Improving transferability strategies for debris flow susceptibility assessment: Application to the Saponara and Itala catchments (Messina, Italy)

Handle URI:
http://hdl.handle.net/10754/623851
Title:
Improving transferability strategies for debris flow susceptibility assessment: Application to the Saponara and Itala catchments (Messina, Italy)
Authors:
Cama, M.; Lombardo, Luigi; Conoscenti, C.; Rotigliano, E.
Abstract:
Debris flows can be described as rapid gravity-induced mass movements controlled by topography that are usually triggered as a consequence of storm rainfalls. One of the problems when dealing with debris flow recognition is that the eroded surface is usually very shallow and it can be masked by vegetation or fast weathering as early as one-two years after a landslide has occurred. For this reason, even areas that are highly susceptible to debris flow might suffer of a lack of reliable landslide inventories. However, these inventories are necessary for susceptibility assessment. Model transferability, which is based on calibrating a susceptibility model in a training area in order to predict the distribution of debris flows in a target area, might provide an efficient solution to dealing with this limit. However, when applying a transferability procedure, a key point is the optimal selection of the predictors to be included for calibrating the model in the source area. In this paper, the issue of optimal factor selection is analysed by comparing the predictive performances obtained following three different factor selection criteria. The study includes: i) a test of the similarity between the source and the target areas; ii) the calibration of the susceptibility model in the (training) source area, using different criteria for the selection of the predictors; iii) the validation of the models, both at the source (self-validation, through random partition) and at the target (transferring, through spatial partition) areas. The debris flow susceptibility is evaluated here using binary logistic regression through a R-scripted based procedure.Two separate study areas were selected in the Messina province (southern Italy) in its Ionian (Itala catchment) and Tyrrhenian sides (Saponara catchment), each hit by a severe debris flow event (in 2009 and 2011, respectively).The investigation attested that the best fitting model in the calibration areas resulted poorly performing in predicting the landslides of the test target area. At the same time, the susceptibility models calibrated with an optimal set of covariates in the source area allowed us to produce a robust and accurate prediction image for the debris flows activated in the Saponara catchment in 2011, exploiting only the data known after the Itala-2009 event.
KAUST Department:
Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division; Physical Sciences and Engineering (PSE) Division
Citation:
Cama M, Lombardo L, Conoscenti C, Rotigliano E (2017) Improving transferability strategies for debris flow susceptibility assessment: Application to the Saponara and Itala catchments (Messina, Italy). Geomorphology 288: 52–65. Available: http://dx.doi.org/10.1016/j.geomorph.2017.03.025.
Publisher:
Elsevier BV
Journal:
Geomorphology
Issue Date:
30-Mar-2017
DOI:
10.1016/j.geomorph.2017.03.025
Type:
Article
ISSN:
0169-555X
Sponsors:
All of the authors of this paper collaborated equally during all of the steps of the study (design, data acquisition and analysis, results discussion) as well as in the writing of the paper itself. The study was carried out in the framework of the SUFRA (SUscettibilità da FRAna in Sicily) project, coordinator Prof. E. Rotigliano. The authors also wish to thank Ms. Cassandra Funsten for editing the paper's English.
Additional Links:
http://www.sciencedirect.com/science/article/pii/S0169555X16307176
Appears in Collections:
Articles; Physical Sciences and Engineering (PSE) Division; Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division

Full metadata record

DC FieldValue Language
dc.contributor.authorCama, M.en
dc.contributor.authorLombardo, Luigien
dc.contributor.authorConoscenti, C.en
dc.contributor.authorRotigliano, E.en
dc.date.accessioned2017-05-31T11:23:09Z-
dc.date.available2017-05-31T11:23:09Z-
dc.date.issued2017-03-30en
dc.identifier.citationCama M, Lombardo L, Conoscenti C, Rotigliano E (2017) Improving transferability strategies for debris flow susceptibility assessment: Application to the Saponara and Itala catchments (Messina, Italy). Geomorphology 288: 52–65. Available: http://dx.doi.org/10.1016/j.geomorph.2017.03.025.en
dc.identifier.issn0169-555Xen
dc.identifier.doi10.1016/j.geomorph.2017.03.025en
dc.identifier.urihttp://hdl.handle.net/10754/623851-
dc.description.abstractDebris flows can be described as rapid gravity-induced mass movements controlled by topography that are usually triggered as a consequence of storm rainfalls. One of the problems when dealing with debris flow recognition is that the eroded surface is usually very shallow and it can be masked by vegetation or fast weathering as early as one-two years after a landslide has occurred. For this reason, even areas that are highly susceptible to debris flow might suffer of a lack of reliable landslide inventories. However, these inventories are necessary for susceptibility assessment. Model transferability, which is based on calibrating a susceptibility model in a training area in order to predict the distribution of debris flows in a target area, might provide an efficient solution to dealing with this limit. However, when applying a transferability procedure, a key point is the optimal selection of the predictors to be included for calibrating the model in the source area. In this paper, the issue of optimal factor selection is analysed by comparing the predictive performances obtained following three different factor selection criteria. The study includes: i) a test of the similarity between the source and the target areas; ii) the calibration of the susceptibility model in the (training) source area, using different criteria for the selection of the predictors; iii) the validation of the models, both at the source (self-validation, through random partition) and at the target (transferring, through spatial partition) areas. The debris flow susceptibility is evaluated here using binary logistic regression through a R-scripted based procedure.Two separate study areas were selected in the Messina province (southern Italy) in its Ionian (Itala catchment) and Tyrrhenian sides (Saponara catchment), each hit by a severe debris flow event (in 2009 and 2011, respectively).The investigation attested that the best fitting model in the calibration areas resulted poorly performing in predicting the landslides of the test target area. At the same time, the susceptibility models calibrated with an optimal set of covariates in the source area allowed us to produce a robust and accurate prediction image for the debris flows activated in the Saponara catchment in 2011, exploiting only the data known after the Itala-2009 event.en
dc.description.sponsorshipAll of the authors of this paper collaborated equally during all of the steps of the study (design, data acquisition and analysis, results discussion) as well as in the writing of the paper itself. The study was carried out in the framework of the SUFRA (SUscettibilità da FRAna in Sicily) project, coordinator Prof. E. Rotigliano. The authors also wish to thank Ms. Cassandra Funsten for editing the paper's English.en
dc.publisherElsevier BVen
dc.relation.urlhttp://www.sciencedirect.com/science/article/pii/S0169555X16307176en
dc.subjectLandslide susceptibility modellingen
dc.subjectTransferabilityen
dc.subjectMultiple debris flow eventsen
dc.subjectSaponara and Itala catchment (Sicily, Italy)en
dc.titleImproving transferability strategies for debris flow susceptibility assessment: Application to the Saponara and Itala catchments (Messina, Italy)en
dc.typeArticleen
dc.contributor.departmentComputer, Electrical and Mathematical Sciences and Engineering (CEMSE) Divisionen
dc.contributor.departmentPhysical Sciences and Engineering (PSE) Divisionen
dc.identifier.journalGeomorphologyen
dc.contributor.institutionDepartment of Geosciences, University of Tübingen, Rümelinstraße 19-23, 72070 Tübingen, Germanyen
dc.contributor.institutionDipartimento di Scienze della Terra e del Mare, Università degli Studi di Palermo, Via Archirafi 22, 90123 Palermo, Italyen
kaust.authorLombardo, Luigien
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.