Digital Elevation Model (DEM) is a representation of elevation data that represent terrain with or without overlaying objects of the Earth. It is the ideal and most widely used method for determining topography. DEMs are generated from various techniques such as traditional Surveying, Photogrammetry, InSAR, LiDAR, Clinometry and radargrammetry. It has been observed that mostly LiDAR-generated DEMs provide the best accuracy. The unavailability of LiDAR data in most of the region restricts global researchers from high-resolution and accurate DEMs. The recent launch of ICESat-2 with a 13m beam footprint and 0.7m pulse interval, promises elevations at high orbital precision. Its accuracy is of the order of few centimeters in complex topography, because of this ICESat-2 proves to be a good source to generate high-accuracy DEMs. ICESat-2 provides discrete photon data with elevations of points on the Earth’s surface. Traditional interpolation techniques tend to over-smooth the estimated space and still are unable to justify the complicated continuity in the topographical data. Machine learning algorithms are widely being used to extract patterns and spatial extent in geographic data. To estimate a DEM from ICESat-2 LiDAR point data, machine learning regression algorithms are implemented in this study. The present study has been performed for a plain region of Ghaziabad, Uttar Pradesh, India. Studies have shown that Cartosat-1 DEM V3 R1 product provides an accuracy of the order of 2m in predominantly plain regions, hence taken for this region. Current work focuses on comparing various regression-based machine-learning techniques to interpolate DEM from ICESat-2 data The RMSE of the interpolated DEM resulted from the Gradient Boosting Regressor, Random Forest Regressor, Decision Tree Regressor, and Multi-Layer Perceptron Regressor was 7.13m, 7.01m, 7.15m, and 3.76m, respectively when evaluated against the TANDEM-X DEM of the same region. The MLP Regressor is found to perform the best among the four algorithms tested.
The Warwan sub-basin in Jammu and Kashmir is a remote glaciated region containing several glacial lakes and has recently experienced population growth and infrastructure development. Due to inaccessibility and geomorphic masking, multiple mass movement events, including avalanches and glacial lake outburst floods (GLOFs), have remained largely unreported. This study analyzes three major avalanche events from the past two decades. The September 2005 and September 2020 avalanches originated from glaciers GL-B and GL-A within the same glacier complex, while the March 2020 rock–ice avalanche initiated from the headwall of glacier GL-F in an adjacent valley and terminated before reaching its ablation zone. Runout mapping shows that the September 2020 avalanche descended from the headwall of GL-A and impacted its proglacial lake, triggering a GLOF. In contrast, the September 2005 event terminated before reaching the glacial lake that began forming in 1999 at the terminus of GL-B. Geomorphic analysis indicates persistent sediment influx from meltwater streams of GL-D and GL-E into the lake associated with GL-A, progressively infilling the lake. Continued glacier retreat caused meltwater to accumulate behind the sediment infill, dividing the lake into two disconnected lakes. Impact from the September 2020 avalanche led to complete drainage of stored water and sediment, generating a downstream debris flow. Pre- and post-GLOF imagery reveal breaching and widening of the outflow channel and deposition of a debris fan downstream. The repeated history of mass movements and continued growth of glacial lakes raise serious hazard concerns in the Warwan sub-basin. GLOF modeling identifies Lake-B as a potentially hazardous lake in its present condition, posing significant downstream risk. Settlements such as Youdu and Qaderna, along with bridges, roads, and residential structures, lie within potential GLOF inundation zones. High sediment availability from past mobilization and deposition along the valley increases the potential of debris-flow cascades, posing elevated risks to downstream infrastructure and hydropower, underscoring the need for preparedness and mitigation planning. GLOF modeling suggests that an early warning system closer to the source will be able to provide a good lead time in case of potential GLOFs in the valley.