Naslov
Relationship between atmospheric conditions and groundwater level on Grohovo landslide

Autori
Volf, Goran ; Žic, Elvis ; Ožanić, Nevenka

Vrsta, podvrsta i kategorija rada
Sažeci sa skupova, sažetak, znanstveni

Izvornik
4th Workshop of the Japanese-Croatian Project on "Risk Identification and Land-Use Planning for Disaster Mitigation of Landslides and Floods in Croatia", Book of abstracts / Vlastelica, Goran ; Andrić, Ivo ; Salvezani, Daša - Split : Fakultet građevinarstva, arhitekture i geodezije Sveučilišta u Splitu, 2013, 46-46

ISBN
978-953-6116-46-1

Skup
4th Workshop of the Japanese-Croatian Project on "Risk Identification and Land-Use Planning for Disaster Mitigation of Landslides and Floods in Croatia"

Mjesto i datum
Split, Hrvatska, 12.12.2013. - 14.12.2013

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
Grohovo landslide ; Rječina valley ; machine learning ; atmospheric conditions ; groundwater level

Sažetak
Grohovo landslide situated on the north-eastern slope of the Rječina valley is the largest active landslide along the Croatian part of the northern Adriatic coast. To contribute to understanding the effect of atmospheric conditions on ground water level fluctuations on Grohovo landslide a machine learning tool for induction of models in form of set of rules was applied on a data set comprising daily atmospheric and groundwater level data measured in 2012. Atmospheric data comprise average daily air temperature, humidity, wind speed, pressure, total evapotranspiration, precipitations and sum of 5, 10, 15, 20, 25, 30, 35, 40 and 45 day precipitations. For the experiment atmospheric data were used as independent variables from which target variable ; groundwater level is modelled. Rule-based regression models for numeric prediction are interpreted as a set of if-then rules where each rule is associated with a multivariate linear model. A rule indicates that, whenever a case satisfies all the conditions, the linear model is appropriate for predicting the value of the target attribute. The algorithms for rule induction mostly represent different variations of the M5 algorithm. The algorithm implemented in a software package Cubist was applied for modelling, in which the basic M5 algorithm was enhanced by combining the model-based and instance-based learning. The model describing groundwater level fluctuations consist of ten rules and have very high correlation coefficient of 0, 99. Results of measured and modelled data of groundwater level are presented on Figure 1. Of all atmospheric parameters (independent variables) presented above the model mostly used for rule induction 45 day precipitations (53%), air temperature (52%) and 35 day precipitations (33%). In equations which describes target variable are mostly used 45 day precipitations (82%), 10 day precipitations (77%), air temperature (74%), 20 day precipitation (72%) and 35 day precipitation (55%). From the given model it can be concluded that the most influence on groundwater level fluctuations have sum of daily precipitations and air temperature.

Izvorni jezik
Engleski

Znanstvena područja
Geologija, Građevinarstvo

POVEZANOST RADA