Abstract
AbstractThe western Nyainqêntanglha Range on the Tibetan Plateau reaches an elevation of 7,162 m and is characterized by an extensive periglacial environment under semi‐arid climatic conditions. Rock glaciers play an important part of the water budget in high mountain areas and recent studies suggest that they may even act as climate‐resistant water storages. In this study we present the first rock glacier inventory of this region containing 1,433 rock glaciers over an area of 4,622 km. To create the most reliable inventory we combine manually created rock glacier outlines with an automated classification approach. The manual outlines were generated based on surface elevation data, optical satellite imagery and a surface velocity estimation. This estimation was generated via InSAR time series analysis with Sentinel‐1 data from 2016 to 2019. Our pixel‐based automated classification was able to correctly identify 87.8% of all rock glaciers in the study area at a true positive rate of 69.5%. In total, 65.9% of rock glaciers are classified as transitional with surface velocities of 1–10 cm/yr. In total, 18.5% are classified as active with higher velocities of up to 87 cm/yr. The southern windward side of the mountain range contains more numerous and more active rock glaciers. We attribute this to higher moisture availability supplied by the Indian Monsoon.
Highlights
We present an inventory of actively moving rock glaciers within the western Nyainqêntanglha Range on the southeastern true positive rate (TP) (Figure 1)
We evaluate the impact of the Topographic position index (TPI) with a radius of 200 m and with a radius of 800 m
After identifying the most suitable features for the classification, we evaluated the performance of the automated classification compared to the entire manually created rock glacier inventory
Summary
Rock glaciers are lobate or tongue-shaped landforms developing through gravity-driven creep of frozen debris and interstitial ice.[1,2,3] Rock glaciers play an important part in the water budget of high mountain areas.[4]. We selected a temporal subset of 14 interferograms for ascending and 13 interferograms for descending geometry from October 2017 to February 2018 This time period displays good coherence values throughout the study area and we used the SBAS results of this shorter period to fill in spatial data gaps in our 3-yr time series results. Multiple studies in recent years have evaluated different classification algorithms and varying input parameters to automatically identify and classify rock glaciers.[40,41,42,43,44] In addition to the manual approach described above, we performed a pixel-based supervised maximum-likelihood. We performed the classification again with the five initial features and different combinations of the remaining features to determine the optimal feature combination
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
