Displays

Land use is one of the most important factor which can promote or reduce the susceptibility of an area towards shallow slope instabilities. Different plant species guarantee different amounts of additional reinforcement to unstable soil covers, thank to the mecahanical effects of their roots as a function of their density and shear strength properties. Furthermore, land use changes and modifications of management practices in cultivated slopes could cause an increase in the proneness towards these phenomena, due to modification on vegetational types and on farming and tillage operations that could reduce the root additional reinforcement in soil. Hilly areas vocated to viticulture are one of the most affected landscapes that suffere of shallow slope instabilities as a consequence of modification in agricultural management and of land use changes for the abandonement of previously cultivated hillslopes. Therefore, this work aims to analyze the effects of the land use changes and of the different agronomical practices occurring in an area vocated to viticulture prone to shallow landslides triggering. From the point-of-view of land use changes, we analyzed especially the linkage between the location of past shallow landslides events and the possible temporal variations of land cover or of agricultural practices in still cultivated areas. For the effect of agricultural practices in vineyards, we quantified the root reinforcement and the probability of occurrence of shallow landslides on vineyards managed with traditional agricultural techniques of tillage and permanent grass cover as well as the alternation of these two practices between adjacent inter-rows. The research was conducted in several test-sites of the Oltrepò Pavese (Lombardy region, north-western Italy), one of the most important Italian zones for wine production in northern Italian Apennines. The results show that the test-site was characterised by pronounced land abandonment and important changes in agricultural practices. In particular, abandoned cultivated lands that gradually recovered through natural grasses, shrubs and woods were identified as the land use change classes that were most prone to shallow landslides. Regarding the features of the grapevine root system, vineyards with alternation management of inter-rows had the highest root density and the strongest root reinforcement, of up to 45% in comparison to permanent grass cover, and up to 67-73% in comparison to tilled vineyards. As a consequence, slopes with medium steepness (10-18°) were unstable if inter-rows of vineyards were tilled, while vineyards with permanent grass cover or alternation in the inter rows promoted the stability of slopes with higher steepness (>21-25° for vineyards with permanent grass cover in the inter rows, 28-33° for vineyards with alternation). The results of this study yielded important information to establish effective land use management practices able to reduce shallow slope instabilities. This work was supported by the project Oltrepò BioDiverso, funded by Fondazione Cariplo in the frame of AttivAree Program.

How to cite: Bordoni, M., Vercesi, A., Maerker, M., and Meisina, C.: Impact of land use changes and of agricultural management in vineyards to shallow landslides susceptibility in a representative area of northern Italian Apennines, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-12608, https://doi.org/10.5194/egusphere-egu2020-12608, 2020.

The intensification of agriculture over the last 50 years together with a constant change in climatic conditions has resulted not least in a deterioration of the aquatic habitat due to sediment input and siltation in the upper reaches of Bavarian streams. Concerned about this development, the  Fisheries Association Bavaria has launched a project to investigate the main causes of erosion on agricultural land.

By comparing aerial photographs from the 1960s with current orthophotos, by means of a detailed GIS-analysis, the size of agricultural plots in five representative catchment areas was first investigated. In a further step, erosion modelling based on the Universal Soil Loss Equation (USLE) was implemented in two catchment areas.

The intersection of the digitalized land uses from the two time steps showed that despite an almost constant proportion of arable land in the catchment area, the length of the fields had been increased by a third on average and their extent had at least doubled, due to wide-ranging changes in the landscape structure.

By considering the soil loss in the 1960s, that under today's conditions, and by modelling scenarios with conserving farming technics and further-reaching retention measures, conclusions can be drawn as to which measures will be necessary in the future to enable effective soil and water protection.

The erosion modelling showed that the average long-term soil loss - as a result of the USLE - currently exceeds a value of 40 t/ha*a under conventional farming in vast areas of the arable land. Likewise, even with conservation tillage (no-till), isolated erosion spots of more than 20 t/ha*a occur. Since a simple change to soil-conserving cultivation (reduction of the cultivation factor C) will not be sufficient to prevent future erosion events (increased precipitation erosivity R) and constant soil loss, targeted measures (improvement of the erosion protection factor P) against soil erosion must be implemented. This includes nature-based retention measures as wetlands, buffer strips or green waterways. All of which will also help to tackle the upcoming impacts of the Climate crisis. The chosen model supports the localization of the source of erosion as well as the selection and implementation of targeted measures.

How to cite: Ringler, G., Höfler, S., Reebs, F., and Gumpinger, C.: Targeted erosion control measures: The backbone of soil-conserving cultivation, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-9623, https://doi.org/10.5194/egusphere-egu2020-9623, 2020.

Soil erosion in Alpine grassland areas is an ecological threat caused by the extreme topography, prevailing climate conditions and land-use practices but enhanced by climate change (e.g., heavy precipitation events, changing snow dynamics) in combination with changing land-use practices (e.g, more intensely used pastures). To increase our understanding of ongoing soil erosion processes in Alpine grasslands, there is a need to acquire detailed information on spatial extension and temporal trends.

In the past, we have successfully applied a semi-automatic method using an object-based image analysis (OBIA) framework with high-resolution aerial images (0.25-0.5m) and a digital terrain model (2m) to map erosion features in the Central Swiss Alps (Urseren Valley, Canton Uri, Switzerland). Degraded sites are classified according to the major erosion process (shallow landslides; sites with reduced vegetation cover affected by sheet erosion) or triggering factors (trampling by livestock; management effects) (Zweifel et al. 2019). We now aim to apply a deep learning (DL) model with the purpose of fast and efficient spatial upscaling(e.g., alpine-wide analysis). While OBIA yields high quality results, there are multiple constraints, such as labor-intensive steps and the requirement of expert knowledge, which make the method unsuitable for larger scale applications. The results of OBIA are used as a training dataset for our DL model. The DL approach uses fully-convolutional networks with the U-Net architecture and is capable of rapid segmentation and classification to identify areas with reduced vegetation cover and bare soil sites.

Results for the Urseren Valley (Canton Uri, Switzerland) show an increase in total area affected by soil degradation of 156 ±18% during a 16-year observation period (2000-2016). A comparison of the two methods (OBIA and DL) shows that DL results for the Urseren Valley follow similar trends for the 16-year period and that the segmentations of eroded sites are in good agreement (IoU = 0.83). First transferability tests to other valleys not considered during training of the DL model are very promising, confirming that DL is a well-suited and efficient method for future projects to map and assess soil erosion processes in grassland areas at regional scales.

References

L. Zweifel, K. Meusburger, and C. Alewell. Spatio-temporal pattern of soil degradation in a Swiss Alpine grassland catchment. Remote Sensing of Environment, 235, 2019.

How to cite: Zweifel, L., Samarin, M., Meusburger, K., Roth, V., and Alewell, C.: Identification of Soil Erosion in Alpine Grasslands on High-Resolution Aerial Images: Switching from Object-based Image Analysis to Deep Learning?, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-2328, https://doi.org/10.5194/egusphere-egu2020-2328, 2020.

Soil erosion under forests occurs if forest layers get disturbed. Disturbances may arise from treefall, forest road works, skid trails or deforestation. In these disturbed areas, both an intact canopy and forest floor cover are missing, so that forest soils lack protection against water erosion. To counteract these negative effects a quick restoration of soil surface covers by vegetation is important. In particular, biological soil crusts (biocrusts) are able to quickly colonize gaps in higher vegetation and they are known to reduce soil erodibility. So far, the focus of biocrust research has been in drylands, whereas biocrusts have proven to be an important factor in mesic environments, especially as a pioneer vegetation in disturbed areas.

In this study, the natural succession of biocrusts in skid trails was observed on four different underlying substrates in a temperate European forest ecosystem (Schönbuch Nature Park in the state of Baden-Württemberg, Germany) and their influence on surface runoff, sediment discharge and nutrient relocation was investigated. Therefore, 144 micro-scale runoff plots (ROPs, 40 x 40 cm) were established with four replicates in the wheel tracks as well as in the center tracks and two replicates on undisturbed forest soil. In order to initiate splash and interrill erosion, four rainfall simulations were carried out from spring to winter with a constant intensity of 45 mm h^-1. With the purpose to compare these small-scale erosion rates with a larger scale, additional turbidity sensors were installed in the catchment area. The biocrust succession was determined by regular vegetation surveys with a classification of mainly mosses and liverworts up to the species level. Additionally, DNA samples of the upper soil layer were collected to conduct DNA extractions specify other potential biocrust organisms such as lichens, cyanobacteria, fungi and algae.

First results show that surface runoff and sediment discharge are higher in the wheel track than in the center track and that both parameters are reduced with a higher developmental stage of soil surface cover. The vegetation survey demonstrates a quick development of moss-dominated biocrusts from April to October with up to ten different species in one ROP. Depending on the location of the skid trail, a quick development of the higher vegetation was observed as well. Lab work on nutrient relocation and DNA analysis is still in progress and further results will be presented at the EGU 2020.

How to cite: Gall, C., Nebel, M., Quandt, D., Sauer, M., Scholten, T., and Seitz, S.: Soil Erosion in Mesic Forests: How do Biological Soil Crusts affect sediment transport and surface runoff?, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-664, https://doi.org/10.5194/egusphere-egu2020-664, 2020.

The practice of coppicing is debated in the literature for the risk factors associated with soil erosion. Although erosion experiments provide useful data for estimating the susceptibility to soil erosion, there are many open questions that cannot be solved in isolated experiments, but which can be assessed by activating a long-term monitoring process. In this way, it is possible to correctly frame the spatial and temporal scale of soil erosion in coppice forests. 

The aim of the work is to evaluate the effectiveness of the use of remote sensing data in combination with field data, for monitoring the evolution of forest stands interested by coppicing in relation to soil erosion. 

We have installed a long-term monitoring network for erosion estimation, while Sentinel-2C satellite data were used for the period 2016-2018. Starting from this dataset, a selection of vegetation indices was calculated and compared to the morphological and topographical parameters of the study area, as well as the above-ground data collected during field activities. Using the Canonical Correspondences Analysis (CCA) the relationships between the matrix of vegetation indices, topographic and vegetational parameters and the respective performances of this protocol have been explored in order to describe the evolution of the forest stands in the study area associated to soil losses.

How to cite: Giadrossich, F., Ganga, A., Campus, S., Murgia, I., Piredda, I., Lovreglio, R., Di Prima, S., Pirastru, M., and Scotti, R.: Remote sensing data and field survey activities for monitoring the evolution of forest systems after coppicing and soil erosion: A case study in Sardinia (Italy), EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-21354, https://doi.org/10.5194/egusphere-egu2020-21354, 2020.

Landslides represent a major threat to humans and result in high costs for the society. Landslide inventory maps depict the areas of past slope instabilities and are a valuable information source for authorities, spatial planners and risk managers. However, existing inventories are rarely complete, especially in sparsely populated and/or areas difficult to access. Previous work based on change detection and using approaches that automatically map distinct landslide events exploiting remote sensing data has shown promising results. The aim of this study was to test the applicability of multi-temporal change indices derived from Sentinel-2 (S2) for landslide detection for two landslide-prone study sites in Italy and China: South Tyrol and Longnan, respectively.

The methodical approach was built upon a change vector analysis applied to annual cloud-free S2-composites at 10m spatial resolution to extract land-cover disturbances. Landslide areas in the time period 2015-2019 were analyzed on the basis of already known landslide location points, downslope-oriented moving windows and supervised classifications using the Receiver Operating Characteristic (ROC) curve.  Subsequently, time-series analysis was applied to the detected landslide-affected areas and to derive temporal breakpoints (i.e. the timing of the landslide occurrence). Finally, applying a multi-temporal revegetation analysis, we accounted for false positives originating from agricultural activities or artefacts on single images. Our findings highlight that out of the 67 already known landslide locations in South Tyrol, only 9 (13.4%) were detectable by means of S2 data. Major challenges resulted from similar spectral characteristics of landslides and other land cover disturbances (especially tree logging). However, larger landslides were detectable both spatially and temporally by means of the multi-temporal change detection approach. By applying a quantitative accuracy assessment for the independent test site in Longnan, China, we are currently assessing the transferability and suitability of the developed approach for efficient spatial-temporal landslide mapping over large areas.

How to cite: Mayrhofer, P., Steger, S., Sonnenschein, R., Cuozzo, G., Atzberger, C., Schneiderbauer, S., Zebisch, M., and Notarnicola, C.: Forest change as a proxy for landslide occurrence - a Sentinel 2 based spatio-temporal landslide detection approach for two test sites, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-4743, https://doi.org/10.5194/egusphere-egu2020-4743, 2020.

Rainfall-induced soil slips are one of the most common and critical natural phenomena affecting the steep slopes in mountainous regions. These soil processes cause - directly and indirectly - huge damages to human-life, infrastructures and properties, especially when evolve into rapid soil movements such as debris avalanches, debris flows, flow slides, and rockslides. In this context, a landslide risk management that includes an accurate and robust real-time landslide early warning system at large scale (catchment or regional) for assessing the triggering soil slips both in space and in time, is necessary. This purpose appears more complicated where the forest covers most of the territory of a region and landslide-triggering thresholds cannot catch the exact process. In addition, most of physically-based models for real-time landslide warning neglect the role of vegetation, which is well-recognised to be fundamental in preventing shallow soil movements. In fact, forests influence hydrological and mechanical properties of the shallower soil layers through the beneficial effects of root systems and the canopy cover (reducing soil moisture, intercepting precipitation, reinforcing the soil resistance, etc.).

The present study proposes a modified version of the physically-based stability model, SLIP (Shallow Landslides Instability Prediction), based on the limit equilibrium method applied to an infinite slope and on a simplified modelling of the water down-flow. SLIP was integrated with a quantification of the rainfall interception by the forest canopy, and of the soil reinforcement provided by root systems as a function of tree species and tree density (which are data available from the forest management plans). The adapted model was applied to two mountainous catchments located in Valsassina (Northern Italy) and almost completely covered by forests (conifers, broadleaves and mixed). The study area was affected by shallow landslides and debris flows occurred after extreme meteorological events during autumn 2018. The model accuracy was tested through a back-analysis on the recent soil slips, mapped into a landslide inventory that was produced comparing high-resolution multi-temporal satellite images. The results provide an accurate risk map, identifying the areas of sediments source that can evolve into more threating soil movements.

The specific development of more accurate physically-based model can reasonably be an important tool for landslide risk management. Combined with a radar rainfall forecasting method, SLIP can be useful for addressing the real-time civil protection response to the emergencies. Moreover, the proposed method can play a key role in identifying the priorities inside the catchment management strategy, e.g. removing accumulated sediments in reservoirs, designing additional geotechnical or soil-bioengineering countermeasures, or evaluating the protection function of the forests.

How to cite: Gatto, M. P. A., Bischetti, G. B., Miodini, C., and Montrasio, L.: Towards an early-warning system for rainfall-induced landslides in forested catchments: a case study in Valsassina (Northern Italy), EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-13631, https://doi.org/10.5194/egusphere-egu2020-13631, 2020.

Both landscape rejuvenation through tectonic uplift and human-induced deforestation are known to increase landslide (LS) activity. Yet, the interaction between deforestation and landscape evolution has hitherto not been explicitly considered. Here, we investigate how shallow LS frequency is impacted by deforestation and landscape rejuvenation through knickpoint retreat in the Kivu Rift (East African Rift) while accounting for rock strength and slope steepness. In the past 12 Ma, the Kivu Rift has been characterized by tectonic uplift which gave rise to knickpoints in the river profiles enforcing topographic steepening. On a much shorter timescale, the rapidly growing population in the Rift has gradually expanded its cultivated and urban land leading to widespread deforestation.

We compiled an inventory of almost 8000 shallow LSs using Google Earth imagery. To quantify LS frequency, we developed a new method that accounts for the temporal and spatial inconsistency of satellite imagery coverage. To characterize long-term landscape evolution, we identified (i) 672 non-stationary knickpoints in the Rift and (ii) quantified the impact of lithology on slope threshold angles (TA). We identified two homogenous lithological groups: one group of younger/weaker lithologies (<540 Ma, TA=19.0 +/- 2.0°) and one group of older/stronger ones (>540 Ma, TA=27.9 +/- 0.3°). Further analysis focused on the latter group since it covers 85% of the study area and contained more than 95% of the observed LSs.

The overall shallow LS frequency in the rejuvenated landscapes inside the rift is 0.039 LS/km²/yr versus 0.010 LS/km²/yr in the relict landscapes outside the rift. Generally, LS frequency on recently deforested slopes increased by 200 to 800% in comparison to forested land. There is no notable difference in LS frequency on equally steep non-forested slopes (i.e. slopes deforested at least several decades ago) inside and outside of the rift. However, forest slopes of similar steepness are 2-3 times more sensitive to landsliding within the rift. We propose two mechanisms that might explain the higher frequency of landsliding on similar topographies within the rift: (i) the active undercutting by rivers may lead to slope destabilization without significantly increasing the average slope gradient as extracted from the SRTM DEM and (ii) tectonic uplift may induce rock and regolith fracturing, leading to weaker, more LS-prone slopes. The fact that we did not observe differences in LS frequency on hillslopes that were deforested long ago suggests that on such slopes a new equilibrium is established whereby these aforementioned mechanisms are no longer important.

In conclusion, one of the key factors why the rejuvenated landscape inside the rift is more sensitive to landsliding is the higher prevalence of threshold slopes due to active incision. However, the impact of rejuvenation cannot be understood by considering only its effects on overall topography. Deforestation dramatically increases LS frequency in both relict and rejuvenated landscapes, in the first decades after forest cover removal.

How to cite: Depicker, A., Govers, G., Jacobs, L., Campforts, B., Uwihirwe, J., and Dewitte, O.: Landslide frequency in the Kivu Rift: impact of landscape evolution and deforestation, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-17974, https://doi.org/10.5194/egusphere-egu2020-17974, 2020.

Estimation of the volumes of potential future debris flows is key for hazard assessment and mitigation. Worldwide, however, there are few catchments for which detailed volume-frequency information is available. We (1) reconstruct volume-frequency curves for 10 debris-flow catchments in Saline Valley, California, USA, from a large number of well-preserved, unmodified surficial flow deposits, and (2) assess the correlations between lobe-volume quantiles and a set of morphometric catchment characteristics. We find statistically-significant correlations between lobe-volume quantiles, including median and maximum, and catchment relief, length (planimetric distance from the fan apex to the most distant point along the watershed boundary), perimeter, and Melton ratio (relief divided by the square root of catchment area). These findings show that it may be possible to roughly estimate debris-flow lobe-volume quantiles from basic catchment characteristics that can be obtained from globally-available elevation data. This may assist with design-volume estimation in debris-flow catchments where past flow volumes are otherwise unknown.

How to cite: Densmore, A. and de Haas, T.: Debris-flow volume quantile prediction from catchment morphometry, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-10327, https://doi.org/10.5194/egusphere-egu2020-10327, 2020.

Landslides occurred in the Northeastern part of Turkey are generally classified to be shallow seated landslides or earthflow type based on Varnes (1984) classification. These landslides are occasionally seen in the weathered  Eocene or Upper Crateceous aged volcanic and volcano-clastic rocks. Although there are considerable studies both directly on these landslides in point of mapping and hazard assessment, there is no any studies concerning size and magnitude characteristics of them. By considering this point, an assessment of size and magnitude characteristics of shallow seated type landslides at an area where is one of most landslide prone area of Turkey was carried out.

The investigation area is totaly covered by Eocene aged volcano-clastic lithology, and the weathering is widespread due to the climatical conditions in the area. The extend of the area is 140 square kilometers. In the area, 120 landslides were mapped by a multiple image interpretation that is from the years of 2000 to 2019. To do this, Google Earth images were used. In the area, the area (A_L) of the landslide mapped differs from 53.28 m² to 902,809 m². The length and width of these landslide were also determined and these characteristics were taken into account for an assessment of relationship between the size and topographical features such as slope gradient, curvature, topographical wetness index and stream power index. The approximate volumes of these landslides were calculated by considering direct depth observations in the field surveys, and then assessed by different relations proposed by different studies. Magnitude (M) of these landslides were also assessed by taken into account of the area (A_L) and volume (V_L) values.

How to cite: Tezel, K., Akgün, A., and Alizadeh, E.: Assessment of shallow seated landslide size and magnitude characteristics: An example from Northeastern Turkey , EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-22033, https://doi.org/10.5194/egusphere-egu2020-22033, 2020.

Landslides are one of the most common and dangerous threats in the world that generate considerable damage and economic losses. The aim of this study is to analyse the suitability of using the ESA G-POD (European Space Agency Grid Processing On Demand) environment to detect landslide incidence.

This free service allows to gain Mean Deformation Maps (mm/year and cm/year) by means of P-SBAS (Parallel-Small Baseline Subset) method, which is a kind of A-DInSAR (Advanced-Differential SAR Interferometry) technique.

The study area is in the Northwest of Spain, where there are and have been some significant well-known active landslides. ENVISAT ASAR satellite data collected from 2003 to 2010, have been contrasted with the slope instabilities inventory of Asturias (BAPA: Base de datos de Argayos del Principado de Asturias - Principality of Asturias Landslide Database), from 2003 to 2010. Afterwards, a new check with instability data registered in BAPA dataset from 2010 to 2017 has been done.

A-DInSAR and BAPA data have been jointly integrated and examined in a GIS. The results obtained indicate that there is consistency between both types of data. In addition, this research has been useful to highlight the G-POD free service as a reliable, economic and adequate tool to analyse movements of terrain during time periods of several years in the North of Spain.

How to cite: Cuervas-Mons, J., Domínguez-Cuesta, M. J., Mateos-Redondo, F., and Valenzuela, P.: LANDSLIDE INCIDENCE ANALYSIS BY MEANS OF A-DInSAR OPEN ACCESS SERVICE AND LANDSLIDE DATABASE, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-354, https://doi.org/10.5194/egusphere-egu2020-354, 2020.

To overcome the food and water shortages and optimize the land use, remote sensing techniques and satellite image processing have utilized our demands. However, with limitations in image processes, the use of such techniques will need further development to overcome related constraints. Shadows, occurred on the opposite side of objects, result from topography and different angles of the emitting light source is one of these limitations. Several topographic correction methods are proposed based on the properties of ground coverage. To suggest and compare methods for imagery topography, this study uses Cosine Correction, C-Correction, Statistical Empirical Correction, and finally the Minnaert Correction. The study area used to compare the introduced methods is located in North West of Isfahan (Ardestan), Iran. The current report has used OLI sensors (LANDSAT 8) combined with ASTER global digital elevation data. After implementing topographic corrections, by optimal index OIF, images are processed. Based on the unsupervised method and the study region, results based on optimal arrangement bands are introduced as a suitable classification. In conclusion, based on imagery and statistical data from the topography corrections, Minnaret shows the most exceptional topographical correction classification for the chosen studied region.

Keywords: aster, c-correction, cosine correction, Isfahan, Landsat-8, land management, Minnaert, oli, topographic correction, unsupervised classification.

How to cite: Karimi, H., Shahabi Ghahfarokhi, S., and Arfania, R.: Comparison of topographic corrections for land‐cover classification in mountainous terrain, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-22085, https://doi.org/10.5194/egusphere-egu2020-22085, 2020.

We present the approach taken to map surface rock exposures in upland areas of Scotland. This has been carried out as a means of enhancing the mapping of superficial sediment thickness which has important applications including the assessment of potential geohazard susceptibility. The presented study includes selected test cases that have been constructed prior to scaling up the approach to upland areas across Great Britain (GB).

The presence of rock at surface acts as a marker of locations with minimal superficial sediment cover (essentially a zero depth). The thickness of superficial sediments across GB are currently estimated based on borehole records which range in both quality and coverage, with limited data particularly for upland regions. Superficial sediment thickness is an integral factor for assessing geohazard processes including landslides. Therefore, by improving datasets detailing rock at surface, we can enhance superficial sediment thickness estimates and enhance the variable inputs to the models used to assess geohazard susceptibility.

The GB landscape has been subject to a range of different environmental processes through time with its current topography being the subject of glacial erosion through to marine incursions. However, these patterns are not uniform and this results in a range of landscapes. The resulting domains are an important consideration when attempting to model the relationship between the presence and absence of natural rock exposures.  With a wealth of information available across GB including high resolution topography, the resulting (often scale-dependent) geomorphometric derivatives, geological datasets as well as satellite imagery, we are able to consider a range of possible relationships that might exist. We combine these datasets coupled with field validation of rock absence/presence to train a random forest classifier for specific domains with the aim being to identify a way of modelling rock exposure in areas of limited data availability as is the case for many upland areas.

The methodology and results of the approach for specific process domains will be presented with a specific focus on the Glen Gyle catchment, at the head of Loch Katrine (the primary water reservoir for the city of Glasgow) in the Trossachs National Park, Scotland. This is an area that has been subject to recent landslides which have affected local properties and infrastructure.

How to cite: Williams, C., Finlayson, A., Palamakumbura, R., Kearsey, T., Cornillon, S., and Whitbread, K.: Mapping surface rock exposures to enhance geohazard susceptibility assessment: a random forest approach , EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-5017, https://doi.org/10.5194/egusphere-egu2020-5017, 2020.

The availability of liquid water plays a primary role in controlling the development of topography. Hillslope asymmetry (HA), or slope differences between terrain aspects, has been well-documented in small-scale and field-based studies throughout the world. In this study, we apply a consistent HA analysis method across the entire globe and find that poleward facing hillslopes are on average steeper than equator-facing hillslopes, with the exception of high-latitude, high-elevation, and low-temperature regions where equator-facing slopes tend to be steeper.

To test the impact of different land cover and climate regimes on HA, we use global and high-resolution elevation, vegetation and land-surface temperature data to examine erosional process differences between poleward- and equator-facing hillslopes. We find that vegetation supports poleward-steepening, and that low temperatures and high freeze-thaw cycle frequencies enhance equator-steepening of hillslopes. We further show that HA is propagated into the size and form of fluvial drainage networks. We posit that insolation plays a key role in controlling soil-water availability and retention, and thus drives asymmetries in vegetation cover, soil formation rates and landscape form at the planetary scale.

How to cite: Smith, T. and Bookhagen, B.: Shaping Planetary Surfaces: The Impact of Liquid and Frozen Water on Hillslope Topography, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-8180, https://doi.org/10.5194/egusphere-egu2020-8180, 2020.