The complexities of assessing volcanic hazards along the Cameroon Volcanic Line using spatial distribution of monogenetic volcanoes

International audienceVolcanic eruptions represent hazards for local communities and infrastructure. Monogenetic volcanoes (usually) erupt only once, and then volcanic activity moves to another location, making quantitative assessment of eruptive hazards challenging. Spatio-temporal patterns in the occurrence of these eruptions may provide valuable information on locations more likely to host future eruptions within monogenetic volcanic fields. While the eruption histories of many stratovolcanoes along the Cameroon Volcanic Line (CVL) are relatively well studied, only fragmentary data exist on the distribution and timing of this region's extensive monogenetic volcanism (scoria cones, tuff rings, maars). Here, we present for the first time a catalog of monogenetic vents on the CVL. These were identified by their characteristic morphologies using field knowledge, the global SRTM Digital Elevation Model (30 m resolution), and satellite imagery. More than ~1100 scoria cones and 50 maars/tuff rings were identified and divided into eight monogenetic volcanic fields based on the visual assessment of clustering and geological information. Spatial analyses show a large range of areal densities between the volcanic fields from >0.2 km-2 to 0.02 km-2 from the southwest towards the northeast. This finding is in general agreement with previous observations, indicating closely spaced and smaller edifices typical of fissure-fed eruptions on the flanks of Bioko and Mt. Cameroon in the southwest, and a more focused plumbing system resulting in larger edifices of lower spatial density towards the northeast. Spatial patterns were smoothed via kernel density estimates (KDE) using the Summed Asymptotic Mean Squared Error (SAMSE) bandwidth estimator, the results of which may provide an uncertainty range for a first-order hazard assessment of vent opening probability along the CVL. Due to the scarce chronological data and the complex structural controls across the region, it was not possible to estimate the number of vents formed during the same eruptive events. Similarly, the percentage of hidden (buried, eroded) vents could not be assessed with any acceptable statistical certainty. Furthermore, the impact of different approaches (convex hull, minimum area rectangle and ellipse, KDE isopaches) to define volcanic field boundaries on the spatial distribution of vents was tested. While the KDE boundary definition appears to reflect the structure of a monogenetic volcanic field better than other approaches, no ideal boundary definition was found. Finally, the dimension of scoria cones (approximated by their basal diameters) across the CVL was contrasted to the specific geodynamic setting. This region presents a complex problem for volcanic hazard analysis that cannot be solved through basic statistical methods and, thus, provides a potential testbed for novel, multi-disciplinary approaches