Abstract

Permafrost is an important component of cryosphere, crucial in shaping landscapes and regulating ecological processes. A thorough analysis of permafrost modeling techniques is presented in this work, highlighting the application of machine learning(ML) and deep learning(DL) techniques in permafrost conditions. From studies investigating permafrost patterns throughout Swiss Alps to the formation of new permafrost in response to environmental changes in lakes like Zonag Lake, we analyze the effectiveness of ML and DL in capturing the complex dynamics of permafrost evolution. Despite notable achievements, challenges persist, including the need for more comprehensive training data, consideration of local drivers, and the integration of multidisciplinary approaches beyond traditional image processing. For all, the paper highlights the possibility of ML and DL to incorporate exploratory variables by leveraging remote sensing data and climate data, paving the way for enhanced understanding and prediction of permafrost dynamics in critical regions worldwide.

Authors

Anayet Ullah Dar, Muzafar Rasool, Assif Assad
Islamic University of Science and Technology, India

Keywords

Permafrost Mapping, Remote Sensing Data, Satellite Imagery, Artificial Intelligence