Determination of probabilities for the generation of high-discharge flows in the middle basin of Elqui River, Chile
Author
dc.contributor.author
Vergara Dal Pont, Ivan
Author
dc.contributor.author
Santibañez Ossa, Fernanda
Author
dc.contributor.author
Araneo, Diego
Author
dc.contributor.author
Ferrando Acuña, Francisco
Author
dc.contributor.author
Moreiras, Stella
Admission date
dc.date.accessioned
2018-11-14T20:39:05Z
Available date
dc.date.available
2018-11-14T20:39:05Z
Publication date
dc.date.issued
2018-08
Cita de ítem
dc.identifier.citation
NatHazards (2018) 93:531–546
es_ES
Identifier
dc.identifier.other
10.1007/s11069-018-3313-0
Identifier
dc.identifier.uri
https://repositorio.uchile.cl/handle/2250/152600
Abstract
dc.description.abstract
The probabilities for the generation of hyperconcentrated flows, and debris and mud flows in the middle basin of Elqui River (Chile) are determined. The objective was achieved collecting, for a period of 14 years, the precipitation events generating high-discharge flows, as well as the larger precipitation events that did not generate this process. For each of these events, data of peak 1-h storm precipitation, temperature (representing the zero-isotherm altitude) and antecedent precipitation of 1, 5 and 10 days were collected from three meteorological stations. Initially, an ordinal logistic regression model for each antecedent precipitation was fitted, but all were discarded due to the low significance of these variables in the generation of the models. This result allowed to infer that the high-discharge flows of the region are generated mainly by runoff and not by deep-seated or shallow landslides. Subsequently, a new model with the remaining variables was performed, which was statistically validated. From this, it was considered prudent to take as thresholds for the occurrence of hyperconcentrated flows, and debris and mud flows, their respective probabilities of 50%. For these thresholds, the model had an efficiency in the prediction of high-discharge flows of 90%. Finally, the partial correlation coefficients of each significant predictor variable with respect to the dependent were calculated, establishing that the temperature has greater influence than the peak 1-h storm precipitation.
es_ES
Patrocinador
dc.description.sponsorship
This research was supported by the Natural Hazards of the Central Andes project from the National University of Cuyo. We are grateful to Simon Higginson for reviewing the English and to Mauricio Vergara for the critical review.