Numerical experiments on the dynamics of channelised lava flows at Mount Cameroon volcano with the FLOWGO thermo-rheological model

Abstract

As for many other effusive volcanoes, Mount Cameroon (MC) is a volcano for which only limited information exists on lava flow properties and emplacement dynamics for recent eruptions. This study provides new quantitative constraints for rheological and dynamic characteristics of lava flow effusion for the 1982 and 2000 eruptive events, used to calibrate the FLOWGO thermo-rheological model for these lava flows. The FLOWGO 1-D physical model is used to simulate down-flow evolution of the geometry and rheology of channel-contained cooling-limited lava flows. Morphometric data from historical lava flows were acquired from the field, e.g. channel geometry, levee and background slope, in order to estimate lava yield strength, velocity and effusion rate. Lava density and viscosity were also estimated from compositional data and laboratory methods. To account for uncertainty in the input rheological and geometrical data, three end-member scenarios were used to bracket the potential range in lava channel initial dimension, initial lava temperature and phenocryst content. For each of these scenarios, two crustal growth models were used: one assuming strong insulation due to lava flow surface crusting, the other a much lower crusting rate. Twelve numerical simulations were made per flow and the results were compared against the channel geometry, microlite content, yield strength and viscosity estimates from field and laboratory investigations. Best-fit models where obtained for both the 1982 and 2000 lava flows using a low rate of surface crusting, high initial temperature and low phenocryst content. Model-predicted lengths were within 5% of the actual lengths. Modelled mean effusion rates for the 1982 (52–64m3s−1) and 2000 (10m3s−1) flows closely matched field data derived estimations (26–68 and 9.5m3s−1 respectively). FLOWGO model results are highly sensitive to initial channel dimensions, phenocryst content and the FLOWGO model is unable to reproduce the observed rapid near-vent increase in microlite content. Near-vent rapid cooling induced by lava degassing is yet to be fully documented and integrated into numerical lava flow models such as FLOWGO, but obtained results already provide valuable insights for lava flow hazard assessment at Mount Cameroon.