Tomographic reconstruction of atmospheric water vapour using a temporal variational approach during central European floods in July 2021

Abstract

GNSS tropospheric tomography is a striking remote-sensing tool to monitor the spatiotemporal behavior of the troposphere, specifically of highly variable water vapour. This technique is generally advantageous compared to conventional data sources such as radiosonde and water vapour radiometers (WVR) due to its capability to continuous monitor Earth’s atmosphere during all weather conditions. Hence, the products of tropospheric tomography can play an influential role in meteorological and climatological studies. In July 2021, a series of severe storms and intense rainfall led to floods in European countries like Germany, Belgium and Luxembourg. This costed hundreds of lives and caused widespread damage to various structures. Therefore, this research aims to define a tomography model with a varying temporal resolution based on the physical atmospheric conditions and satellite’s geometry. In order to optimize the temporal resolution of the tomography model, four distinct features are taken into consideration: horizontal gradients (East-West and North-South components), water vapour pressure variations, and percentages of intersected voxels derived from the resolution matrix. The tomography solution is also reconstructed hourly to provide a better understanding of the retrieved wet refractivity field with dynamic temporal resolution. Due to the heavy rainfall between July 13 and 15, we consider two days before and after this core period when assessing tomography model temporal variability. To evaluate the effectiveness of the tomography model, we compare the retrieved wet refractivity tomography profiles with radiosonde profiles, as well as comparing SWD retrieved by ray-tracing through the tomography model with GNSS SWDs.