Hazards Mitigation of Lahar Flows on Semeru Volcano after the 4 December 2021 Eruption Based on PS-InSAR

,

rainwater and form debris flow or "lahar." Research by the Earth Observatory of Singapore -Remote Sensing Lab (EOS-RS) shows that the area on the surface hit by pyroclastic flows and lahars in the southeastern sector of Semeru reaches Curah Kobokan. Compared to the eruption in 2020, the eruption on December 4, 2021, is classified as a fairly large eruption, even though it has almost the same hot cloud glide distance of 11 km. Satellite monitoring of volcanic ash in Darwin (Darwin Volcanic Ash Advisory Centre, VAAC Darwin) recorded that the eruption column reached 15 km above sea level at 4:30 PM (Sennert, 2021).
The changing volcanic edivice after eruption and accumulation of the deposit area near the flank should be monitored over time as a mitigation plan. In most cases, edifice instability develops as a result of one or more of several causes such as magma emplacement, slope overloading or oversteepening, and peripheral erosion.
Understanding the possible size of structural collapses and the related eruptive activity has significant consequences for the development of monitoring and hazard mitigation techniques for vulnerable volcanoes, which must now account for the probability of future collapse events (McGuire, 1996). Since 1980, Interferometry Synthetic Aperture Radar (InSAR) has been widely used to monitor earth deformations such as earthquakes (Massonnet et al., 1993), movement of permafrost (Goldstein, et al., 1993), volcanoes (Massonnet, et al., 1995), landslides (Amelung, et al., 1999), and underground water flows. In volcano monitoring, the practical application of lithological and geomorphological changes to the Earth's surface in damage assessment and disaster monitoring is crucial (Wadge et al., 2002;Stramondo et al., 2006;Saepuloh et al., 2010;Bignami et al., 2013;Solikhin et al., 2015). Here, we present our new results from PS-InSAR of surface deformation after the eruption of 2021. The highest deformation rate was sampled from the upslope and downslope, and its changes due to the eruption were investigated. High deformation is assumed to be the most rapid sedimentation in catchments, and field observations are used to validate the data to gain information related to the sediment package. Finally, we propose a possible hazard-related lahar that may occur as a consequence of unconsolidated sedimentation material.

Methods
The data required for PS-InSAR processing is Sentinel-1 Single Look Complex (SLC) data, DEM data, and orbital data, which can be downloaded free of charge from https://vertex.daac.asf.alaska.edu/. The satellite data range of 18 data starts from January 2021 to June 2022, with one data selected as master, namely November 2021, and the other data as slaves. All these data come from the Sentinel-1A satellite with ascending orientation (southnorth), which has a wavelength of 0.56 cm or is classified as C-band type and polarization in the form of VV and has a resolution of 5 × 20 m because it is an interferometric wide swath mode (Flores et al., 2019).
The calculation of the coregistration and interferogram produces coregistration data, which is a match between two images, and an interferogram that contains the phase difference between the master and slave. The phase difference in question is the phase difference between the emitted wave and the reflected wave which describes the delay time between the two waves (Ferreti et al, 2007). Phase difference from one pair wave can be used to calculate the difference of acquisition object by using the equation (1) below: where Ø is the phase difference for each interferogram, λ is the wavelength, and R is the distance between the acquisition objects. In addition to the surface deformation components, the interferogram (ϕLOS) phase also contains other components, namely ∆ϕgeom, which is related to the difference in the geometry of the satellite's view, phase difference from the topography ∆ϕtopo, and phase difference from the atmospheric refraction ∆ϕatm. As described in Equation (2), Phase difference from one pair wave can be used to calculate the difference of acquisition object by using t Where Ø is the phase difference for each interferogram, λ is the wavelength, and R is the distance of the acquisitio Figure 1.

Data Processing Steps and Results
The geometry factor is solved by applying the exact orbit and estimating the perpendicular baseline between satellite orbits. This effect was eliminated by assuming ellipsoidal smooth earth. Errors in predictable satellite orbits due to solar radiation pressure. DEM was used in processing to eliminate the phases introduced by the topography. The largest source of phase error was caused by variations in atmospheric refraction between ∆ϕatm acquisitions. The ϕ error is an additional noise that describes the phase noise caused by scattering and heat noise from instruments (Parker, 2017).
The results of coregistration and interferogram were exported so that they could be further processed using the Stanford Method for Persistent Scatterers (StaMPS) method with MATLAB software. StaMPS is a PS-InSAR processing method that uses a set of C++-based software and Matlab scripts regularly according to the Hooper et al. (2012) algorithm, which was developed so that it can be used even if applied in terrain without man-made structures or irregular deformations (Hooper et al., 2018).
The StaMPS used in this study consists of seven stages. The first stage is data loading, which is used to input the coregistration and interferogram data in the format and location corresponding to the StaMPS algorithm. Phase noise value estimation of each pixel candidate of each interferogram was performed based on several parameters, including maximum topographic error, pixel size, the weighting of each pixel, filter combined low-pass, adaptive phase, and the maximum number of iterations (Hooper et al., 2010). The third step in stamps is the PS selection process, which is selected based on the phase-noise criteria in the previous stage. This stage also estimates the percentage density of other pixels (non-PS). The fourth step is the PS weeding stage, which sorts and eliminates pixels that are not selected. Next, phase correction corrects the phase of spatially correlated pixels. Sixth-stage stamps for converting interferograms that are unitary -π to π are converted into absolute phases of integer multiplier 2π, because they are more related to topography and deformation. Phase variation between two points on a flattened interferogram provides a measurement of the actual height variation, having removed all integers from the height (Ferreti et al, 2007). Next, the estimated stage spatially correlates the look angle error to calculate the error look angle that has been correlated in stages three and five. From the processing of stamps 1-7, a map of the distribution of LOS over time was obtained. Based on the results of the data processing, interpretation and mitigation were carried out by considering geological data. The entire data-processing stage is shown in Figure 1.

Results
The results of Sentinel satellite image data processing using the PS-InSAR method are shown in Figure 2. The range of data used in this study was from January 2021 to June 2022, covering the Mount Semeru area. The selection of the data range was based on the last eruption of Mt. Semeru, which occurred on December 4, 2021. Data before and after the eruption were used to observe the deformation that occurred.
The deformation that occurs in the body of Mount Semeru based on the results of processing using the PS-InSAR method shows that the speed of deformation that occurs at the location of this study is indicated by the LOS displacement value ranging from -40 mm/year to 40 mm/year. From the figure, it can be seen that the North, West, and South areas are relatively not deformed. The eastern area shows little deformation as indicated by the yellow to light green PS-InSAR points, which reveals that the LOS displacement rate ranges from 0 to -10 mm/year.
The area in the southeastern part shows significant deformation, as indicated by the dark red to dark blue PS-InSAR points with LOS displacement values ranging from -40 mm/year to 40 mm/year. The results of this study correspond with the eruption that occurred on December 4, 2021, where the crater opening and pyroclastic material flowed southeast. The Southeast area is the focus of the discussion of the results of this study.

Significant Deformation in the Southeast Area
The zones with significant deformation in the southeast can be grouped into two categories. The first category is the zone with yellow to dark red PS-InSAR dots that are relatively close to the center of the Semeru eruption.  size from gravel to boulders (diameter up to 1.5 m), while the matrix is composed predominantly of sand-togravel-sized materials (< 4 mm). Unconsolidated materials have the potential to produce lahar flows when mixed with water (Parfitt and Wilson, 2008).

Mitigation of Potential Lahar Flow Hazard Zones
Lahar or debris flows consist of a mixture of water and solids that move toward a lower area. In volcanic areas, the solid material that makes up the lahar is generally composed of material from a volcanic eruption. According to Parfitt and Wilson (2008), lahars are a long-term threat after a volcanic eruption, especially if the material resulting from the eruption is dominated by unconsolidated material. The materials easily mix with water and become a lahar flow. In Semeru, lahar flows are often triggered by rain, which causes runoff to mix with unconsolidated pyroclastic material moving along the river channels. Therefore, the amount of material accumulated and the distance to the river flow were used as parameters to determine the potential hazard zone against lahars in this study.
PS-InSAR deformation data overlaid with a map of the drainage pattern and the existence of residential areas were used to mitigate the potential hazards of lahars. Based on these considerations, the locations that have a high-risk potential of being affected by lahar flows are Curah Kobokan, Supiturang Village, Pronojiwo District, and Tulungrejo, Pasropan Village, Pasrujambe District, Lumajang Regency. The area marked with a yellow circle in Figure 6 is located around the river, where there is a large accumulation of pyroclastic material upstream, based on the results of the PS-InSAR deformation processing. The results of this study are expected to be considered by authorities when handling potential lahar flows around Semeru Volcano.

Conclusions
Based on this research, it can be concluded that the zone of significant deformation was located in the direction of the opening of the Semeru eruption (southeast). The area with negative deformation was closer to the peak indicated by the LOS displacement from -10 mm/y to -40 mm/y. This was probably caused by the scoured material coming out of the crater. This material then accumulates in the zone of positive deformation located in the downstream area, which is indicated by the LOS displacement from 10 to 40 mm/y. An area with positive deformation was located downstream. The deformation appears significant, especially after the eruption on December 4, 2021. Areas with high potential to be affected by debris flow (lahar) are Curah Kobokan, Supiturang Village, Pronojiwo District, then Tulungrejo, Pasropan Village, Pasrujambe District, Lumajang Regency.