Abstract
Lahars are mass flows containing variable concentrations of water and volcanic debris that can cause catastrophic impacts to life, livelihoods and infrastructure downstream from their volcanic origin. Accurate and quantitative information on lahar hazards are essential for reducing the impact of these events. Lahar hazard assessments often focus on the use of numeric or empirical models to describe flow behaviour and inundation areas, which rely on historic lahar events and expert elicitation to define model inputs. This results in qualitative or semi-quantitative estimates of hazard that do not account for the mechanics of lahar initiation or, in the case of rain-triggered lahars, the dependence of rainfall intensity and duration on initiation. Here we develop a method for calculating rain-triggered lahar susceptibility, defined as the occurrence probability of a particular lahar initial volume at a specific location. The model relies on terrain and deposit characteristics and a probabilistic measure of rainfall in the form of rainfall intensity-frequency-duration relationships. Results for a case study of the October 28, 1995 lahar at Mangatoetoenui stream, Ruapehu Volcano, New Zealand, indicate lahar volume is controlled by a characteristic timescale, relating the deposit depth H to the hydraulic diffusivity D0 in the ratio H2/D0. The timescale describes the transmission of positive pore pressures within the deposit, leading to shallow failure. As a consequence of this timescale, rainfall duration is the most important factor determining initial lahar sediment volume. Rainfall intensity plays a minor role, controlling the volume of water in the lahar mixture. This observation is consistent with power-law relationships used to determine lahar triggering rainfall thresholds. The rain-triggered lahar susceptibility approach developed here is anticipated to improve probabilistic lahar hazard assessments by providing quantitative, reproducible estimates of initial lahar volumes.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have