Open AccessQuantifying the damage susceptibility to extreme events of mountain stream check dams using Rough Set Analysis
Author(s) -
Mazzorana B.,
TrenkwalderPlatzer H.,
Heiser M.,
Hübl J.
Publication year - 2018
Publication title -
journal of flood risk management
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.049
H-Index - 36
ISSN - 1753-318X
DOI - 10.1111/jfr3.12333
Subject(s) - masonry , rough set , event (particle physics) , computer science , set (abstract data type) , residual , hazard , data mining , identification (biology) , structural engineering , engineering , algorithm , programming language , chemistry , physics , botany , organic chemistry , quantum mechanics , biology
In this work, a Rough Set Analysis‐based approach is proposed to quantify the damage susceptibility of check dams through specific indexes, which all require expert judgment to be quantified. The indexes are the Post‐Event Damage Condition ( DamPost ), the induced Condition Change ( DamCh ), the Residual Condition ( RC ), and the Post‐Event Functionality ( FPost ). Preliminarily, an existing data set, containing a quantification of the damage indexes, the associated characterisation of the flow process type of three torrential hazard events occurred in South Tyrol (Italy), the identification of construction material and the determination of age of the structure, was statistically analysed. To predict the damage indexes based on Rough Set Analysis, a general model, which considered all check dams regardless of their construction material, two specific models, for concrete and for masonry structures respectively, and a simplified version of the general model were set up. The derived rule bases exhibited satisfactory prediction accuracies only when the post‐event functionality, FPost , was chosen. Prediction accuracies were 68% for the general model, 79% for the material category concrete, 60% for the material category masonry, and 86% could be obtained by simplifying the decision attribute to a binary form (functionality given or not).