Hidetsugu Yoshida

Decrease of size of hummocks with downstream distance in the rockslide-debris avalanche deposit at Iriga volcano, Philippines: similarities with Japanese avalanches

A morphometric investigation of the longitudinal distribution of hummocks at the southeastern foot of Iriga volcano in the Philippines showed that hummock size decreases away from the volcano. Aerial photographs and GIS analysis revealed that the size–distance relationship can be expressed as the exponential function A = α exp (−β D), where A is the area of a hummock and D is its distance from the source. This relationship is the same as that observed previously for freely spreading debris avalanches in Japan, including two avalanches at Bandai volcano. This size–distance relationship provides information about the physical characteristics of the event: the α value shows a strong correlation with the volume of the collapsed mass of the volcanic edifice, and the β value shows a strong correlation with the coefficient of friction of the debris avalanche. Thus, morphometric analysis of hummocks created by a volcanic avalanche illuminates both the physical properties of the volcanic body and the mobility of the avalanche. For the Iriga debris avalanche, the observed longitudinal hummock distribution is clearly a function of the volume of the collapsed mass and the coefficient of friction of the avalanche. The relationships so defined appear to be a geometric effect related to the areal extent of freely spreading hummocky avalanche deposits, especially their longitudinal dimensions.

Geomorphic reconstruction of formation and recession processes of waterfalls of the Kaminokawa river basin on Osumi Peninsula, southern Kyushu, Japan

To assess the geomorphological importance of waterfall recession in volcanic bedrock, we examined recession rates of six waterfalls in the lower reaches of Kaminokawa river basin on the Osumi Peninsula in southern Kyushu. The examination was performed with an empirical equation that uses a dimensionless parameter obtained by dimensional analysis of relevant measured factors, including erosive force, size of waterfall, and bedrock resistance. Welded Ata ignimbrite, formed at approximately 110 ka, may have played an initiating role to maintain such waterfalls because it resists weathering more than other local rocks. Estimated recession rates for the six waterfalls range from 0.2 to 3.0 cm/y, which compare with estimated rates for waterfalls in another region characterised by welded ignimbrite. Comparison of equation-derived recession rates of waterfalls with actual recession distances from confluences supports the idea that an original waterfall will subsequently split into two distinct waterfalls when it recesses past an upstream junction of two channels. Our findings revealed that all six waterfalls likely would have been at almost the same point lower in the watershed in the past, marking the general site of the original waterfall. Moreover, the ancestral original waterfall is highly likely to have started from a point between the caldera rim and present river mouth. There, a knickpoint was likely caused by the river dropping into an inner part of the caldera, possibly just after the eruption of the Ata ignimbrite. The waterfall erodes upstream away from the caldera basin, and this happens to be to the east.