Development of regional landslide early warning system based on rainfall threshold
Landslides are mainly triggered by rainfall. Daily as well as antecedent (cumulative) rainfalls are responsible for saturation of soil and subsequent landslide in hilly areas. Therefore, a rainfall intensity-duration relationship was developed by NRSC for experimental landslide early warning. Bhuvan platform is currently used to disseminate daily landslide early warning for selected road corridors in the Himalayas.
Daily 72 hrs rainfall forecast received from IMD or SAC (MOSDAC) is used as an input to the landslide early warning model. Then the available landslide susceptibility map is integrated with rainfall probability to generate landslide early warning in Bhuvan. Early warning in four categories (Warn, watch, advisory and no warning) are issued regularly through Bhuvan. Currently the landslide earlywarning system is operational experimentally in selected routes of Himachal Pradesh, Uttarakhand, Meghalaya, Assam and Mizoram. Further development of rainfall intensity-duration relationship is required specific to different landslide prone areas of India.
Automatic identification of earthquake and landslide induced terrain changes from high resolution satellite data
Earthquake cause subsidence or upliftment of land in the seismic prone areas. Similarly, landslides, particularly the creeps (slow moving landslides) cause persistent damage to raod and buildings in hilly areas. Object-based change detection technique was used to identify new landslides using Resourcesat-2/2A LISS-IV Mx satellite data. CartoDEM (10 m) was additionally used to eliminate several false positives using slope criteria. Characterisation and identification of terrain changes due to earthquakes and creeps can be done effectively by differential SAR interferometric techniques. The outputs of the same in the terms of displacement maps and velocity fields with are used to model the prediction of debris flow path in the case of complete slope failure. Advanced processing methods like PsINSAR or STamPS will be attempted to overcome problem related to coherence loss or atmospheric interferences.
