Precipitation observation in mountainous areas is quite important, not only for monitoring long-term effects on water resources or ecosystem changes but for short-term impacts of disasters and weather on local economies, recreation, and tourism. Especially in Monsoon Asia, increases in extreme weather events are a concern due to global warming. In Japan, mountain ranges sometimes cause heavy precipitation in the upwind areas; they also block the disturbances and protect lee side areas against heavy precipitation. However, the stagnation of the Baiyu front in 2018 caused extreme precipitation in Hiroshima, leaving 200 dead and 7000 people evacuated, despite recognition that the area, surrounded by islands, has a climate with less precipitation. Also, recent typhoons sometimes hit Hokkaido, the second largest island at the north end of Japan, causing unexperienced disasters. Given the progress of domestic depopulation and aging with globalization from abroad in the mountain society, we must share knowledge and precautions about how to cope with extreme weather events.

For short-term precipitation forecasts, observation by weather radar with a gauge network is fundamental. However, gauges are scarce and sometimes shaded by mountain ranges. Especially, continental mountainous areas are remote from cities, and gauges of the surface observation network are scarce. As a part of the Global Precipitation Mission (GPM), JAXA/EORC started the real-time estimation of the distribution of hourly global precipitation using a combination of multiple satellite-based microwave radiometer data with gestationary Infrared (IR) information, called GSMaP (https://sharaku.eorc.jaxa.jp/GSMaP/index.htm) . Experiments performed to validate data and assess data utilization for the local society are needed in mountainous/remote areas. Collaboration that share knowledge of weather change dynamics with the local community is quite important for mitigating the hazard and preparing for future climate impacts.

Seventy percent of Japanese islands are mountains with forests; even the weather variability of the coastal big cities is indirectly controlled by the thermodynamic effects of the surrounding mountains. According to global climate changes, natural hazards may be enhanced in the intermediary areas between plains and mountains due to extreme rains associated with typhoons or Baiyu fronts. Modulation of the activity of extra-tropical cyclones in winter may increase the frequencies of heavy snows or rain on snow in the mountains. This poster presentation introduces recent weather extremes in Japan and how satellite data could capture/estimate the precipitation-related phenomena over the mountains. Especially, the validation experiments of GSMaP with JALPS mountain weather data will be demonstrated with the causes of discrepancies between gauge and satellite estimates. Also, we would like to discuss how local societies or visitors from abroad can cope with digital weather/climate data from the mountains to mitigate hazard and develop ecotourisms.

The ability to observe the Earth from space at relevant spatial and temporal scales is key to understanding changes and related responses of species and water systems to climate change. The recent perfusion of commercial earth imagery with high spatiotemporal resolution may be able to bridge the gap between ground-based instrumentation and coarsely-captured satellite data. In particular, observations of snow-covered area in montane areas at high resolution (meter scale) are of critical interest as snow drives much of the seasonal hydrological regimes and can have significant ecological impacts on phytocenosis. Currently, remotely-sensed snow cover observations with adequate temporal resolution (daily) are either captured at a spatial scale far too large (e.g. MODIS, 500m), or are appropriate in spatial scale (1-10 m) but have inadequate temporal resolution (e.g. LIDAR). Planet Labs, Inc. (Planet) is a promising source of high-resolution Cubesat imagery that can be used in environmental science studies, as it has both high spatial (0.7-3.0 m) and temporal (1-2 day) resolution. Planet imagery is acquired by a constellation of nanosatellites collecting data mostly in visible and by some at near-infrared bands. However, its immediate utility with respect to inferring snow cover is limited due to one near infrared band which makes distinguishing snow from clouds difficult using a radiometric index (such as the NDSI), and therefore requires an alternative approach. Here we employ a machine learning (ML) classification algorithm to discriminate between snow-containing pixels and snow-free pixels. A cloud-based approach is used to process a large subset of Planet imagery and training and executing the ML model requires large computing power. We apply our procedure in the Tuolumne River basin, USA, where 3-m resolution airborne LiDAR-derived datasets are available as part of the NASA Airborne Snow Observatory (ASO) program, which are used for training and validation of the ML algorithm.

Mountain forests offer effective, natural and cost-efficient protection against avalanches. Trees reduce the probability of the avalanche formation and may also decelerate small to medium sized avalanches through snow detrainment. Remote sensing data are promising tools for an efficient assessment of forest structural parameters on large scales. The aims of this study were: (i) to test relevant forest parameters obtained from remote sensing methods; and (ii) to evaluate effects of forest parameters and forest cover changes on avalanche runout. We compared control assessment of maximum tree height, crown coverage and surface roughness with a DTM in combination with a photogrammetry-based vegetation height model (VHM_P) and with a LiDAR-based vegetation height model (VHM_L). We then simulated two avalanche case studies near Davos (Switzerland) with forest parameters estimated by the remote sensing and control data. The RAMMS simulation outputs as runout distance were compared. Tree height and crown coverage as assessed with both remote sensing methods were not significantly different from the control method. However, surface roughness was underestimated using the DTM compared to the control classification. For the wet-snow avalanche Teufi, runout distances of simulated avalanche did not differ significantly, but runout was increased for an avalanche scenario with less forest cover in the release area and/or less forest cover after forest destruction by a preceding avalanche event. For the dry-snow avalanche Schatzalp, the forest cover was underestimated by the VHM_P, which led to longer runout distance. Our study indicates that available remote sensing methods are increasingly suitable for the determination of forest parameters which are relevant for avalanche simulation models, but that more research is needed on the precise estimation of forest cover in release areas and on effects of forest cover changes on avalanche runout.

The analysis of river systems at appropriate spatial and temporal scales is essential to support a sustainable river management and to develop solutions to mitigate natural hazard impacts. Recently, new perspectives have been opened up for river monitoring thanks to emerging remote sensing technologies which are providing an unparalleled amount of data at spatial and temporal resolution not available in the past. Our research aims to investigate the potential to integrate the Sentinels satellite data with Unmanned Aerial Systems (UAS) derived river datasets to monitor river forms and processes consistently at large scale. Specifically, the project focuses on five hydromorphological indicators: mapping of in channel geomorphic units, sediment grain size, sediment budgets (through DoD analysis from Drone), and deriving proxy of discharge from water channel mapping. UAS datasets create the ground-truth whereas multispectral information from Sentinel 2 and SAR data from Sentinel 1 are explored to assess what can be observed and with which accuracy from space. Drone and satellite data were collected once a year for two years, on eight sites selected along Italian rivers (all with a channel width > 20 m) from north to the south. We present results obtained so far to map river geomorphic units and sediment grain size from UAS and how similar parameters can be observed from Sentinel 2 datasets. Thanks to their characteristics, the fusion of Sentinel 2 and UAS river data opens to a new generation of cross-scale hydromorphological indicators, where the ability to explore historical trajectories of channel processes is paving the way for a more comprehensive and consistent characterization and monitoring of river systems. Such pan-scale river observatory is suitable also for mountain area where probably UAS derived indicators will be the main source of information, whereas satellite data will be challenged due to their limited spatial resolution.

There is currently a growing concern in the mountain communties about changing glaciers, and particularly growing lakes and the probability of GLOF events. We currently lack standardized , integrated remote sensing and field-based tools for assessing whether these lakes are of concern, Furthermore, local knowledge is not often integrated in scientific assessments.

Current glacier hazard ranking schemes need to be updated with new data which are more readily available, and better communication is needed among scientists. It is key we develop tools that are easily transferred to local communities to assess the risks from changing glaciers, and the Mountain hazards session is an opportunity to bring together scientists to address these issues.

In this session I will present current efforts to develop a standardized scheme within the context of IGCP project 672 "Himalayan glaciers and risks to local communities'. The emphasis is on developing open source tools that are disseminated to local communities through a series of capacity building workshops.

Deep-seated gravitational slope deformations (DSGSDs) are abundant geomorphological features on Alpine slopes. Monitoring their activity is an important task in order to deepen the understanding of the main drivers, their spatial behavior and to prevent potentially devastating impacts. Remote sensing techniques and multi-temporal laser scanning in particular have become cost-efficient methods for area-wide monitoring of DSGSDs. In the present study 3D point clouds acquired with various remote sensing techniques were used for reproducing past and monitoring current changes of the Reissenschuh landslide in the Schmirn valley (Tyrol, Austria). Data acquisition techniques include airborne laser scanning (ALS), terrestrial laser scanning (TLS), unmanned aerial laser scanning (ULS) and photogrammetric techniques based on historic and recent aerial imagery. The multi-temporal 3D data spanning the period from 2008 to 2018 allow to assess the landslide’s spatio-temporal behavior. Extracted 3D displacement vectors for the period from 2016 to 2018 were validated based on periodical measurements using a differential global navigation satellite system (DGNSS) and are well in agreement. The magnitude of 3D displacement vectors varies spatially within one observation period in the order of 10 cm to 110 cm per year, but also temporally between different observation periods indicating zones of landslide acceleration and deceleration in the order of ±15 cm per year. The results allow to conclude on the various used techniques’ reliability for monitoring DSGSDs displacement.

Remotely Piloted Aerial Systems (RPAS) are becoming a standard tool for the Geosciences. With the increasing availability of diverse platforms, from fixed wing to multirotors, RPAS equipped with sensors are able to deliver near-real time data at very high spatial resolution for a number of applications including precision agriculture, land cover classification, and topographic changes, among others. Mountain environments represent a particular challenge for obtaining data from RPAS. Harsh atmospheric conditions and remoteness can limit the time for flying and surveying, requiring meticulous planning as well as adaptability during fieldwork. Here we present several RPAS applications currently in development jointly by the Mountain Geoscience Group and the GOAIR (Grupo de Operaciones, Aplicaciones e Investigación en RPAS) initiative at Universidad de Concepción, Chile. Our RPAS include multirotors, fixed wing, and hybrid (vertical take-off landing or VTOL) platforms developed and assembled in the Universidad de Concepción, and equipped with photogrammetric cameras, multispectral cameras, lidar, and temperature sensors. We show results of surveys over (a) glacierized Andean watersheds aimed to determine short term volumetric changes, aerodynamic roughness length, and air temperature profiles; and (b) forested regions in the Chilean Coastal Range to study flooding dynamics, land cover change and landslides. Among other aspects, we discuss lessons we have learned from fieldwork preparation and operation, data processing and analysis. We also propose opportunities for further development of RPAS for atmospheric data retrieval in mountain regions and integration with other observational platforms, such as satellite imagery.

The mountainous region affected by the 2008 Wenchuan earthquake was frequently analyzed to assess its vegetation recovery. However, the dynamic of surface thermal environment was rarely investigated but directly influenced local geophysical and biophysical processes. Under this background, this research aimed to analyze the trend of surface temperature based on Moderate Resolution Imaging Spectroradiometer (MODIS) daily land surface temperature (LST) product from 2000 to 2017 and understand its change mechanism. An annual temperature cycle (ATC) model was applied to obtain the daytime and nighttime annual cycle parameters (ACPs), including mean annual surface temperature (MAST), yearly amplitude of surface temperature (YAST), and phase shift. The trend analysis with the Mann-Kendall (MK) test illustrated the strong inverse relationship between the two ACPs (MAST and YAST) and elevation, and the nighttime terms showed a stronger connection than those of the daytime. The areas with significant decreases in vegetation coverage induced by the earthquake exhibited an obvious upward trend in daytime MAST and annual maximum temperature, while the vegetation improved areas presented a cooling effect. Comparatively, the nighttime ACPs was little sensitive to vegetation change and the change magnitude was relatively small. For the high mountain area, with the shrinkage of snow cover duration, the global warming effect resulted in a warming trend by increasing the annual minimum temperature at both the daytime and nighttime. Overall, this study provided insights to the impact of the earthquake to mountain thermal environment and its different responses to changes in vegetation cover and climatic environment.

Rock slope destabilization due to warming or thawing permafrost poses a risk to the safety of local communi-ties and infrastructure in populated mountain regions. The analysis of fracture kinematics in the context of localtemperature evolution in the longer-term is a common approach aiming to identify its forcing (e.g. Wegmann andGudmundsson, 1999, Matsuoka and Murton, 2008, Blikra and Christiansen, 2014). Hasler et al. (2012) and We-ber et al. (2017) analyzed fracture dilatation data measured at Matterhorn Hörnligrat at 3500 m a.s.l. and suggestthawing related processes, such as meltwater percolation into fractures to cause irreversible displacement. How-ever, this finding so far has not been backed up by data from different instruments or analysis methods. Hence,misinterpretation of the existing data can not reliably be excluded. Based on further data consisting of surfacedisplacements measured with D-GPS, inclinometers, ambient seismic vibrations and ground resistivity capturedand compiled over a period of ten years, we apply a multi-data cross validation technique to detect and quantifytemperature-induced rock slope dynamics and identify the components of derived process knowledge that predictbehaviour across differing observation methods. The combined analysis of this multi-modal dataset allows to fur-ther develop and analyse our limited understanding of the dominant processes governing rock slope stability, inour case a steep bedrock mountain permafrost buttress.

Based on this evidence we conclude that the kinematics observed at the surface in the winter/re-freezing period isnegligible compared to those observed during spring initiated by the thawing and mobilization of fluid water w.r.t.to destabilization and precursory signs of rockfall at a larger scale. Therefore, future research should focus on thequantification of water supply, distribution and mobility both in the frozen and fluid state.