Glaciers are under pressure in the current climate warming trend. The aerial extent and mass balance, a measure of mass loss or mass gain of a glacier, represent the “health state” of a glacier. They are recognized as essential climate variables by the world meteorological organization. Due to the sheer number of glaciers of roughly 200.000 worldwide and their inaccessibility, satellite data have been a valuable source to map glacier area and glacier mass balance around the globe during the last decades.
Earth observation satellites, including optical and Synthetic Aperture Radar (SAR) sensors, allow for systematic and continuous glacier monitoring worldwide. Satellite images have enabled researchers to map the extent of glaciers and glacier surface types (glacier facies, i.e., snow, firn, ice). Multi-temporal information on the glacier surface elevation and proxies (snow cover, albedo) which are highly correlated with the glacier mass balance, have been used to estimate mass changes. Today, the ever-growing amount of satellite data motivates the development of new methodologies based on machine learning to optimize the information retrieved from large data sets of multi-modal and multi-temporal satellite data.
The PhD research activity aims to design deep learning methods to automatically analyze temporal series of remotely sensed data and map glacier area, glacier surface characteristics, and infer glacier mass balance. The PhD candidate will take advantage of the latest developments in convolutional neural networks to design deep architectures capable of fusing optical and SAR data for glacier facies analysis within the same network. Combining multiple data sources within a big geodata analysis framework, the candidate will design a deep learning approach to infer the glacier mass change.