Predicting the occurrence of a landslide event, both where and when it will occur, remains a challenge. However, using remote sensing techniques, satellite data and numerical simulation, broad predictions about the day and path of landslides can be estimated, which can assist with disaster management and subsequent recovery.

The National Remote Sensing Centre (NRSC) is one of the primary centers of Indian Space Research Organisation (ISRO), Department of Space (DOS). The NRSC has ground stations that receive satellite data, generate data products, disseminate information to users and develop techniques for remote sensing applications. This world-renowned center is primarily focused on disaster management support and geospatial services.

Priyom Roy and his colleagues from the NRSC have been researching landslide behavior and the ability to predict landslide time and path from satellite data. In their recent article, “Time and Path Prediction of Landslides Using InSAR and Flow Model ,” published in Remote Sensing of Environment, January 17, 2022, they present a novel method for predicting landslides using ENVI® SARscape and Sentinel-1 data over two landslide regions in India (Figure 1).

ENVI SARscape works with all commercially available SAR data, as well as many noncommercial SAR datasets, and has automated tools to help quickly and easily prepare, view and analyze SAR data. SAR data not only provides amplitude, or the intensity of the backscatter response, but also phase, which allows for measurement of height and displacement – a unique benefit of SAR.

ENVI SARscape integrates point and area-based analysis techniques to measure displacement and deformation over time. This approach makes it possible to analyze deformation that affects both extended and localized structures related to natural or other phenomena, such as landslides.

Fig. 1. a. Pre-event satellite image of Kikruma Landslide b. Post event satellite image of Kikruma Landslide. Pre-event satellite image of Kotropi Landslide. d. Post event satellite image of Kotropi Landslide. Insets show location of landslide on India map. (Source: Google Earth).

The availability of open-source Sentinel-1 data has revolutionized the study involving landslides and predicting the time of failure. In their paper, the authors present a novel method for time and path prediction of landslides using two large landslides (Kikruma and Kotropi) located in the Himalayas in India. The Sentinel-1 data stack was processed using the Persistent Scatterer and Small Baseline Subset interferometric techniques from ENVI SARscape to analyze the trend of ground deformation leading to slope failure.

Since the image covers a 250 km swath, the subset covering the landslide was selected using the “sample selection SAR geometry tool,” which is a predefined module in ENVI SARscape and is used to subset the data using a user defined area of interest.

Figure 2. The process flowchart from SAR data to displacement map.

The authors demonstrated the ability to use SAR analysis techniques to aid in the prediction of landslides. They observed that in both cases, methods were able predict the day or a few days prior to actual landslide. In addition, they were also able to model the probable geometry and flow path of the landslides. As a result, they were able to understand that the instability had commenced almost a year prior to failure in both landslide instances. Further, due to the regularity in the acquisition of SAR imagery from Sentinel-1 satellites, it is now possible to continuously monitor ground deformation in a region.

Fig. 3. Material acceleration map for a. Kikruma and b. Kotropi landslides; Classified acceleration maps for c. Kikruma and d. Kotropi landslides and Release area. Maps for e. Kikruma and f. Kotropi landslides.

In summary, SAR data and ENVI SARscape can be used to detect potential landslides, select sites for hazard mitigation and provide early warnings to trigger engineering mitigation treatment.