Satellite-Based Rainfall Thresholds for Landslide Initiation in Indonesia: A Systematic Literature Review

Abstract

Rainfall thresholds are widely recognized as a key element in landslide early warning systems, as they help identify rainfall conditions under which slope failure is likely to occur. In Indonesia, the need to develop reliable rainfall thresholds is particularly urgent, given the high frequency of landslide events, the country’s highly diverse physiographic characteristics, and the uneven distribution of rain gauges across many landslide-prone areas. At the same time, recent advances in satellite precipitation products, including IMERG, GSMaP, PERSIANN, and CMORPH, provide new opportunities to develop rainfall thresholds using datasets with broader spatial coverage. This study employs a qualitative approach through a Systematic Literature Review (SLR) to examine the evolution of satellite-based rainfall threshold research for landslide initiation in Indonesia. Relevant studies were systematically selected following the PRISMA 2020 framework. The selected literature was then extracted and analyzed using thematic synthesis, with particular attention to study context, satellite precipitation products, threshold formulations, validation approaches, and their potential application in landslide early warning systems. The synthesis identifies five major patterns. First, research in Indonesia has gradually shifted from the use of a single national rainfall threshold toward more localized and dynamic threshold models. Second, satellite precipitation data have become increasingly important as an alternative data source for areas with limited ground based observations. Third, the performance of satellite precipitation products is highly context-dependent, varying across regions, spatial and temporal resolutions, and validation procedures. Fourth, the Bali/Badung research cluster reflects a more advanced methodological trajectory, as shown by the evaluation of gridded rainfall datasets, high-resolution datasets, double-fusion datasets, and comparative assessments between IMERG and MSWEP. Fifth, the main challenge is no longer merely to establish rainfall thresholds, but to translate them into early warning systems that are both scientifically robust and operationally practical. This study concludes that satellite based rainfall thresholds for landslide initiation in Indonesia should be developed in a context-specific manner, validated using local evidence, and oriented toward multi-level warning system designs that can bridge scientific accuracy with institutional usability.