Landslide size matters: A new data-driven, spatial prototype

Lombardo, Luigi; Tanyaş, Hakan; Huser, Raphaël; Guzzetti, Fausto; Castro‐Camilo, Daniela

doi:10.1016/j.enggeo.2021.106288

Cited by 54 publications

(36 citation statements)

References 122 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Guzzetti et al (1999) proposed to alternatively model landslide areas, which can be easily extracted from a polygonal inventory. Nevertheless, the first spatially-explicit model able to estimate landslide areas has only been recently proposed by Lombardo et al (2021). In their work, the authors exclusively estimated the potential landslide size at a given location, without informing whether the given location would have been susceptible in the first place.…”

Section: Introductionmentioning

confidence: 99%

Space-time landslide hazard modeling via Ensemble Neural Networks

Dahal¹,

Tanyaş²,

Westen³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

For decades, a full numerical description of the spatio-temporal dynamics of a landslide could be achieved only via physics-based models. The part of the geomorphology community focusing on data-driven model has instead focused on predicting where landslides may occur via susceptibility models. Moreover, they have estimated when landslides may occur via models that belong to the early-warning-system or to the rainfall-threshold themes. In this context, few published research have explored a joint spatio-temporal model structure. Furthermore, the third element completing the hazard definition, i.e., the landslide size, has hardly ever been modeled over space and time. However, the technological advancements of data-driven models have reached a level of maturity that allows to model all three components (Where, When and Size) mentioned above. This work, takes this direction and proposes for the first time a solution to the assessment of landslide hazard in a given area by jointly modeling landslide occurrences and their associated areal density per mapping unit, in space and time. To achieve this ambitious task, we have used a spatio-temporal landslide database generated for the Nepalese region affected by the Gorkha earthquake on the 25th of April 2015. The model relies on a deep-learning architecture trained using an Ensemble Neural Network, where the landslide occurrences and densities are aggregated over a squared mapping unit of 1 x 1 km and classified/regressed against a nested 30 m lattice. At the nested level, we have expressed predisposing and triggering factors. As for the temporal units, we have used an approximately 6-month resolution depending on the mapped inventory dates. The results are promising as our model performs satisfactorily both in the classification (susceptibility) and regression (density prediction) tasks. We believe that the model we propose brings a level of novelty that has the potential to create a rift with respect to the common susceptibility literature, finally proposing an integrated framework for hazard modeling in a data-driven context.

show abstract

Section: Introductionmentioning

confidence: 99%

Space-time landslide hazard modeling via Ensemble Neural Networks

Dahal¹,

Tanyaş²,

Westen³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…For this reason, we already envision future models that would take the measured extent of TS and TEG as the response variable, this time solving a regression task rather than a classification, one as per susceptibility requirement. Such direction has recently been explored for landslides occurring at lower latitudes (Lombardo et al, 2021;Moreno et al, 2022). And, an even better extension has already been tested where the expectation of locations prone to landslides are modelled together with the expectation of the resulting landslide size (Aguilera et al, 2022;Bryce et al, 2022).…”

Section: Considerations Within and Beyond Svalbard: Supporting And Op...mentioning

confidence: 99%

Multi-hazard susceptibility mapping of cryospheric hazards in a high-Arctic environment: Svalbard Archipelago

Nicu

Elia

Rubensdotter

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Abstract. The Svalbard Archipelago represents the northernmost place on Earth where cryospheric hazards, such as thaw slumps (TS) and thermo-erosion gullies (TEG) could take place and rapidly develop under the influence of climatic variations. Svalbard permafrost is specifically sensitive to rapidly occurring warming and therefore, a deeper understanding of TS and TEG is necessary to understand and foresee the dynamics behind local cryospheric hazards' occurrences and their global implications. We present the latest update of two polygonal inventories where the extent of TS and TEG is recorded across Nordenskiöld Land (Svalbard Archipelago), over a surface of approximately 4000 km2. This area was chosen because it represents the most concentrated ice-free area of the Svalbard Archipelago and, at the same time, where most of the current human settlements are concentrated. The inventories were created through visual interpretation of high-resolution aerial photographs, as part of our ongoing effort toward creating a pan-Arctic repository of TS and TEG. Overall, we mapped 562 TS and 908 TEG, from which we separately generated two susceptibility maps using a Generalized Additive Modelling (GAM) approach, under the assumption that TS and TEG manifest across Nordenskiöld Land, according to a Bernoulli probability distribution. Once validating the modelling results, the two susceptibility patterns were combined into the first multi-hazard cryospheric susceptibility map of the area. The two inventories are available for the whole community at https://doi.org/10.1594/PANGAEA.945348 (Nicu et al., 2022a) and https://doi.pangaea.de/10.1594/PANGAEA.945395 (Nicu et al., 2022b).

show abstract

“…LGCPs are widely used in the modeling of spatial point patterns and have risen as a counterpart of Gaussian processes (used for continuous variables) for modeling discrete spatial phenomena. Recently, LGCP models have drawn attention in modeling landslide occurrences; see, e.g., Lombardo et al (2018Lombardo et al ( , 2020Lombardo et al ( , 2021 and Opitz et al (2020). In this work, we take LGCP models as a basis and in § §3.2-3.4, we build joint models for landslide occurrences and sizes (here defined as the square root of landslide areas), where sizes are viewed as numerical marks, each associated with exactly one point.…”

Section: Poisson Processes Andmentioning

confidence: 99%

“…Precisely, we shall model the square root of the landslide area, which is lighter-tailed than the area itself, and can be measured and intuitively interpreted on the same scale as the landslide diameter. In the literature, other attempts have already been made to model and predict landslide sizes or "magnitudes" in addition to the modeling of landslide counts (see, e.g., Guo et al, 2017;Roback et al, 2018;Valagussa et al, 2019;Vanani et al, 2021;Lombardo et al, 2021). For example, Guo et al (2017) used power-law relationships of the size distribution to study earthquake-induced landslides in both the Himalayan and Lesser Himalayan regions.…”

Section: Introductionmentioning

confidence: 99%

“…In some related work, Vanani et al (2021) studied the correlation between the landslide counts and sizes using a bivariate model and a linear automated modeling procedure using the SPSS software. More recently, Lombardo et al (2021) proposed a statistically-based model to estimate the area of landslides aggregated over slope units, though they completely ignored their occurrence probability. Specifically, they proposed a Bayesian version of generalized additive models where both the maximum landslide size per slope unit and the sum of all landslide sizes per slope unit were modeled using a log-normal model.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Joint modeling of landslide counts and sizes using spatial marked point processes with sub-asymptotic mark distributions

Yadav¹,

Huser²,

Opitz³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

To accurately quantify landslide hazard in a region of Turkey, we develop new marked point process models within a Bayesian hierarchical framework for the joint prediction of landslide counts and sizes. To accommodate for the dominant role of the few largest landslides in aggregated sizes, we leverage mark distributions with strong justification from extreme-value theory, thus bridging the two broad areas of statistics of extremes and marked point patterns. At the data level, we assume a Poisson distribution for landslide counts, while we compare different "sub-asymptotic" distributions for landslide sizes to flexibly model their upper and lower tails. At the latent level, Poisson intensities and the median of the size distribution vary spatially in terms of fixed and random effects, with shared spatial components capturing cross-correlation between landslide counts and sizes. We robustly model spatial dependence using intrinsic conditional autoregressive priors.Our novel models are fitted efficiently using a customized adaptive Markov chain Monte Carlo algorithm. We show that, for our dataset, sub-asymptotic mark distributions provide improved predictions of large landslide sizes compared to more traditional choices. To showcase the benefits of joint occurrence-size models and illustrate their usefulness for risk assessment, we map landslide hazard along major roads.

show abstract

Landslide size matters: A new data-driven, spatial prototype

Cited by 54 publications

References 122 publications

Space-time landslide hazard modeling via Ensemble Neural Networks

Space-time landslide hazard modeling via Ensemble Neural Networks

Multi-hazard susceptibility mapping of cryospheric hazards in a high-Arctic environment: Svalbard Archipelago

Joint modeling of landslide counts and sizes using spatial marked point processes with sub-asymptotic mark distributions

Contact Info

Product

Resources

About