Rainfall-triggered volcanic debris flows (VDFs) occur because of reduced infiltration capacity after a volcanic eruption and are the primary source of sediment export. Even after the infiltration capacity of volcanic slopes has recovered, VDFs can persist for several decades. However, no models have been developed to estimate long-term sediment export more than a decade after an eruption of any volcano. We developed generalized linear models to estimate sediment export (Se) over a long term in the Tansandani and Gokurakudani gullies at Mount Unzen, which erupted between 1990 and 1995. The variable Se, calculated using digital elevation models from 2003 to 2022, was set as the dependent variable. The rainfall index, decline indicator for sediment export, proxy of constant sediment supply, and gully difference were set as independent variables. In addition to 40 rainfall index patterns, we tried three decline indicators: the normalized difference vegetation index for the growing and non-growing seasons (Model 1a and 1b) and the cumulative dates (Model 2). Based on stepwise processes, the rainfall index and decline indicator were selected for all models. The gully difference was selected for Models 1b and 2. For all models, the estimated Se well followed the observed Se, although the estimation error was the smallest for Model 2, followed by Models 1b and 1a. Thus, we successfully estimated Se over two decades using easily obtainable variables. Further studies would be effective to confirm the broader applicability of our models.

The risk of glacial lake outburst floods (GLOFs) in high mountain regions such as the Himalayas is increasing due to the effects of global warming. In this study, we focused on Tsho Rolpa, the largest glacial lake in Nepal, and conducted on-site observations and numerical simulations based on physical processes to quantitatively assess how rising temperatures due to global warming are affecting the occurrence of GLOF. During on-site observation, moraine surface temperature was measured for 1 year from November 2018 to November 2019. In the numerical simulation, using the temperature changes obtained from on-site observations as input values, changes in the temperature distribution of the moraine from the past to the future were calculated, and the progression of the melting surface layer of the moraine was determined. Next, assuming the occurrence of GLOF due to spontaneous collapse, we estimated temporal changes in the risk of GLOF occurrence through slope stability analysis with the critical slip surface method, considering the melting surface layer of the moraine as a collapsible soil layer. The simulation results showed a significant decrease in the stability of moraines over several decades from the past to the future, which is consistent with the increasing trend of GLOF occurrence observed in recent years.