Surface deformation monitoring and forecasting of sinabung volcano using interferometry synthetic aperture radar and forest-based algorithm

Abstract

The Sinabung volcano on Sumatra Island stands out as one of the most active volcanos, having recorded the highest number of eruptions since it resumed activity in 2010. The eruptive activities have caused significant deformations on the volcano's surface. This research aimed to analyze, cluster, and forecast its deformation patterns based on Sentinel-1 A time series data from 2016 to 2023. The differential interferometry synthetic aperture radar (DInSAR) technique was used to monitor monthly deformations and to create time series data. A forest-based forecast (FBF) model was used to predict the rate of changes in volcano surface inflation from January 2024 to December 2027. The deformation times series patterns were also analyzed and clustered into three regions to reveal areas with similar deformation behaviors. The results indicated that Mount Sinabung's deformation is an overall continuous sporadic phenomenon where random ground inflation and deflation were recorded throughout the area with an average deformation rate ranging from 0.06 to 0.32 cm/month and an overall average of 0.197 cm/month with a standard deviation of 0.96 cm, confirming that the volcano is inflating. The highest single-pixel monthly inflation of 4.62 cm was recorded in 2023, while the highest deflation occurred in 2018 at −4.58 cm. The FBF model predicted a gradual and increasing inflationary pattern at the rate of 0.54 cm/month for 2024–2027, higher than the average of the observed data. The deformation within the lava dome and caldera poses a significant risk and could lead to wall collapses and landslides in the crater dome, potentially triggering explosive eruptions. The outcomes of this research serve as valuable supporting information and offer an early warning of potential volcanic disasters in the future.