Susceptibility assessment of shallow slides failure and run-out

Melo, Raquel; Zêzere, José; Rocha, Jorge; Oliveira, Sérgio

Melo, Raquel,Zêzere, José,Rocha, Jorge,Oliveira, Sérgio · 2019

Abstract

The research is focused on the susceptibility assessment of shallow slides in the region north of Lisbon (Portugal), by modelling the failure and run-out areas separately. The shallow slides failure is evaluated using a statistical method (logistic regression). The existence of shallow slides inventories occurred in distinct periods allowed the separation of data into two independent groups (training and validation) and the adoption of the temporal criterion for the independent validation. The latter revealed an Area Under the Receiver Operating Characteristic curve of 0.90, which reflects a very good predictive capacity of the logistic regression model. For the run-out assessment, a simple cellular automata model is implemented through the following sequential steps: a) pre-processing and establishment of transition rules; b) integration of variables; and c) temporal indexing and simulation. The pre-processing step includes the creation of a database with the modelling inputs. The transition rules are directly related with the motion of the displaced mass. In this context, the likely traveling directions are identified, both horizontally and vertically. The integration of transition rules is performed using the algorithm Path Distance, from ESRI. For the temporal indexing, we use the Markov chains analysis to estimate a transition area matrix, which records the number of cells that is expected to change location over a specified time. The last stage refers to the cellular automata model simulation, i.e. to the spatial distribution of the landslide displaced mass. The run-out modelling, using the cellular automata model proposed, provided good results, with an overlap between the simulation and the real cases of 77%. Lastly, a final shallow slide susceptibility map was constructed including both failure and run-out areas. This work accomplished a combination of low-cost methodology with limited input data that allowed a good performance of the landslide susceptibility assessment and can be easily applied to other regions.

Read accessible full text

Susceptibility assessment of shallow slides failure and run-out

Author: Melo, Raquel,Zêzere, José,Rocha, Jorge,Oliveira, Sérgio

Publisher: European Geosciences Union

Year: 2019

Source: https://repositorio.ulisboa.pt/bitstream/10451/41293/1/EGU2019-2983.pdf

Geophysical Resea ch Abs ac s
Vol. 21, EGU2019-2983, 2019
EGU Gene al Assembly 2019
© Au ho (s) 2019. CC A ibu ion 4.0 license.
Suscep ibili y assessmen o shallow slides ailu e and un-ou
Raquel Melo, José Luís Zêze e, Jo ge Rocha, and Sé gio C uz Oli ei a
Uni e sidade de Lisboa, Cen o de Es udos Geog á icos, IGOT, Lisboa, Po ugal
The esea ch is ocused on he suscep ibili y assessmen o shallow slides in he egion no h o Lisbon (Po ugal),
by modelling he ailu e and un-ou a eas sepa a ely. The shallow slides ailu e is e alua ed using a s a is ical
me hod (logis ic eg ession). The exis ence o shallow slides in en o ies occu ed in dis inc pe iods allowed he
sepa a ion o da a in o wo independen g oups ( aining and alida ion) and he adop ion o he empo al c i e ion
o he independen alida ion. The la e e ealed an A ea Unde he Recei e Ope a ing Cha ac e is ic cu e o
0.90, which e lec s a e y good p edic i e capaci y o he logis ic eg ession model.
Fo he un-ou assessmen , a simple cellula au oma a model is implemen ed h ough he ollowing sequen ial
s eps: a) p e-p ocessing and es ablishmen o ansi ion ules; b) in eg a ion o a iables; and c) empo al indexing
and simula ion. The p e-p ocessing s ep includes he c ea ion o a da abase wi h he modelling inpu s. The
ansi ion ules a e di ec ly ela ed wi h he mo ion o he displaced mass. In his con ex , he likely a eling
di ec ions a e iden i ied, bo h ho izon ally and e ically. The in eg a ion o ansi ion ules is pe o med using he
algo i hm Pa h Dis ance, om ESRI. Fo he empo al indexing, we use he Ma ko chains analysis o es ima e
a ansi ion a ea ma ix, which eco ds he numbe o cells ha is expec ed o change loca ion o e a speci ied
ime. The las s age e e s o he cellula au oma a model simula ion, i.e. o he spa ial dis ibu ion o he landslide
displaced mass. The un-ou modelling, using he cellula au oma a model p oposed, p o ided good esul s, wi h
an o e lap be ween he simula ion and he eal cases o 77%. Las ly, a inal shallow slide suscep ibili y map
was cons uc ed including bo h ailu e and un-ou a eas. This wo k accomplished a combina ion o low-cos
me hodology wi h limi ed inpu da a ha allowed a good pe o mance o he landslide suscep ibili y assessmen
and can be easily applied o o he egions.
Acknowledgmen s: This wo k was suppo ed by he p ojec BeSa eSlide—Landslide Ea ly Wa ning so
echnology p o o ype o imp o e communi y esilience and adap a ion o en i onmen al change (PTDC/GES-
AMB/30052/2017).