⼭体不安定によるもの︖ もしこの部分(~0.5 km3)が将来⼭体崩壊を起こすとし たら(⼤きな仮定︕),⼭体崩壊体積は最⼤で~107 m3. (24 ka⿊斑︓ 4x109 m3, ⻤押溶岩流︓ 1.7x108 m3 ) gure 2b], which show that t‐As scaling may t / As m, with m = 0.31 ± 0.19 (i.e., there is ield data, yet on average, large slides are er). Predictions for t versus As scaling may om our model assuming model derived t (equation (7)) and As / w · l. However, as ction 3.2, slide width, w, is not well con- chanical considerations (in theory it could laterally infinite slope). Two end‐members or slide widths: a lower bound arises from uggesting w / t, so that As / t · l, (this is trained end‐member) and an approximate r width suggests w / l, so that As / l2. (7), our model thus predicts scaling rang- As 0.39 (for w / t) and t / As 0.27 (for w / l), the scatter of t versus As observed by 2010] may be explained via the mechani- le range in landslide widths, w, which a range in mechanically permissible areas compared with the lower range of power exponents found from the analysis (i.e., d = 1.36) the values of the scaled cohesions c/g are an order of magnitude larger (between 0.005 and 1 m) leading to cohesion ranging between 10−1 and 20 kPa. When comparing with the model results using the w / l scaling (Figure 15), and a power of 1.32, the cohesion is found to be in the range of 0.6–200 kPa. [32] The very large range of obtained permissible cohe- sions (more than 5 orders of magnitude, 10−3 to 200 kPa) versus the very narrow range of permissible power ex- ponents (d ranging by a few percent between 1.32 and 1.38) indicates that the value of the power exponent is not very sensitive to the specific cohesion value of the slope material, and also not very sensitive to the width of the slide, but is instead mostly sensitive to the aspect ratio t/l which really dictates the scaling between volume and surface area. This clearly supports, and partially explains, the observations that landslides from diverse conditions follow similar power laws. 地すべりの⾯積と体積のスケーリング Klair et al. (JGR, 2011)