Search

Lack of global data inventories obstructs scientific modeling of and response to landslide hazards which are oftentimes deadly and costly. To remedy this limitation, new approaches suggest solutions based on citizen science that requires active participation. In contrast, as a non-traditional data source, social media has been increasingly used in many disaster response and management studies in recent years. Inspired by this trend, we propose to capitalize on social media data to mine landslide-related information automatically with the help of artificial intelligence techniques. Specifically, we develop a state-of-the-art computer vision model to detect landslides in social media image streams in real-time. To that end, we first create a large landslide image dataset labeled by experts with a data-centric perspective, and then, conduct extensive model training experiments. The experimental results indicate that the proposed model can be deployed in an online fashion to support global landslide susceptibility maps and emergency response.

The Derbyshire Platform is a Mississippian aged flat-topped, steep sided platform that forms the westernmost expression of the Derbyshire-East Midlands Platform. On the south-east platform margin, 60 km2 of Visean limestone has been dolomitised, forming two distinct bodies. One of these bodies forms along a major NW–SE trending basement fault and smaller, associated, N–S trending faults and fractures. This study uses outcrop, petrographic and geochemical analysis to better constrain the timing and mechanism for this fault-controlled dolomitisation. Field relationships demonstrate dolomitisation was multi-phase and initiated after the main phase of matrix pore-occluding calcite cementation on the Derbyshire Platform and terminated prior to the main phase of mineralisation. Fluids are interpreted to have fluxed from adjacent basins, primarily along strike-slip crustal faults that were reactivated during basin inversion at the onset of the Variscan Orogeny. Fluid supply was episodic and progressively confined to fractures as matrix porosity became occluded. The study demonstrates the complex interplay between basin kinematics, host rock permeability and timing of fluid supply through seismic valving along faults that connect the carbonate platform to basin compartments. This ultimately controlled the position of dolomite geobodies along faults and provides a record of fluid flow during the transition from thermal subsidence to post-rift basin inversion. The findings have implications for the exploration of both minerals and hydrocarbon within dolomitised host rocks and can inform studies of fluid transfer and reaction on carbonate platforms within the burial realm.

Landslide susceptibility modelling in tropical climates is hindered by incomplete inventory due to rapid development and natural processes that obliterate field evidence, making validation a challenge. Susceptibility modelling was conducted in Kuala Lumpur, Malaysia using a new spatial partitioning technique for cross-validation. This involved a series of two alternating east-west linear zones, where the first zone served as the training dataset and the second zone was the test dataset, and vice versa. The results show that the susceptibility models have good compatibility with the selected landslide conditioning factors and high predictive accuracy. The model with the highest area under curve (AUC) values (SRC = 0.92, PRC = 0.90) was submitted to the City Council of Kuala Lumpur for land use planning and development control. Rainfall-induced landslides are prominent within the study area, especially during the monsoon period. An extreme rainfall event in December 2021 that triggered 122 landslides provided an opportunity to conduct retrospective validation of the model; the high predictive capability (AUC of PRC = 0.93) was reaffirmed. The findings proved that retrospective validation is vital for landslide susceptibility modelling, especially where the inventory is not of the best quality. This is to encourage wider usage and acceptance among end users, especially decision-makers in cities, to support disaster risk management in a changing climate.

Ripon is a historic cathedral city and market town located in North Yorkshire, UK. It is the only city in the UK enforcing classified planning zones based on the potential for subsidence caused by bedrock dissolution. The underlying geology comprises Permian limestone, marls and gypsum, all capped by superficial deposits. It is the rapid rate of dissolution of the gypsum (CaSO4 2H2O) that enables karst features to form and evolve on human timescales. Following a collapse that formed a sinkhole in the garden of a property on Magdalens Road, Ripon, in November 2016, the British Geological Survey (BGS), funded by a Natural Environment Research Council (NERC) urgency grant, initiated a programme of research. Monitoring began on a depression feature forming in a nearby recreation area that had been reported by residents to have increased in dimensions. A geophysical survey was undertaken utilising integrated ground-based LiDAR imaging with geophysical techniques (Williams et al. 2020). Alongside this work, a 3D geological ground model was built to provide the wider geological context. The Ripon 3D geological model was one of the first applied geological models to be piloted in the BGS Groundhog Geological Software. Initially created using GSI3D (Geological Surveying and Investigation in 3 Dimensions), (Mathers et al. 2011), the original data was imported into Groundhog, testing the interoperable capabilities of the software (Wood et al, 2015). Setting the sinkhole and the geophysical surveys in a spatial 3D context, it may be possible to allow interpretations to be made with respect to the processes associated with the formation of sinkhole geohazards. BGS, as the UK’s premier provider of objective and public-good geoscientific data, works with the local community, councils, and geotechnical consultants to communicate a better understanding of where sinkhole hazards may occur. The model will become a useful and integral part of planning and geological investigations and

Tropics make up about 40% of the earth’s surface area and are home to approximately 40% of the world’s population [1]. Significant and variable depths of residual soils occur across tropical regions for instance Southern India, South-West China, Malaysia, Africa, Taiwan, Hong Kong, Singapore, Sri Lanka, the Philippines, and Northern Brazil. Rapid development in these countries, coupled with changing climate, results in an increasing number of landslides affecting people, property, and livelihoods. Step changes in landslide risk management can be achieved through landslide susceptibility mapping in conjunction with the determination of thresholds at which landslides might be initiated. The landslide trigger thresholds can be defined as a condition that when overcome can cause a change in the state of a system or landscape resulting in one or more landslides [2]. These thresholds can be developed at different spatial scales using approaches such as process-based, conceptual, empirical, statistical, or probabilistic methods [3]. The existing review of studies on trigger thresholds by [2] highlights the challenges of limited landslide inventories, quality of datasets, need for objectivity and reproducibility, and validation. The landslide trigger threshold in tropical countries mostly uses empirical or data-driven methods, which are usually based on incomplete and/or impact-biased landslide inventories. The incomplete or limited landslide inventory further limits the development of the threshold considering the regional or domain-specific ground-based information such as geology, geomorphology, climate, and hydrology. Therefore, the resulting trigger thresholds at regional-scale in a tropical setting are less reliable for landslide forecasting-based applications. The landslides in tropical residual soils may not always directly correlate to rainfall amount or rainfall intensity owing to the importance of soil properties [4], derived from the underlying geology and anthr

This paper presents an online system that leverages social media data in real time to identify landslide-related information automatically using state-of-the-art artificial intelligence techniques. The designed system can (i) reduce the information overload by eliminating duplicate and irrelevant content, (ii) identify landslide images, (iii) infer geolocation of the images, and (iv) categorize the user type (organization or person) of the account sharing the information. The system was deployed in February 2020 online at https://landslide-aidr.qcri.org/landslide_system.php to monitor live Twitter data stream and has been running continuously since then to provide time-critical information to partners such as British Geological Survey and European Mediterranean Seismological Centre. We trust this system can both contribute to harvesting of global landslide data for further research and support global landslide maps to facilitate emergency response and decision making.

The aim of this report is to provide insight into community-based methods, approaches, and actions for reducing risk to landslides. More specifically, it presents a broad overview of recent studies on landslide risk reduction at community level, exploring the various landslide risk reduction measures recommended and/or implemented, and discusses the challenges and opportunities for the implementation of these measures. The report draws on a wide range of applications investigating distinct case studies in different areas around the Globe, as well as more general studies on equitable resilience (how communities are really engaged) and landslide risk management within the context of Disaster Risk Reduction (DRR) in the Global South's vulnerable communities. The documented strategies are analysed from the perspectives of scale of intervention, their timing and sustainability, and the resources required for their implementation. The review suggests that there is no one-size fits all solution for community-based landslide risk reduction. The integration of local knowledge into landslide risk reduction practices is context-specific and varies both in relation to the time of implementation (ex-ante or ex-post disaster) and historically, due to the dynamic nature of communities’ structure and functioning. Its contribution to resilience (including coping and adaptation capacity) depends on the interaction with other types of knowledge (e.g., science based) and the general institutional setting (legal and governmental framework). Moreover, the scale of organisation and action, from individual to household and community level, influences the impact and long-term sustainability of mitigation measures. Nevertheless, overcoming barriers of knowledge, trusts, resources, and power at local level could enhance co-development and collaboration between communities and governmental and non-governmental organisation, communities Landslide risk reduction cannot be addressed in isolation.

Fault-controlled dolomitization has been documented in Lower Carboniferous (Viséan) platform carbonates at various localities in the Pennine Basin and North Wales. The largest of these dolomite bodies (approx. 60 km2) occurs on the Derbyshire Platform, on the southern margin of the Pennine Basin. This study tests the hypothesis that dolomitization occurred at this locality during deposition, platform drowning, and the earliest stages of burial, coincident with the transition from a late syn-rift to post-rift regime. It also assesses the importance of syn-rift volcanism on dolomitization. Planar, fabric-retentive dolomite with single-phase (i.e., low temperature) fluid inclusions occurs along NW–SE and E–W oriented faults, and in platform margin facies and in proximity to the Masson Hill Volcanic Complex. Oxygen isotope data are consistent with dolomitization from seawater, but slightly depleted δ13C values reflect mixing with magmatic fluids. Volcanic activity is likely to have produced a thermal drive for fluid circulation on the platform margin, and post-depositional alteration of basalts by CO2-rich fluids could have led to alteration of olivine and release of magnesium to convecting seawater. Consequently, the large volume of dolostone on the southern margin of the Derbyshire Platform is attributed to the increased geothermal gradient and a localized increase in the Mg/Ca ratio of dolomitizing fluids at this locality, compared to elsewhere in the Pennine Basin. The results suggest that syn-rift carbonate platforms in volcanically active areas of rift basins have a greater potential for dolomitization from seawater than non-volcanic platforms in the same basin.

Swelling and shrinking soils are soils that can experience large changes in volume due to changes in water content. This may be due to seasonal changes in moisture content, local site changes such as leakage from water supply pipes or drains, changes to surface drainage and landscaping, or following the planting, removal or severe pruning of trees or hedges. These soils represent a significant hazard to structural engineers across the world due to their shrink–swell behaviour, with the cost of mitigation alone running into several billion pounds annually. These soils usually contain some form of clay mineral, such as smectite or vermiculite, and can be found in humid and arid/semi-arid environments where their expansive nature can cause significant damage to properties and infrastructure. This chapter discusses the properties and costs associated with shrink–swell soils, their formation and distribution throughout the UK and the rest of the world, and their geological and geotechnical characterization. It also considers the mechanisms of shrink-swell soils and their behaviour, reviewing strategies for managing them in an engineering context, before finally outlining the problem of trees and shrink–swell soils.

The Brassington Formation of the Pennines is the most extensive onshore Miocene succession in the UK. It is preserved as outliers in Lower Carboniferous Limestone. During the Cenozoic, central England underwent uplift, with erosion of post-Mississippian strata from the Pennine axis in the Peak District. The Brassington Formation is hence significant in reconstructing Cenozoic geological history. It is non-marine, derived from Triassic sandstone and of Mid–Late Miocene age. The c. 60 outliers occur in three clusters over c. 220 km2. They are remnants of a sedimentary prism at least 75 m thick and with a volume of c. 10 km3. Suffosion of bedrock, conditioned by aggressive precursor fluids of hypogenic origin, was the major control on subsidence. The calculated volume of the fills in the Bees Nest and Green Clay outliers is 3.7–5.0 × 106 m3. Referenced to a sub-Miocene surface at c. 450 m OD in pre-subsidence times, the original volume of this subsidence complex was c. 21.8 × 106 m3 and the aggregated volume of the Late Neogene White Peak palaeokarst was at least 0.66 km3. Subsidence was concomitant with Pliocene uplift of the Pennine Axis, suggesting suffosion accommodation of 3 × 106 m3 per km2 or 244 m3 per annum.