Prediction of landslide displacements as a function of rainfall is a hot topic for the management of landslide hazards. The availability of long time series of in situ network encourages the use of data-driven models. These approaches, starting from measured data, allow to identify models, which can predict the behaviours of the investigated phenomenon in a pragmatic, but reliable way and can also give indirect information about the dynamics of the system. In particular, data-driven approaches are of particular success whereas the knowledge about the system to be modelled is incomplete, and unknown boundary conditions and/or unmeasured extra inputs exist. This work introduces the use of an evolutionary modelling technique, namely EPRMOGA in predicting landslide displacements as function of antecedent displacements and rainfall. EPRMOGA is a methodology based on a Genetic Algorithm that returns models as closed-form equations, characterized by relatively simple structures, thus permitting to make some realistic speculations about the terms and the input of the models. These speculations can be used to advance some careful physical interpretation about the constructed models, from which it is possible to get some scientific knowledge discovery. In summary, the use of EPRMOGA allows to get both reasonably accurate predictions of landslide displacements and scientific knowledge about how landslide displacements occur. Here a case study is presented about a landslide located in north Italy, on the east side of the lake of Como, in Valsassina valley. The outcome of EPRMOGA is an equation, which can predict the displacements with a high accuracy level, up to several months ahead. Moreover, the identified equation permits to highlight those rainfall cumulative values, which apparently have a major influence on displacements.

Prediction of landslide displacement as a function of antecedent landslide by EPRMOGA tool / Doglioni, Angelo; Crosta, Gb; Frattini, P; Melidoro, Nl; Simeone, Vincenzo. - (2013), p. 311. (Intervento presentato al convegno GEOITALIA 2013 - IX FORUM di Scienze della Terra tenutosi a Pisa nel 16-18 settembre 2013).

Prediction of landslide displacement as a function of antecedent landslide by EPRMOGA tool

DOGLIONI, Angelo;SIMEONE, Vincenzo
2013-01-01

Abstract

Prediction of landslide displacements as a function of rainfall is a hot topic for the management of landslide hazards. The availability of long time series of in situ network encourages the use of data-driven models. These approaches, starting from measured data, allow to identify models, which can predict the behaviours of the investigated phenomenon in a pragmatic, but reliable way and can also give indirect information about the dynamics of the system. In particular, data-driven approaches are of particular success whereas the knowledge about the system to be modelled is incomplete, and unknown boundary conditions and/or unmeasured extra inputs exist. This work introduces the use of an evolutionary modelling technique, namely EPRMOGA in predicting landslide displacements as function of antecedent displacements and rainfall. EPRMOGA is a methodology based on a Genetic Algorithm that returns models as closed-form equations, characterized by relatively simple structures, thus permitting to make some realistic speculations about the terms and the input of the models. These speculations can be used to advance some careful physical interpretation about the constructed models, from which it is possible to get some scientific knowledge discovery. In summary, the use of EPRMOGA allows to get both reasonably accurate predictions of landslide displacements and scientific knowledge about how landslide displacements occur. Here a case study is presented about a landslide located in north Italy, on the east side of the lake of Como, in Valsassina valley. The outcome of EPRMOGA is an equation, which can predict the displacements with a high accuracy level, up to several months ahead. Moreover, the identified equation permits to highlight those rainfall cumulative values, which apparently have a major influence on displacements.
2013
GEOITALIA 2013 - IX FORUM di Scienze della Terra
Prediction of landslide displacement as a function of antecedent landslide by EPRMOGA tool / Doglioni, Angelo; Crosta, Gb; Frattini, P; Melidoro, Nl; Simeone, Vincenzo. - (2013), p. 311. (Intervento presentato al convegno GEOITALIA 2013 - IX FORUM di Scienze della Terra tenutosi a Pisa nel 16-18 settembre 2013).
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11589/25108
Citazioni
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact