FROM INFRARED TO PASSIVE MICROWAVE: DEEP LEARNING EMULATION WITH UNCERTAINTY DIAGNOSTICS
Marzo 25, 2026ON THE DETECTION OF THE ROTATIONAL CENTER OF MEDICANES: COMPARISON BETWEEN MERCAD AND ARCHER
Marzo 25, 2026H. Xu1
1Beijing University of Technology
Mass loss from polar ice sheet is one of the major contributors to the global mean sea level rise. Studies have shown that half of the mass loss from Greenland Ice sheet is due to melt water run-off. However, the question of how the melt water goes into the sea still needs to be answered. Perennial Firn aquifer was first discovered in the Arctic Circle Transverse(ACT) mission in 2011. In situ and airborne Ground Penetration Radar(GPR) in the consecutive studies have revealed its extensive distributions over the percolation zone. Studies over the aquifer water table level changes have shown that this change is highly related to the surface melt events. Such studies are valuable but are limited by the coarse weather, high cost and distance travels to the polar regions. Microwave remote sensing satellite data has shown its sensitivity to the existence of firn aquifer, but efforts are still needed in the modeling studies of the microwave data signatures both in polarization and angular dependence. With the launch of NISAR, using the combined active and passive method for the study of firn aquifer becomes possible.
This paper presents the L-band modeling work for the active and passive observables of the firn aquifer region. Vector radiative transfer model based on 3-D random media description is used for the incoherent emission and scattering modeling. Results have shown that the forward models can explain the microwave data well. This work provides a theoretical basis for the quantitative study of firn aquifers with microwave remote sensing data.
