Evaluation of Double Fusion Satellite Rainfall Dataset in Establish Rainfall Thresholds for Landslide Occurrences Over Badung Regency-Bali

Rainfall stations provide reliable rainfall data, but their availability is limited in mountainous areas, complex terrain, and remote areas. Satellite rainfall datasets (SRDs) provide high-resolution worldwide rainfall estimation, which has the potential to be used in identifying rainfall conditions that trigger landslides. Landslides can be predicted through rainfall threshold modeling, serving as an early warning system. It is essential to validate the chosen threshold model to assess the accuracy of forecasting landslide occurrences triggered by rainfall events. This study aims to evaluate the effectiveness of a dual fusion approach, utilizing two SRDs, in establishing rainfall thresholds for landslide prediction in the Badung regency over the period from 2015 to 2022. Rainfall threshold analysis in this investigation focuses on cumulative rainfall events occurring 5, 7, 10, and 15 days prior to the onset of landslides. The first fusion was established through the application of the cumulative distribution functions method, involving a comparison between the SRDs and the datasets from rain gauges. Subsequently, the analysis transitioned to the second fusion, where a weighted correlation coefficient function was employed to assess the connections between rain gauges and individual SRDs. The results illustrate that the second fusion of SRDs yields an area under the curve value of 0.82 for a 15-day cumulative rainfall, surpassing the performance of the first fusion. The first quartile approach demonstrates the highest accuracy compared to alternative methods, providing a reliable estimation of landslide occurrences with minimal error.