Characterisation of erosion/deposition processes due to lahars at Volcán de Colima, Mexico, with remote sensing techniques

Abstract

Lahars are the most abundant and dangerous secondary hazard in tropical volcanic environments. These rapid volcaniclastic flow events have the potential to destroy and bury pre-existing structures tens of kilometres away from their source. A recent study revealed that 17 % of all global historical volcano-related fatalities occurred due to lahars emphasising the importance of understanding the characteristics of those processes. Hence, investigations have been performed in many volcanic areas using different strategies and techniques including remote sensing. In this thesis, we will present a detailed study on erosion and deposition processes of lahars by the use of Synthetic Aperture Radar (SAR) data provided by the ESA’s SENTINEL-1 mission. As study area, the Montegrande ravine south of Volcán de Colima in western Mexico has been selected which is an optimal natural laboratory for studying lahars due to their high frequency during the rainy season and the existence of a systematic lahar detection program by Universidad Nacional Autónoma de México (UNAM). Furthermore, it provides perfect conditions for satellite observations because of an undisturbed line of sight into the ravine since a pyroclastic density current run down in July 2015 deforesting the entire ravine. In combination with the continuous SENTINEL-1 dataset since late 2014, it is possible to observe the evolution from a freshly filled to a deeply incised channel morphology due to the erosion by lahars. The proposed analysis methods are based on the principles of change detection by the use of SAR coherence data, which is the correlation coefficient between a pair of SAR images. To check the SENTINEL satellite’s monitoring capabilities in the ravine, a capability analysis considering multiple non-event image pairs from the rainy season 2017 has been performed confirming excellent conditions for monitoring applications along Montegrande due to SAR. Afterwards, a lahar event analysis has been performed by processing multiple non-events and event image pairs from the rainy seasons 2015 and 2017 according to an event catalogue revealed by seismic data provided by UNAM. To identify changes due to lahars, these datasets have been normalised using two methods. First, a normalised coherence difference ( ) has been calculated combining either a pre- or post-event with an event image pair. The second method combined pre- and post-event image pairs within the calculation of the logarithmic coherence ratio ( ). The obtained results showed that the -method is a more suitable solution for the detection of the spatial evolution of lahars. To observe the ravines evolution between 2015 to 2017, a seasonal analysis has been performed using dry and rainy season image pairs to obtain seasonal changes by the use of the - method as well. From this analysis, the process of decreasing lahar activity in the years after an eruptive phase has been confirmed, which has also been observed for recent volcanic activity at Volcán de Colima and also in other volcanic environments. Additionally, to these coherence approaches, seismic and optical data has been used for verification and validation of the obtained coherence results. Seismic data, however, has been used mainly for the selection of appropriate SAR images and a rough estimation of lahar magnitudes for interpretational purposes. Optical data has been obtained from two UAV datasets, used for morphological analysis in a specific portion of the ravine, and a terrestrial time-lapse camera (TLC), mounted on the UNAM monitoring station. Due to problems in the processing of the UAV data as well as limitations of applications of the TLC data, both datasets were used for interpretational purposes only.