Land subsidence and coastal retreat as observed using Sentinel-1, tide dynamics, and socio-cultural survey of Sayung, Demak, Indonesia

The coastal zone of northern Java experiences rapid land subsidence. Satellite-derived InSAR data and GPS had provided evidence  displaying the spatio-temporal rates of the subsidence. Interestingly, inhabitants do not always recognize this phenomenon. However, they do recognize that flooding from the sea (called ‘rob’) has changed from gift to disaster, viz. from shallow infrequent flooding that carry inshore fishes, to regular deep and widespread disastrous floods. In the Demak regency, more than 300 families have moved to locations further inland. In some locations rob flooding has caused a coastline retreat of up to 5kms.

Our initial Sentinel-2 analysis comparing 2016 and 2022 images have showed the change of ecosystem in the area. In addition to that, social surveys revealed timely change of the ecosystem and point places where coastlines have moved.

For this study, we use Sentinel-1 imagery for SAR coupled with tide data to determine the behavior of coastline change from 2018-2022. Sentinel-1 is selected due to the temporal resolution and performance in  all-weather. Ground Range Detected (GRD) of Sentinel-1, is single look complex data projected using an earth ellipsoid, will be used as main input in spatial analysis to observe shoreline condition. There are 4 types of polarization of GRD data: HH,VV, HH+HV and VV+HV. Polarization type will be tested and determined to get less noise images result. Image masking and pixel analysis will sequentially be conducted to segregate the land and the sea. The analysis will be managed in Sentinel Application Platform (SNAP).

Tide data plot from nearby station is used to assign at what phase Sentinel-1 imagery is acquired. Additional analysis from tide data plot is applied to predict land/inundation at important sites and when imagery is not available. 

The social research component of the study through preliminary survey data , our initial survey showed residents relocation had been started before the period covered by the SAR data sets. Hence, random purposive sampling will be used for a follow-up mixed-methods survey to gather more data on social and economic dynamics arising from the geomorphological shifts in coastline. 

Ultimately, we hope  our study will help unravel land subsidence and the resulting coastline retreat to provide a strong basis for mitigation and adaptation options by the people and the government.