Decoupling of small-volume pyroclastic flows and related hazards at Merapi volcano, Indonesia

The November 1994 eruption at Merapi volcano provided good evidence of decoupling of dome-collapse pyroclastic flows and of large-scale detachment of an ash-cloud surge (ACS) component from the basal block-and-ash flow (BAF). Timing and stratigraphic relationships of the largest 1994 ACS indicate that this escaped from the valleys, travelled well ahead of the BAF, arrived at the termination tens of seconds before it and deposited a discrete ACS deposit beneath the BAF unit. This suggests that the ACS detachment mostly occurred relatively high on the volcano slope, likely at the foot of the proximal cone. Later pyroclastic flow eruptions in January 1997 and July 1998 also showed evidence of ACS detachment, although to a lesser extent, suggesting that ACSs could be a frequent hazard at Merapi volcano. Based on an extensive review of the available literature and on field investigations of historical deposits, we show here that flow decoupling and ACS detachment in the way inferred from the 1994 eruption is a common process at Merapi. The ACS-related destructions outside valleys were frequently reported in the recent past activity of the volcano, i.e. in at least 16 pyroclastic flow eruptions since 1927. Destruction occurred systematically in eruptions where maximum runout of the BAFs was 6.5 km or more, and occurred rarely for BAF runouts of 4.5 km or less. The ACS deposits have been recognized beneath some valley-filling BAF units we attribute to some recent destructive eruptions, i.e. the 1930, 1954, 1961 and 1969 eruptions. Topographic conditions at Merapi volcano favouring ACS detachment include: (a) the high slope (30 degrees) of the proximal cone, leading to high proximal velocities of the pyroclastic flows and thus to the transfer of large amounts of particles into the ash cloud; (b) the strong break in slope at the foot of the proximal cone, where the velocity of the basal BAF is strongly reduced and a major ACS component is thought to form and detach by shearing over the BAF; and (c) the small depth of most valleys in the first kilometres beyond the foot of the cone, which allows minor ACS components to escape from the valleys during travel of the BAF; however, flow decoupling and ACS detachment occur for only some of the numerous pyroclastic flows that follow the same path in a given eruption. This indicates that topography alone cannot lead to flow decoupling. We suggest two factors that control flow decoupling and its extent. The main one is flow volume (and thus flux, as both are correlated in almost instantaneous, dome-collapse events), as suggested by the observed relationship between flow decoupling and the travel distance of the pyroclastic flows. The second factor is the amount of available ash in the flow at its early stage, which influences the mass and thus momentum of the ash cloud. The amount of ash in the pyroclastic flows of Merapi may depend on several factors, among which are (a) the physical and thermal state of the part of the active dome that collapses, and (b) the proportion of older, cold rocks incorporated in the flow, either by undermining of surrounding summit rocks by the current pyroclastic flow activity or by erosion on the upper slopes.