IMPACT OF LAND USE CHANGES ON THE TSUNAMI HAZARDS IN PART OF COASTAL KEBUMEN

This research was conducted to analyze land use change from 2016 to 2022 and their impact on the tsunami hazard zone on the coast of Kebumen Regency. For land use change analysis, remote sensing, geographic information systems (GIS), and statistical tests were applied to Sentinel-2A satellite imagery. The tsunami hazard was simulated using tsunami inundation modeling based on land use spatial data and DEMNAS processed using GIS. Land use changes occurred significantly in the study area, especially in Mirit District. The Southern Cross Road (JJLS) and coastal morphological conditions influence land use change patterns. Land use changes impact changes in the tsunami hazard zone, especially in the use of fir forests and shrimp ponds. The research findings can be used as input for developing a tsunami disaster mitigation plan and detailed spatial planning on the coast of Kebumen Regency.


INTRODUCTION
The south coast of Java Island has a tsunami threat that needs to be watched out for.Tsunamis are a series of sea waves with a large-scale impact from earthquakes, landslides, and volcanic eruptions that occur spontaneously in the sea [1]- [3].Some of the southern coasts of Java Island have records of occurrence and potential for tsunamis with run-up heights of up to 15 m [4]- [9].On July 17, 2006, a tsunami-triggered earthquake (Mw 7.7) occurred at a depth of 34 km and, centered 225 km off the south coast of Pangandaran Regency (9,2220 o S, 107,3200 o E), had an impact extending to the south coast of Kebumen Regency [6], [10], [11].On the coast of Kebumen Regency, 12 people died, 46 people were missing, 27 people were injured, and 1388 people were evacuated to a safe place [6], [11].Tsunami inundation modeling for tsunami hazard assessment on the coast of Kebumen Regency conducted by several researchers resulted in a tsunami inundation of 600 m to 3,300 m from the coastline [12], [13].Several tsunami risk studies have shown that parts of the coast of Kebumen Regency are at high risk [14]- [16].
The coast is a strategic area for people's lives and livelihoods.The coast is a meeting zone of land, sea, and atmospheric systems with unique physical conditions and natural resources [1], [17].Utilization of natural resources results in the development of infrastructure to support economic and social activities [18]- [20].Road infrastructure is one of the factors triggering land use changes in regional development [21], [22].The Southern Cross Road (JJLS) was built in parts of the coast of Kebumen Regency, Central Java Province, from 2016 to 2019 [23].JJLS causes an increase in land prices and land use changes on the Kebumen coast [24].
JJLS is designated as a strategic area in the 2011-2031 Regional Spatial Plan (RTRW) of Kebumen Regency, which can be a trigger for regional development [25], [26].Land use change is an impact of regional development that can be identified with multitemporal spatial data.This study aims to analyze land use changes from 2016 to 2022 and their impact on the tsunami hazard zone on the coast of Kebumen Regency.A further aim is to identify the types and spatial patterns of land use that affect changes in the tsunami hazard zone.It is hoped that the findings of this research can become input for developing a tsunami disaster mitigation plan and detailed spatial planning for the coast of Kebumen Regency.

METHOD
The research location is part of the coast of Kebumen Regency, covering the coastal area to the JJLS area (figure 1).JJLS at the research location is approximately 1 km from the coastline.Coastal areas can be identified from the topography, lithology, and hydrology conditions resulting from fluvial, marine, and aeolian geomorphological processes [27].Specifically, the research location is located between the estuary area of the Lukulo River to the estuary area of the Wawar River with flat to gentle slope conditions [16], [28], [29].The coastal lowland lithology is a segregated loose sand beach sediment formation zone.Common landforms at the study site include beach ridges, swales, lagoons, coastal plains, and floodplains [30].The research locations are included in three districts, namely Buluspesantren, Ambal, and Mirit.The first spatial data used in this research is Sentinel-2A satellite imagery in 2016 (before JJLS was built) and 2022 (after JJLS was completed), which were downloaded via https://earthexplorer.usgs.gov/.Visual interpretation of Sentinel-2A satellite imagery with a spatial resolution of 10 m is used to identify land use.The 2014 national land cover classification standard [31] is used for land use classification.The confusion matrix method was used to test the accuracy of land use data, and the stratified proportionate random sampling method was used to determine field check samples [32], [33].The first analysis used an overlapping technique based on a geographic information system (GIS) to determine land use changes in area and distribution.The second analysis was carried out by logistic regression statistical tests to determine the effect of JJLS on changes in land use [34]- [36].
The second spatial data used in this research is the National Digital Elevation Model (DEMNAS) provided by the Government of Indonesia obtained from the Geospatial Information Agency (BIG) via https://tanahair.indonesia.go.id/demnas/#/.DEMNAS with a spatial resolution of approximately 8.3 m or 27-arc-second is converted to a Digital Terrain Model (DTM) using GIS filtering techniques.The DTM accuracy test was conducted according to the 2019 base map accuracy guidelines [37] based on field elevation measurements using a GNSS Receiver [38].
DTM and Sentinel-2A were processed using GIS to identify coastal morphology and then used as material for analyzing the third pattern of land use change through the GIS overlapping technique.DTM is convert-  1) based on expert judgment [38]- [41].
Data on slope gradient, surface roughness (2016 and 2022), and a 15 m wave height scenario on the shoreline based on previous research [6], [13], [14] are calculated using the formula developed by Berryman (2006) to create a tsunami inundation model.A spatial model of tsunami inundation refined using GIS to generate tsunami hazard zones for 2016 and 2022 [41], [42].The 2016 and 2022 tsunami hazard zones were compared with the GIS overlay technique to analyze hazard zone change patterns.The Hawke's Bay equation is shown in equation 1 [39].
) + 5sinS ...... (Equation 1) where H loss is the loss of tsunami water height per 1 m of inundation distance, H 0 is the height of the tsunami waves at the shoreline (m), S is the slope (in degrees), and n a is the surface roughness coefficient [10], [39].The Manning coefficient for surface roughness (n) in several land use classes is shown in Table 1.
The results of the analysis of land use change, tsunami hazard, suitability of land use to the RTRW, and morphology are used as the basis for preparing a land use pattern structuring plan.Literature studies on mitigation strategies by selecting land use in areas that have experienced a tsunami and environmental impacts are also considered in preparing a plan for structuring land use patterns [43]- [46].

RESULTS AND DISCUSSION Land Use Change
The results of land use identification in 2016 obtained 16 classes and increased to 17 land use classes in 2022.The accuracy test of land use data based on 210 field check samples has an overall accuracy of 84.8%.Using Sentinel-2A satellite imagery for land use identification has the same accuracy value as several similar studies [47], [48].The analysis of land use change found that there had been extensive changes in several land use classes in the three districts of the research area (figure 2).The highest percentage of land use change occurred in Mirit District is 5.4%.While the Ambal District is 2% and the Buluspesantren District is 1.7%.The land use that increased the most in the area was a 107ha shrimp pond.Agricultural land is the land use that has decreased the most in the area (174 ha).Many shrimp ponds were built in Mirit District (84 ha) and Ambal District (32 ha) on land previously serving as agricultural fields.There has been a change in agricultural land use to become shrimp ponds because the productivity potential of shrimp ponds is more economically profitable [49].The second order of land use that has increased in the area is the trade-industry sector of chicken farming and commercial buildings in coastal tourism (33 ha).Industrial trade development is evenly distributed in Mirit, Ambal, and Buluspesantren districts.Many chicken-trading industries were built on land that previously functioned as coconut plantations and mixed gardens.Other land uses that experienced an increase in area sequentially from largest to smallest were shrubs (30 ha), pavement (19 ha), fir forests (19 ha), shrubs (14 ha), mangrove forests (8 ha), salt ponds (5 ha), and settlements (3 ha).Analysis of the effect of JJLS on land use changes used a sample of land use that changed (Y=1) and did not change (Y=0) as the dependent variable (Y) and the distance (m) of the sample land use to JJLS as the independent variable (X).In Figure 3, 210 samples of land use are distributed (105 land use changes and 105 land use does not change) and used for statistical logistic regression tests.The simple logistic regression test results can be interpreted based on Table 2 (A and B).The significance value (sig.) in Table 2-A shows a value of 0.000 or less than 0.05 (sig <0.05), which informs that there is a significant influence of the independent variable (X) on the dependent variable (Y) [34].Then the Negelkerke R Square value in Table 2-B shows a coefficient of determination of 0.559, meaning that the independent variable (X) is a factor that influences 55.9% of changes in the value of the variable (Y) [34].The overall meaning based on statistical tests shows a significant effect of JJLS on land use changes that occur in parts of the coast of Kebumen Regency.Other factors that trigger changes in land use should be taken into account because the focus of the research object is JJLS as a strategic infrastructure built in the vulnerable period of 2016 to 2022.JJLS is an infrastructure with great potential to support progress in various fields of economic and social activity [50], [51].DEMNAS is converted to DTM through a selection process (GIS-based filtering) based on land use data and maintains a spatial resolution of 8.3 m.Land uses with significant differences in elevation from the elevation of the ground surface, such as forests, plantations, and settlements, are subject to elevation adjustments.Land uses such as shrubs, agricultural fields, and beach sand are considered to have an elevation close to the ground level.The DTM accuracy test based on 131 elevation measurement points in the field (figure 4A) shows a vertical accuracy (LE90) of 4.8 m and a horizontal accuracy (CE90) of 5.1 m.The DEMNAS accuracy test is based on the same 131 elevation measurement points, resulting in a vertical accuracy (LE90) of 6.2 m and a horizontal accuracy (CE90) of 5.1 m.Guided by SNI 8202-2019, DTM accuracy results are relevant to produce 1:25,000 scale maps, while DEMNAS accuracy is relevant to produce 1:50,000 maps.DTM is used to identify coastal morphology based on Standard Procedure Norms for Land System Mapping Criteria Scale 1:25,000/1:50,000 [52].DTM is converted to hillside, slope, and geomorphic to help identify coastal morphology [53], [54].The dominant morphology identified at the research site was ridges and valleys, alternating in an elongated pattern parallel to the coastline (figure 4B-4D).The elongated morphology of the basin/valley is sometimes flooded during the rainy season, as seen in the Sentinel-2A satellite image (figure 4D).The morphological stretches of alternating ridges and basins for approximately 3 km from the shoreline can be classified as beach ridges and swales.The pattern alternates between the beach ridge and swale approximately 100 m to 300 m apart.Then there is a dominant morphology of broad plains with flat to gentle slopes, and there are small river flow patterns that can be classified as alluvial plains.Identified the morphology of the relatively narrow plains on the right and left sides of the river, which are sometimes stagnant with water, can be classified as a floodplain.Then identified lagoons and spit at the estuary of the Wawar River and the estuary of the Lukulo River.The pattern of residential land use develops following the morphology of beach ridges starting from 1.5 km from the coastline.Most of the basin (swales) is used for agricultural land.Newly developed land uses such as industrial-chicken trading, and JJLS were also developed to suit the morphology of beach ridges.During the pattern of land use development for shrimp ponds, the majority lies in the morphology of elongated valleys (swales) starting from 500 m from the shoreline and partly located on a beach ridge near the estuary of the Lukulo River.Mangroves develop in lagoons and floodplains near the estuary of the Wawar River.Fir forests develop on the coastal plains associated with tourism sites.

Tsunami Hazard
The results of the tsunami inundation modeling in 2016 and 2022 resulted in a change in the area of the tsunami hazard zone (table 3).The 2016 tsunami inundation modeling resulted in a tsunami hazard zone of 1606.20 ha, and the results for 2022 increased to 1629.13 ha.The farthest range of tsunami inundation was 3.5 km from the coastline towards the Wawar River channel (figure 6).Spatial analysis of tsunami inundation (2016 and 2022) shows that there have been several changes in the extent of the tsunami hazard zone from the coastline in several areas (figure 6).The farthest range of the tsunami hazard is 3.5 km through the Wawar River channel and inundating the area on the west side of the Wawar River for 600 meters.The reach of the tsunami hazard zone through the Lukulo River channel is 1.5 km.
The difference in the range of the tsunami hazard in the two river channels is because the estuary of the Lukulo River is not parallel to the main river channel (the channel turns west near the coast).In contrast, the estuary of the Wawar River tends to be parallel to the main river channel.The farthest reach of the tsunami hazard zone is on the coast, which is not in contact with the river channel, about 780 m from the coastline.7E) occurred at 50 m to 100 m, where previously agricultural fields became fir forests and trade tourism; agricultural fields also turned into shrimp ponds at 550 m to 720 m.Analysis of the tsunami hazard zones from three different locations found that changes in the type and spatial pattern of land use impacted changes in the extent of the tsunami hazard zone.Types of land use, such as fir forest with a suitable surface roughness coefficient as a barrier to tsunami inundation, can be developed on the coast parallel to the coastline [45].Then shrimp ponds that have a poor surface roughness coefficient as a barrier to the rate of tsunami inundation should be developed in the safe zone.Most developments in land use from 2016 to 2022 followed the RTRW of Kebumen Regency, especially for agricultural fields, mangrove forests, and fir forests.Part of the research area (approximately 500 m from the coastline) is a defense and security area within the RTRW of Kebumen Regency, which has a strategic function for Indonesia.Some time ago, there was a conflict over land ownership and use between the community and the Indonesian National Armed Forces (TNI) in the area of defense and security [55].According to facts from an interview with a community member in the field, most of the conflicts between the community and the TNI have been resolved peacefully and fairly.
Then the development of shrimp ponds and the chicken-trading industry still needs to be considered regarding its environmental impact and space allocation [25].However, based on the results of an interview with one of the informants in the field, it is known that business actors and the local government have agreed on the area for the development of shrimp ponds and the chicken-trading industry to be allocated in the south side of JJLS.The results of community participation and agreement can be used to compile land use arrangements to reduce disaster risk [56].Open shrimp ponds would be better devel-oped with the concept of side-by-side with forestry/plantation plants (silvofishery) as a tsunami mitigation strategy [43].A comprehensive environmental impact assessment is needed by related parties so that environmental damage does not occur due to the development of shrimp ponds and the chickentrading industry [57].Then the development of mangrove forests, fir forests, and community forests in the suitable locations can function as a greenbelt/natural barrier to tsunami waves [43], [45].
According to the results of the 2016 and 2022 tsunami inundation model, JJLS is located in a relatively safe location parallel to the coastline.The road network can function as a barrier to tsunami waves, thereby minimizing losses in vulnerable areas based on the concept of structural mitigation [43], [44].Then the alternating pattern of coastal morphology between the beach ridges-sand dunes and the valleys (swales) that extend parallel to the coastline can serve as a natural barrier to tsunami waves.Wise use of the natural topography of an area can be useful for preventing a tsunami disaster [46].Conceptually and based on literature studies, an alternating pattern of land use that inhibits tsunamis and productive land use can be developed as an alternative input for tsunami mitigation and spatial planning (figure 8).

CONCLUSION
Road infrastructure is a factor that has a significant effect on land use changes.Multitemporal satellite imagery is very effective in identifying land use changes.Land use change can be used as a parameter for tsunami hazard analysis.Therefore, changes in the tsunami hazard zone are influenced by changes in land use on the coast.Changes in coastal morphology are physical parameters for tsunami hazard assessment that can be used for more comprehensive analysis in future studies.
The development of new land uses can have both positive and negative impacts on the potential tsunami hazard.Land use, such as shrimp ponds near the coast, is a land use that is not good at preventing tsunamis.If the land use for shrimp ponds develops in the tsunami hazard zone, the tsunami hazard zone will further expand inland.Therefore, the concept of structuring land use patterns can be used as a strategy in tsunami mitigation.The concept of land use arrangement can also be developed in more detail in the detailed spatial planning document.

Figure 2 .
Figure 2. Graph of Changes in Land Use Area in The Three Sub-districts in The Research Area S ource: Author Analysis (2022)

Figure 3 .
Figure 3. Spatial Distribution of Land Use Change Sample Points for Statistical Tests Source: Author Analysis (2022)

Figure 5 .
Figure 5. Spatial Distribution of Land Use Change to Coastal Morphology Source: Author Analysis (2022) The most expansive tsunami hazard zone is in Mirit District.Changes in the tsunami hazard zone (2016-2022) in Mirit District increased by 16.60 ha, Ambal District increased by 7.03 ha, and Buluspesantren District decreased by 0.7 ha.Agricultural land is the most significant land use in the tsunami hazard zone.Settlements that are included in the tsunami hazard zone are only found in Mirit District.Most industrial-commercial land uses included in the tsunami hazard zone are commercial buildings in tourism areas.Most shrimp ponds are included in the tsunami hazard zone because of their location close to the coastline.

Figure 7 .
Figure 7. Cross Section of The Tsunami Inundation Range at Three Locations on The South Coast of Kebumen Source: Author Analysis (2022)

Figure 8 .
Figure 8.The Concept of Structuring Land Use Patterns in Parts of The Coast of Kebumen Regency Source: Author Analysis (2022)

Table 2 .
Logistic Regression Statistical Test Results of The Effect of JJLS on Land Use Changes in 2016-2022 (A) Variables in the EquationSource: Data Analysis (2022)

Table 3 .
Changes in The Area of Tsunami Hazard Zone from 2016 to 2022

Landuse Hazard Area -District (ha) Hazard Area - Land Use (ha)
BAGUS PAMUNGKAS, DJATI MARDIATNO, AND ARRY RETNOWATI  IMPACT OF LAND USE CHANGES ON THE TSUNAMI HAZARDS IN PART OF COASTAL KEBUMEN