Nyiragongo crater collapses measured by multi-sensor SAR amplitude time series

Crater morphology at active volcanoes can change rapidly. Quantifying changes during the course of a volcanic unrest episode may help assess the level of volcanic activity. However, limitations such as crater accessibility, cloud cover or intra-crater eruptive activity may hamper regular optical or on-site crater monitoring. Here we use multi-sensor satellite Synthetic Aperture Radar (SAR) imagery to produce dense time series of quantitative indicators of crater morphological changes. High temporal resolution is achieved by combining images from a variety of sensors and acquisition modes, though the diversity of acquisition geometries (incidence angle, viewing direction, resolution…) prevents direct comparison between the different images. Using basic trigonometry assumptions, we develop PickCraterSAR, an open-access tool written in Python to measure crater radius and depth from SAR amplitude images in radar geometry. We apply our methodology to study the crater collapse associated with the May 2021 and January 2002 eruptions of Nyiragongo volcano. After the 2021 collapse, we estimate the maximum depth of the crater to be 850 m below the rim and the total volume to be 84$\pm$10 Mm$^3$ (270 m deeper but only 15-20 \% more voluminous than the post-2002 eruption crater). We also show that the 2021 crater collapse occurred progressively while a dike intrusion was migrating toward the south.