Yohei Arata

Mountainous rural regions are facing population decline and aging water infrastructure, increasing the need for decentralized water supply systems. Implementing these systems requires identifying watersheds suitable as water sources and their spatial distribution. In snowfall-dominated regions, where snowmelt runoff constitutes a major seasonal water resource, understanding the controls on snowmelt runoff is essential for identifying such watersheds. Although geological setting is often a primary control on runoff and therefore a key factor in identifying such watersheds, its influence on snowmelt runoff and its interactions with climatic conditions remain poorly understood. We evaluated the controls of climate, watershed geology, and their interactions on quick and delayed discharge during the snowmelt period across 33 forested watersheds (2.1–21.9 km2) in the Sorachi River basin, Hokkaido, northern Japan, using 153 watershed-year observations. Linear mixed models revealed that watershed geology exerted the strongest control on delayed discharge. In particular, delayed discharge increased with volcanic rock proportion, which also interacted significantly with snowfall water equivalent, indicating stronger geological effects under high-snowfall conditions. In contrast, quick discharge was controlled primarily by climatic conditions. Although volcanic rock proportion proved to be an effective indicator for identifying suitable source watersheds, watersheds with high volcanic rock proportions were concentrated in the eastern basin. In contrast, high-relief watersheds, the second most important control on delayed discharge, occurred mainly in the central and western parts of the basin, where non-volcanic geology predominates. Considering these complementary spatial patterns provides an important framework for decentralized water supply systems in mountainous rural regions.

Sheet erosion is an important process in soil degradation and surface soil loss, but quantification remains difficult because soil is removed in a thin layer. In this study, we tested a method for tracking sheet erosion at high temporal resolution by combining Structure-from-Motion (SfM) with time-lapse cameras capable of nighttime imaging and a colored surface-soil tracer. A study plot measuring 7.5 m in width and 10 m in length was established, and white tracer lines were applied. Images were acquired at 30-min intervals from 24 August to 6 November 2025. Daytime and nighttime orthomosaics were generated using a UAV-derived digital elevation model (DEM) and camera position and orientation data. A sheet erosion index, E, was derived from temporal changes in normalized tracer brightness. Nighttime images acquired in infrared mode allowed identification of the white tracer lines and tracking of their temporal changes even with few tie points. During non-rainfall periods, the standard deviation of normalized tracer brightness showed positive relationships with solar radiation and negative relationships with relative humidity during daytime, whereas no clear relationships with meteorological variables were observed at night. This indicates that nighttime observations under infrared imaging were less affected by meteorological conditions and allowed more stable tracer detection. In addition, E showed strong relationships with rainfall duration and total rainfall, but weak relationships with rainfall intensity and microtopographic conditions. Although the proposed method does not directly quantify absolute erosion, it has potential for relative assessment of sheet erosion dynamics, especially under stable nighttime conditions.

Vegetation and subsequent ecosystem services can recover over time in forest headwaters devastated by massive disasters. However, in cold regions, their recovery rates are typically slow and often imperceptible, which makes it difficult to evaluate how much ecosystem services have recovered. This study targeted dissolved organic matter (DOM), which plays a central role in biogeochemical processes in forest ecosystems, and aimed to examine whether vegetation conditions affect the quality of stream DOM from cool-temperate forest headwaters in northern Japan. To achieve this, hydrological observations and stream water sampling were conducted monthly from May to December 2021 in three small forest catchments with different landslide coverage. Dissolved organic carbon (DOC) concentration in stream water was measured, and the molecular composition of DOM was analyzed using ultrahigh-resolution mass spectrometry and compared among the three catchments. The peak-intensity-weighted average aromaticity index (AI_wa) increased with DOC concentration. We found that AI_wa was the highest in the undisturbed catchment, followed by the catchments with landslide coverages of 16% and 52% at a given DOC level. These results indicate that the quality of DOM could dramatically change depending not only on DOC concentration but also on vegetation disturbance in cool-temperate forest headwaters.

The urgency to conserve and restore forests for their multifaceted benefits is escalating. We spotlight Japan's new Forest Environment Tax, a novel fiscal measure crafted to finance public‐beneficial ecosystem services through enhanced forest management. To convey the expert perceptions of the policy, we present the results of a survey targeting individuals immersed in Japan's forest policies, which aimed to assess attitudes toward the various benefits that forests provide. We classified forest functions into five core areas: wood production, soil and water conservation, mitigation of anthropogenic global warming, wildlife conservation and cultural utilities. We found that stakeholders closely involved in forest policies in Japan prioritize soil and water conservation as the paramount function over the mitigation of anthropogenic global warming. The results of the survey underscore the necessity of evaluating forest management practices and the importance of recognizing the multiple values that can be derived from forests. While there has been much attention to the carbon benefits of forests in the region and beyond, we emphasize the need to avoid an excessive focus on this single ecosystem service and to ensure that the other important multifunctional values of forests are not overlooked. Policy implications: We call for a more holistic approach that recognises the interdependence of the different functions of forests and the importance of valuing forests as natural capital in all their dimensions.

Forest vegetation and soils in headwaters can control runoff and surface erosion. However, it remains unclear how vegetation affects nutrient exports from cool-temperate forest headwaters during intense rain events that transport sediment-associated nutrients, such as phosphorus (P). To clarify this, we targeted an upstream landslide area and analyzed P contents in surface soils and total P (TP) in stream water of the undisturbed (UF) and landslide-bearing forest (LB) catchments. The soil P content was higher in the UF catchment than in the LB catchment, but differences in the average TP concentration and load during low flows between these catchments were not significant. Conversely, the overall runoff and the TP load were three and ten times higher in the LB catchment than in the UF catchment, respectively, during a rain event with daily precipitation of 49 mm, despite the soil P content being much lower in the LB catchment. Particulate P (PP) accounted for more than 90% of the TP load during the rain event in the LB catchment, whereas dissolved P accounted for more than 80% of the TP load in the UF catchment. Therefore, soil surface mobility strongly affected P transport in the forest catchments. Our study suggests that vegetation not only reduces PP loads by controlling runoff, but also influences stream P forms in cool-temperate forests.

We examined soil‐water response in a hillslope with multi‐layered volcanic soils affected by fissures formed by intense shaking during the 2016 Kumamoto earthquake in Japan. Pressure head and volumetric water content responses in a 6 × 20 m hillslope ridgeline plot were monitored by tensiometers and capacitance‐based soil moisture sensors, respectively. The plot contained seven seismic fissures (0.3–0.8 m deep and 0.4–2.2 m wide) formed parallel to the ridgeline that exposed underlying soil layers, while areas with topsoil occurred between the fissures. Among the alternating layers, the andisol (ca. 1.0 m deep) consistently experienced high pressure heads (>−150 cm H₂O) and volumetric water contents (>0.58 cm³ cm⁻³), indicating high water‐holding capacity. This andisol was a key layer for storing soil water and causing slow drainage into the deeper matrix, resulting in abrupt increases in volumetric water content during storms. Rainwater directly reaching the bottom of fissures without percolating through the topsoil caused rapid soil‐water response compared to the soil layer where water percolated via topsoil. Analysis of in‐situ soil water retention curves, recession rates and water storage changes revealed that the second andisol had unique characteristics for retaining water in the hillslope. The findings of this study show that spatial and temporal variability of soil water responses varied associated with earthquake‐induced fissure formation and soil water characteristics of multi‐layered volcanic soils.

Individually delineated landslide inventories are essential in analyzing post-earthquake-induced landslides (EIL) hazard assessments, particularly examining statistical correlations between landslides (e.g., frequency and size) and physical parameters. Despite rapid advances in remote sensing technology, previous recorded EIL inventories still have limitations in carrying out fine quality inventories, mainly due to limitations in delineating individual landslides manually over large areas by low-resolution satellite images. To be specific, fine quality inventory requires the ability to detect landslide scars and deposits separately over whole affected areas, recognizing smaller landslide sizes (<103 m2) under canopies, as well as avoiding amalgamations, i.e., a combination of several individual landslides in a single polygon, which can lead to severe distortion of landslide statistics. The latest technology from LiDAR-Digital Terrain Model (DTM) allows geomorphologists to manually delineate landslides precisely, but most studies had only focused on deep-seated landslides. Thus, the main objective of this study was to delineate the recent EIL based on LiDAR-DTM visualization over whole landslide-affected areas and test preliminary statistics between our manual LiDAR-based inventory (MLI) with automatic aerial-based inventory (AAI) in the same areas, in addition to NASA’s global EIL database.We manually delineated the recent landslides affected by the 2018 Eastern Iburi earthquake in the Atsuma basin in Hokkaido within an area of 266 km2, accounting for about 90% of the total area affected by landslides. Shaded relief derived from LiDAR-DTM (0.5 m), and aerial photograph (0.2 m) were used to identify landslide morphometrics. AAI collected in the same study area (Kita, 2018) was used to compare with MLI. As a result, our MLI was able to detect a total of 17,160 landslides (total landslide area: 27.5 km2) while the automatic AAI was only 4241 landslides (total landslide area: 33 km2), probably because our MLI was able to recognize more small landslides and separate individual landslides from amalgams. The mean landslide density for MLI is four times greater (64 landslides/km2) compared to AAI (16 landslides/km2), also considered the densest landslide inventory recorded so far in 20 years based on NASA's global EIL inventory database. Based on the binned frequency area distribution (FAD), MLI has a power-law exponent (β) of 3.4 and a rollover point of 800 m2, whereas AAI is 2.7 and 3×103 m2, respectively, probably because AAI's inventory overestimates its delineation by inserting channels and depositional regions in the delineated polygons. Compared with all global EIL inventories (mean β: 2.4), the value of the MLI was found to be larger, indicating that the Iburi EIL is the smallest size EIL so far in history (50% landslides are smaller than 103 m2), but very dense. Our findings suggest that MLI might reveal hidden unexpected statistics of the number and size of EILs, including exposing smaller landslides under the canopy and splitting amalgams.

The spatial variability of landslides and associated sediment deposits induced by earthquakes alters both short- and long-term sediment dynamics in watersheds. Linkages between landslide occurrence and sediment accumulations within channels are important for evaluating spatial and temporal dynamics of sediment from headwaters to downstream. To evaluate spatial variability of landslides, we examined landslide-area density (LAD: landslide area divided by watershed area) in different sub-watersheds (areas 0.01 to 4.4 km2) of Habiugawa watershed (40 km2), which was affected by the 2018 Hokkaido Eastern Iburi earthquake, Japan. The watershed is located 13 km north of the epicenter and is covered by secondary conifer and deciduous forest. The topography is hilly associated with long-term landform development by paleo-glacial erosion. Altitude ranges from 30 to 440 m; mean hillslope and channel gradients are 30° and 10°, respectively. Landslides mostly occurred at depths from 1 to 2 m below pumice layers formed by the Mt. Tarumae eruption 9000 yr ago (Ta-d), with total soil depths from 2 to 3 m. The 0.5 m LiDAR-based DEM and 0.2 m post-earthquake orthophotos were used to calculate LAD by GIS analysis. To examine spatial variability of in-channel sediment deposited by landslides, we used deposit-length ratio (DLR: total length of sediment accumulations within channels divided by total channel length within sub-watersheds). Sediment deposition in channels was assessed as rough surface topography by DEM and orthophotos.We identified 2941 landslides: mean area=1620 m2; range from 20 to 34710 m2. LAD in the entire Habiugawa watershed was 0.12 km2 km-2, which is high compared to the other earthquake-induced landslides (e.g., Wenchuan earthquake: 0.03 km2 km-2). Sub-watersheds < 0.1 km2 had wide ranges in LADs (0.0 to 0.8 km2 km-2), while sub-watersheds from 0.1 to 0.5 km2 ranged from 0.2 to 0.5. Sub-watersheds > 0.5 km2 had LADs from 0.1 to 0.3. Seventy-four percent of small watersheds (< 0.5 km2) with high LADs (> 0.3) also had high sediment accumulations within gentle channels (DLR ≥ 0.8). This suggests that poorly mobilized sediments that initiate in headwaters rapidly deposit in channels. Conversely, the other small watersheds (26%) had lower sediment accumulation within steeper channels (DLR < 0.8), suggesting that these high-mobilized sediments traveled longer and were evacuated from watersheds to some extent. Such differences in sediment mobility in small sub-watersheds (< 0.5 km2) may cause sporadic sediment accumulations within channels of larger watersheds (> 0.5 km2). Our findings suggest that geomorphic features of watersheds associated with long-term legacies of geomorphic evolution possibly affect the spatial variability of landslide occurrence and the associated in-channel sediment accumulation induced by the earthquake.

Vegetation is one of key factors controlling landslide occurrence, including frequency, size, and depth. Both horizontal and vertical root networks have important roles in stabilizing hillslopes. For instance, landslide density can be moderated by dense and deep tree root reinforcement below the potential soil slip surfaces. Landslide size can be reduced by extended dense and thick tree root networks, providing cohesive and lateral hillslope reinforcement. Vegetation conditions such as density and composition also alter the landslide occurrence because they are linked to root network density and strength, which are affected by different biogeoclimatic conditions. These findings regarding landslide-vegetation interactions were mostly based on rainfall-induced landslide cases. Preliminary, but yet to be confirmed, findings in Eastern Iburi Earthquake-Induced Landslides (EIL) showed that lateral root reinforcement might moderate the size of landslide scars in forested areas compared to logged areas.  Therefore, our primary objective was to examine the effect of different vegetation composition on EIL based on global data and supplemental analysis.Our global database of EIL was compiled for a 20-yr period using a literature review and GIS analysis. Documented landslides were restricted to shallow mass movements with depths approximately less than 3 m. For vegetation-related analysis, we used Net Primary Production (NPP) and Leaf Area Index (LAI) derived from MODIS-Terra satellite images. Twenty-seven EIL cases were recorded in our database occurring from 2002 to 2018. Among these, 26% of the total cases occurred in Japan, followed by 18% for both in China and New Zealand. Based on climate types, 22% of total EIL cases occurred in temperate oceanic climate (Cfb) dominated by New Zealand EIL cases, and 15% cases occurred in humid subtropical climate region (Cfa), such as Japan. Moreover, 7% cases occurred in tropical rainforests (Af) and 7% cases in hot desserts climate regions (BWh). Among the 27 recorded cases of EIL, we selected eight EIL cases based on biomass classes, which are low (0-2 gC/m2/day), moderate (3-5 gC/m2/day), and high (>5 gC/m2/day). A power-law cumulative-area distribution of landslide areas showed that low biomass sites had the largest landslides (11,000 m2), followed by moderate biomass (3000 m2), and high biomass (200-3000 m2) with the smallest landslides, possibly associated with the density of vegetation. In low biomass regions, the average LAI was 1.8 m2/m2, which was three times lower compared to regions with higher biomass. This indicates that in regions with sparse vegetation, slope reinforcement by dense lateral root networks was minimal. Future research is focusing on compiling information on landslide scars and root depth to assess the effects of vegetation density and vertical root reinforcement on landslide characteristics in each biomass class.