured erosion–accretion mosaic. By applying this workflow to the 1973−2021 shoreline record, the study provides a long-term spatial baseline, identifies priority erosion and accretion sectors, and translates the results into management pathways for mangrove rehabilitation, coastal protection, and spatial-planning integration. This contribution supports evidence-based coastal management in Ujung Pangkah and provides a reproducible approach for other deltaic environments facing similar shoreline-change dynamics.

Ujung Pangkah District is located in Gresik Regency, East Java, Indonesia, and covers approximately 9,483.23 ha across 13 villages. The area is predominantly low-lying and forms part of a dynamic deltaic coast influenced by the Bengawan Solo River. This study analyzed approximately 30 km of coastline. The coastline was divided into six 5-km analysis areas based on coastal profile characteristics and extreme net shoreline movement (NSM) values. The western and eastern coastal sectors, represented by Pangkah Kulon and Pangkah Wetan, are shown in Figure 1. The station names, representative coordinates, DSAS areas, and brief descriptions are listed in Table 1.

The stations served as reference points for the six DSAS analysis areas along the Ujung Pangkah coastline. Coordinates are reported in decimal degrees. The area codes A−F correspond to the analysis domains shown in Figure 1 and are used consistently throughout the shoreline-change analysis.

This study used secondary datasets to quantify shoreline change and interpret the physical and land-use context of coastal dynamics. The datasets, providers, access points, and analytical roles are summarized in Table 2. The analytical workflow consisted of shoreline extraction from Landsat imagery, DSAS-based calculation of shoreline-change metrics, and contextual interpretation using tidal, wave, current, beach-slope, and land-use datasets.

Multi-epoch Landsat imagery was used as the primary dataset for shoreline extraction and DSAS analysis. The shoreline epochs included 1973, 1982, 1994, 1996, 2001, 2006, 2011, 2016, and 2021. Tidal data were used to support tidal normalization and shoreline interpretation. Wave, ocean-current, beach-slope, and land-use datasets were not used to compute DSAS rates directly. Instead, these datasets provided interpretative context for explaining spatial differences in erosion and accretion patterns across the study area.

Satellite imagery was pre-processed in ENVI 5.1 before shoreline extraction. The pre-processing steps included geometric, radiometric, and atmospheric corrections. Atmospheric correction was performed using the Fast Line-of-sight Atmospheric Analysis of Spectral Hypercubes (FLAASH) module.

Because the Landsat sensor configurations differed across the observation period, sensor-consistent water indices were applied. The Normalized Difference Water Index (NDWI) was used for Landsat MSS imagery from 1973 and 1982. The Modified Normalized Difference Water Index (MNDWI) was used for Landsat TM, ETM+, and OLI imagery from 1994 to 2021 [10], [11]. MNDWI uses the shortwave infrared band instead of the near-infrared band and generally improves land–water separation. Scene-specific thresholds were then applied using Otsu classification and visual refinement. The resulting water masks were vectorized to generate shoreline positions. Quality control included visual inspection against true-color composites and morphological cleaning of the extracted shoreline features.

Shoreline-change statistics were computed using the DSAS v4.3 with a baseline–transect approach [12]. All shoreline epochs listed in Table 2 were merged into a single shoreline feature class. A shore-parallel baseline was digitized using the study-area frame shown in Figure 1. From this baseline, 451 shore-normal transects were generated to sample shoreline displacement across the six analysis areas. The primary end-point metrics are:

(1) NSM (end-point displacement), defined as the end-point displacement between the youngest and oldest shoreline for each transect.

(2) End point rate (EPR), the annualised rate computed as

where, T represents 48 years for 1973−2021.

For additional context, the Linear Regression Rate (LRR) was considered as an indicator of longer-term shoreline tendency. Shoreline-change magnitudes were classified following [12], [13] into high accretion ($>$4 m$\cdot$yr$^{-1}$), low accretion (1−4 m$\cdot$yr$^{-1}$), stable shoreline position (-1 to 1 m$\cdot$yr$^{-1}$), low erosion (-4 to -1 m$\cdot$yr$^{-1}$), and high erosion ($<$-4 m$\cdot$yr$^{-1}$). Quality control followed standard DSAS procedures. Shoreline positions were co-registered and interpreted using consistent tidal-stage information. Therefore, NSM and EPR values were interpreted with caution in relation to the 30-m spatial resolution of Landsat imagery and tidal-normalization uncertainty.

Tidal predictions from the Indonesian Geospatial Information Agency were used to guide image selection and shoreline interpretation. Where possible, satellite scenes were selected under similar tidal-stage conditions. When tidal stages differed, shoreline positions were interpreted relative to a common reference level using predicted tidal levels at the satellite overpass time. The tidal regime was described using the Formzahl number, and mean sea level was used as the reference datum [14]. These tidal data supported shoreline interpretation and uncertainty control but were not used as direct inputs for DSAS rate computation.

Wave and wind data from Copernicus were cropped to the study area and used to describe seasonal wave-forcing conditions. The data were transformed into effective fetch, and significant wave height (Hs) and peak wave period (Tp) were estimated using the Simplified Manning–Boussinesq formulation [15]. Seasonal and representative monthly wave conditions were summarized to support interpretation of potential longshore transport and erosion–accretion patterns. These wave data provided physical-forcing context and were not used to compute NSM or EPR values.

Surface-current data from the NASA Physical Oceanography Distributed Active Archive Center (PO.DAAC) were processed in Surfer 10.1 [16]. The processing generated annual means, seasonal composites, and a November snapshot to support interpretation of sediment-transport pathways. The resulting vector and contour maps were used to identify potential zones of sediment convergence and divergence. These current data were used as interpretative evidence for explaining shoreline-change patterns and were not used directly in DSAS calculations.

Digital Elevation Model Nasional (DEMNAS) data were clipped to the coastal corridor and buffered to 1 km inland [17]. Beach slope was calculated in degrees and averaged for each analysis area. The slope values were used to characterize coastal profile conditions and to support interpretation of shoreline response. They were not used to compute shoreline-change rates.

Landsat-derived land-use data were clipped to the study boundary, radiometrically corrected, and sharpened where applicable. Unsupervised classification was used to classify land use into settlement, rice fields, ponds, vegetation, and rivers [18]. The classified maps were used to summarize land-use dynamics across the study period. In the main analysis, percentage values were emphasized because they are more robust to differences in mapping extent. Absolute area values in hectares were reported as supporting information after checking the consistency of study-area clipping and unit conversion. Land-use information was used to interpret possible anthropogenic drivers of shoreline change, not to calculate DSAS shoreline-change metrics.

Seasonal surface currents in Ujung Pangkah ranged from approximately 0.015 to 0.37 m$\cdot$s$^{-1}$ and generally moved from west to east (Figure 2). During the East Monsoon, winds from Australia toward Asia reinforced the eastward flow. During the West Monsoon, the circulation became more variable and could redirect local current patterns. These seasonal current patterns indicate that monsoon-driven circulation may influence alongshore sediment redistribution and shoreline response in northern Java [19].

Wave conditions also varied seasonally (Figure 3). Significant wave height showed relatively small seasonal variation, whereas peak wave period varied more strongly among the West Season, East Season, Transition Season 1, and Transition Season 2. Periods of stronger seasonal forcing and large-scale climate anomalies may increase wave energy and raise the potential for shoreline retreat in exposed coastal reaches [19].

The tidal regime was diurnal, with a Formzahl value of approximately 7.64 (Figure 4). Sea-level interpretation in this study was based on the analysis datum, with mean sea level used as the reference level. Extreme high tides may temporarily shift the observed waterline and expand inland inundation, thereby influencing short-term shoreline position [20].

Beach slopes were very gentle across all analysis areas, ranging from 0.02° to 0.05° over a 1,000 m inland distance, with an overall mean of approximately 0.04° (Table 3). These low-gradient profiles favor energy dissipation and provide accommodation space for sediment deposition where sediment supply is sufficient.

Land use was classified into settlement, rice fields, ponds, vegetation, and rivers (Table 4 and Table 5). Over the 1973−2021 period, settlement and rice-field classes generally increased, indicating expanding urban and agricultural land use. Pond areas fluctuated over time, while vegetation, including mangrove-associated cover, declined during several intervals. The percentage-based land-use results are emphasized because they are more robust to differences in mapping extent than absolute area values. The observed decline in vegetation during some periods is relevant to shoreline interpretation because reduced vegetated cover can lower sediment-trapping capacity and increase exposure to erosion. Therefore, the land-use results provide contextual evidence for interpreting spatial differences in shoreline behavior rather than direct inputs for DSAS rate computation.

Shoreline positions from nine epochs, namely 1973, 1982, 1994, 1996, 2001, 2006, 2011, 2016, and 2021, were analyzed using DSAS across 451 transects. Area-level summaries are presented in Table 6, while the temporal and spatial patterns are shown in Figure 5 and Figure 6. Positive values indicate accretion or shoreline advance, whereas negative values indicate erosion or shoreline retreat. To ensure consistency across the 48-year observation period, EPR values were calculated as NSM divided by 48 years.

The DSAS results show a clear spatial contrast among Areas A−F. Five areas, namely Areas A−C and E−F, were dominated by net accretion, while Area D was the only sector with net erosion (Table 6; Figure 5 and Figure 6). This pattern indicates that shoreline change in Ujung Pangkah did not occur as a uniform coast-wide trend. Instead, it formed a segment-structured mosaic of accretion and erosion.

Area-level results further clarify this spatial contrast. Area A was accretion-dominated, with average accretion of 592.82 m in NSM and 12.35 m$\cdot$yr$^{-1}$ in EPR, compared with erosion of -263.85 m in NSM and -5.50 m$\cdot$yr$^{-1}$ in EPR. Area B showed a similar accretion-dominated pattern, with accretion of 553.90 m in NSM and 11.54 m$\cdot$yr$^{-1}$ in EPR, compared with erosion of -277.05 m in NSM and -5.77 m$\cdot$yr$^{-1}$ in EPR.

Area C showed the strongest accretion signal among all sectors. This area recorded accretion of 736.29 m in NSM and 15.34 m$\cdot$yr$^{-1}$ in EPR, although erosion also occurred locally, reaching -532.80 m in NSM and -11.10 m$\cdot$yr$^{-1}$ in EPR. Area D showed the opposite pattern and formed the main erosional hotspot. It recorded erosion of -416.53 m in NSM and -8.68 m$\cdot$yr$^{-1}$ in EPR, compared with localized accretion of 384.76 m in NSM and 8.02 m$\cdot$yr$^{-1}$ in EPR. Thus, despite the presence of local accretion, the area-level signal in Area D remained erosion-dominated.

Overall, the long-term DSAS record identifies five accretion-dominated sectors and one persistent erosion sector. This spatial pattern highlights the importance of segment-scale interpretation because coast-wide averages would obscure localized erosion risk in Area D and sustained accretion gains in Areas A−C and E−F.

The shoreline-change mosaic has important implications for mangrove habitat configuration, estuarine connectivity, and coastal livelihoods. In accretion-dominated sectors, shoreline advance can create additional accommodation space for mangrove establishment and foreshore expansion. These changes may increase habitat continuity and nearshore structural complexity, which are important for fish reproduction and juvenile nursery functions [21]. Areas A−C and E−F therefore represent sectors where accretionary gains could support mangrove maintenance or expansion if land conversion is controlled. These habitat gains are also consistent with the broader role of mangroves in coastal protection, carbon storage, and biodiversity support in deltaic environments [22], [23].

In contrast, the persistent retreat in Area D indicates a zone of shoreline instability. Erosion in this sector can reduce mangrove-edge stability, fragment habitat connectivity, and increase exposure to wave energy and sediment resuspension. These processes may weaken nursery functions for estuarine fishes and invertebrates and increase the vulnerability of pond-based livelihoods [22], [23], [24].

The contrast between accreting and eroding sectors is directly relevant to local fisheries and aquaculture systems. Stable mangrove–estuary linkages can support ecosystem services and livelihood opportunities, whereas shoreline retreat can increase maintenance costs, habitat loss, and exposure risk for coastal communities [25], [26]. Previous studies in Ujung Pangkah and other Indonesian coastal systems have further emphasized that mangrove ecosystems support multiple management-relevant functions, including ecotourism potential, habitat-condition assessment, marine-debris monitoring, mangrove-density evaluation, and economic sustainability [27], [28], [29].

The land-use results further support this interpretation. Periodic declines in the vegetation class, including mangrove-associated cover, coincided with intervals where erosion was more evident in the shoreline-change maps (Table 4 and Table 5; Figure 5 and Figure 6). This pattern suggests a potential feedback between shoreline position, mangrove cover, and coastal exposure. It is also consistent with broader evidence that mangrove degradation can reduce natural buffering capacity and increase shoreline instability [30]. In livelihood terms, erosion can increase the vulnerability of fishers and aquaculture farmers by exposing ponds, infrastructure, and productive coastal land to recurrent losses [31].

Taken together, the results indicate two management directions. First, accretion-dominated sectors, especially Areas A−C and E−F, should be protected to maintain shoreline gains and mangrove-related ecosystem services. Second, Area D should be prioritized for targeted erosion control, mangrove rehabilitation, and livelihood-risk reduction.

The shoreline-change pattern in Ujung Pangkah from 1973 to 2021 shows a stable, segment-structured contrast rather than a single coast-wide trajectory. DSAS metrics indicate that most sectors were accretion-dominated, particularly Areas A−C and E−F, where the accretion component reached approximately +12 to +15 m$\cdot$yr$^{-1}$. In contrast, Area D formed the main erosional hotspot, with an erosion component of -8.68 m$\cdot$yr$^{-1}$ and an NSM erosion value of -416.53 m. At the area level, Area D remained erosion-dominated, with a net EPR of -0.66 m$\cdot$yr$^{-1}$ and a net NSM of -31.77 m (Table 6; Figure 5 and Figure 6). This “accretion belt + erosion pocket” pattern suggests that shoreline change in Ujung Pangkah was controlled by segment-scale differences in sediment availability, hydrodynamic exposure, and coastal land-use context. Therefore, interpreting shoreline change only from an overall mean value would obscure localized erosion risk and intervention priorities.

Comparable segment-specific shoreline behavior has been reported along the northern coast of Java using remote sensing and DSAS-based approaches. DSAS-based analysis along the Sayung–Demak coast of Central Java documented strong spatial heterogeneity in shoreline trajectories and showed that severe erosion can be concentrated in specific coastal segments, while accretion is commonly associated with local variations in waves, storm surges, tidal currents, bathymetry, and mangrove cover [32]. Similar shoreline-change studies along the Java coast also showed that coastal change can occur at rates of meters to tens of meters per year, reinforcing that the magnitudes observed in Ujung Pangkah are plausible for actively modified northern Java coastal systems [33], [34].

The Ujung Pangkah results are also consistent with local evidence from the Bengawan Solo deltaic system. Previous satellite-based assessments reported measurable progradation and land-cover transformation in the Bengawan Solo–Ujung Pangkah coastal plain, indicating a strong relationship between sediment delivery, shoreline advance, pond expansion, and mangrove or vegetation dynamics [35]. In the present study, the dominance of accretion in Areas A–C and E–F is consistent with sediment retention and delta-front development, while the persistence of erosion in Area D indicates that sediment supply and shoreline stabilization were spatially uneven.

Beyond Java, DSAS-based assessments in other deltaic environments also show that accretion and erosion frequently coexist within the same coastal system. In the Indus Delta, long-term shoreline records showed dominant retreat at the delta scale, but localized accretion pockets were still detected in specific sectors [36], [37]. A similar pattern was reported in the Meriç Delta, where erosion and accretion alternated across sub-periods and coastal segments [38]. These comparisons support the interpretation that deltaic shorelines rarely respond uniformly to regional forcing. Instead, shoreline trajectories are shaped by the local balance among sediment supply, wave–tide processes, nearshore morphology, vegetation cover, and human modification.

The DSAS results indicate that shoreline change in Ujung Pangkah is organized as a segment-structured accretion–erosion mosaic. Accretion dominated most sectors, particularly Areas A−C and E−F, whereas Area D remained the main erosional hotspot. Rather than indicating a single coast-wide trend, this pattern shows that shoreline response varied systematically alongshore. This finding is consistent with deltaic coasts where sediment retention and effective wave–current exposure are not evenly distributed among segments. Similar DSAS-based delta assessments have also reported coexisting accretion and erosion that cluster into localized hotspots over multi-decadal periods [32].

The physical setting of Ujung Pangkah helps explain this spatial sensitivity. The coast has very gentle beach slopes, ranging from approximately 0.02° to 0.05° over 1 km, with an overall mean of about 0.04° (Table 3). It is also influenced by a diurnal tidal regime, with a Formzahl value of approximately 7.64 (Figure 4). In such a low-gradient and tide-influenced setting, even small gains or losses of sediment can produce visible shoreline displacement. The position of the waterline can also vary with tidal stage, which reinforces the need to interpret DSAS-derived shoreline changes at the segment scale.

The differences among shoreline segments can be explained by the combined influence of seasonal coastal forcing and land-use change. Monsoon-driven currents and changing wave conditions may alter the direction and intensity of alongshore sediment transport. As a result, some sectors can receive and retain more sediment, while others remain more exposed to erosion. At the same time, land-use change on the coastal plain, especially conversion between mangrove or vegetated areas and aquaculture ponds, can reduce sediment trapping and wave-buffering capacity. Similar DSAS-based studies in other deltaic systems have shown that variations in sediment supply and human modification can reinforce localized erosion and accretion, producing a persistent alongshore mosaic rather than a uniform shoreline trend [34].

These findings support the interpretation that shoreline-change metrics should not be used in isolation. DSAS provides a quantitative baseline of shoreline displacement and change rates, but hydro-oceanographic conditions, coastal slope, and land-use dynamics are needed to interpret why erosion and accretion differ among sectors. This integrated interpretation is important for deltaic environments such as Ujung Pangkah, where shoreline change reflects the combined effects of sediment supply, hydrodynamic exposure, vegetation condition, and human land use.

The DSAS-based spatial mosaic provides a basis for segment-specific coastal management. The results indicate strong accretionary tendencies in Areas A−C and E−F, whereas Area D showed persistent net erosion, with a net EPR of -0.66 m$\cdot$yr$^{-1}$ and a net NSM of -31.77 m. The erosion component in Area D reached -8.68 m$\cdot$yr$^{-1}$ and -416.53 m in NSM, confirming this sector as the main erosional hotspot (Table 6; Figure 5; Figure 6). Therefore, management interventions should not be applied uniformly along the coast. Accretionary sectors should be protected to maintain shoreline gains, while Area D should be prioritized for erosion control, mangrove rehabilitation, and livelihood-risk reduction.

Mangrove restoration should be prioritized because mangroves can reduce wave energy, enhance sediment retention, and stabilize low-gradient shorelines [39]. In Ujung Pangkah, this function is particularly important because local aquaculture expansion and vegetation changes have altered the mangrove–pond landscape [40]. Restoration should therefore be hydrology-aware, site-specific, and community-engaged to sustain both ecological and livelihood benefits [41], [42], [43].

In Area D, mangrove-based stabilization should be treated as a high-priority intervention. Restoration should focus on improving hydrologic exchange, reconnecting tidal creeks where feasible, and establishing a protective mangrove belt in areas with suitable sediment and inundation conditions. In contrast, Areas A−C and E−F should be managed to maintain accretionary gains. These areas require protection from further conversion of accreting fringes, supported by enrichment planting along prograding edges where mangrove establishment is ecologically suitable.

Community-based conservation can strengthen shoreline management because local fishers, pond farmers, and coastal residents are directly affected by mangrove condition and shoreline instability [44], [45]. In erosion-prone sectors, community involvement can support restoration maintenance, local monitoring, and early detection of shoreline retreat. In accreting sectors, community stewardship can help ensure that newly formed or expanding mangrove fringes are maintained as ecological assets rather than rapidly converted into ponds or other land uses.

For Area D, community-based monitoring groups could be linked with mangrove rehabilitation activities to support maintenance after planting and to document shoreline changes over time. In Areas A−C and E−F, community stewardship should focus on protecting accreting mangrove fringes so that shoreline gains translate into durable nursery habitat, coastal protection, and blue-carbon benefits [46], [47].

Hybrid coastal infrastructure may be considered where mangrove-only stabilization is insufficient. Such measures should complement nature-based rehabilitation rather than replace it. Hybrid structures can help reduce wave energy and support sediment deposition, but poorly planned hard structures can alter sediment transport and degrade coastal habitat quality [48], [49], [50].

In Ujung Pangkah, any hybrid intervention should be spatially targeted and aligned with the shoreline-change mosaic. Area D may require combined mangrove rehabilitation and low-impact supporting structures if erosion pressure remains high. However, infrastructure placement should consider seasonal current and wave patterns to avoid downdrift sediment starvation. Post-intervention monitoring should use the established remote-sensing and DSAS workflow to evaluate whether shoreline stabilization and mangrove recovery are occurring as intended.

The DSAS-based evidence can support spatial planning by identifying where protection, rehabilitation, and controlled-use zones should be prioritized. Area D can be designated as a rehabilitation and coastal-protection priority zone because it represents the main erosional hotspot. In contrast, Areas A−C and E−F can be managed as accretionary conservation or controlled-use zones to prevent uncontrolled conversion of newly formed coastal land.

Integrating this spatial evidence into regional coastal zoning can help reduce land-use conflicts and align shoreline-change information with mangrove management, aquaculture planning, and livelihood protection [51], [52]. Participatory implementation is also important because local compliance and long-term maintenance are more likely when community actors are involved in planning and monitoring [53]. Complementary habitat-risk screening can further support the prioritization of vulnerable areas and reinforce why Area D requires urgent intervention [54].

The shoreline-change pattern has direct relevance for community resilience and food security. Accreting sectors, especially Areas C and E, provide opportunities to maintain or expand mangrove fringes that support nursery habitat and stabilize shore-adjacent aquaculture systems. These conditions may help reduce exposure and maintenance costs for pond-based livelihoods if land conversion is controlled.

In contrast, persistent retreat in Area D increases exposure of ponds, fishing infrastructure, and productive coastal land to recurrent losses. This condition supports the need to link shoreline rehabilitation with livelihood-risk reduction, aquaculture adaptation, and community resilience planning [55], [56], [57]. Therefore, coastal management in Ujung Pangkah should not only focus on shoreline stabilization, but also on protecting the ecological functions that support fisheries, aquaculture, and household income.

Taken together, these management pathways form an evidence-to-action sequence. DSAS diagnostics identify priority segments, physical and land-use datasets explain the likely drivers, and targeted interventions translate the shoreline-change evidence into coastal rehabilitation, protection, and monitoring actions. This integrated approach is consistent with social-ecological sustainability strategies for mangrove conservation, where ecological rehabilitation, community participation, and adaptive management need to be linked within a single management framework [58]. Figure 7 summarizes this evidence-to-action framework by linking the mapped shoreline-change mosaic to segment-specific management options and enabling conditions.

This study demonstrates that shoreline evolution in Ujung Pangkah from 1973 to 2021 was not spatially uniform but was organized as a segment-structured accretion–erosion mosaic. Most coastal sectors showed persistent accretionary tendencies, whereas Area D remained the main erosional hotspot. This finding confirms that coast-wide average shoreline-change values may obscure localized erosion risk and sustained accretion gains in dynamic deltaic environments.

The main significance of this study lies in its integration of multi-decadal satellite-derived shorelines, DSAS metrics, hydro-oceanographic context, coastal slope, and land-use information into a management-oriented assessment. This integrated workflow provides an operational baseline for identifying priority intervention zones, protecting accretionary sectors, targeting erosion control, and supporting mangrove rehabilitation and spatial-planning decisions in Ujung Pangkah.

Several limitations should be acknowledged. Shoreline extraction relied on 30-m multisensor Landsat imagery, which may introduce positional uncertainty, particularly in low-gradient and tide-influenced coastal settings. Although tidal normalization and quality-control procedures were applied, residual uncertainty related to tidal-stage variation may remain. In addition, area-level DSAS summaries may mask within-area transect variability, and land-use classification was used as contextual evidence rather than direct causal attribution.

Future research should integrate higher-resolution satellite imagery, UAV-based shoreline and mangrove-edge mapping, field validation of sediment and hydrodynamic conditions, and more detailed transect-level shoreline statistics. Post-intervention monitoring is also needed to evaluate the effectiveness of mangrove rehabilitation, hybrid coastal protection, and livelihood-oriented adaptation strategies under future coastal-management scenarios.