Sanja Bernat Gazibara

Preliminary Investigations and Numerical Simulations of a Landslide Reactivation

Open image in new window A large landslide occurred near the Grohovo Village in outback of the City of Rijeka, Croatia, on 13 February 2014 after long term period of heavy rain. The preliminary surface observation was carried out immediately after sliding appearance enabled estimation of site condition, dimension of the landslide so as an assessment of a hazard of further landslide movements. The estimated dimension of landslide body are length of 350 m, width of 135 and 20–30 m of depth to the slip surface. The movement of approximately 12–15 m down the slope caused the complete damage of the local road over the landslide body. The toe of the landslide reached the bank of the Valici Reservoir 250 m away from the downstream located Valici Dam. Based on hazard assessment of further landslide movements, a lowering of the water level in reservoir and surface drainage from the landslide body were conducted as emergency landslide mitigation measures but the main reason for further landslide movements reduction was the end of heavy rain. Although the detailed filed investigations were not carried out, the analysis of possible further development of sliding was conducted. To establish the engineering geological model of the landslide, the analyses of existing LiDAR imagery and engineering geological mapping were carried out. It was identified that the recent landslide is a reactivation of the dormant landslide. Since a slip surface position was not identified by field investigation, a numerical analysis of a slope using the strength reduction method was introduced to determine the shape of a zone of rupture which would be used in landslide simulation. As a main hazard of further landslide movements, the filling of the Valici Reservoir, forming a landslide dam so as possible overflow of the dam and forming the wave that can reach the center of the City of Rijeka downstream the Rjecina River were identified. To determine possible scenarios those could be realized in case of new long term rainy period and raising of ground water level in the landslide body, a numerical simulations of further landslide development were conducted using LS-Rapid simulation software. In case of high reservoir water levels correspondent to the dam overflow, the sliding mass would significantly fill the reservoir and cause the landslide dam, while the water level rising and landslide caused waves (tsunamis) would overflow the Valici Dam and cause significant damage downstream the Rjecina River channel. Conducted LS-Rapid simulation results enabled a selection of relatively safe reservoir water level at which, in case of further landslide movements, no current harmful consequences would be realized. The reservoir filling by sliding mass would cause significant reduction of reservoir volume, disturbances in the Rjecina River and Valici Reservoir flow regimes so as long term disruption of hydro power plant work that imply on necessary landslide remediation before further landslide reactivation.

Identification and mapping of shallow landslides in the City of Zagreb (Croatia) using the LiDAR-based terrain model

Landslides in the hilly zone of Medvednica Mt. were identified visually using LiDAR DTM with a spatial resolution of 15 × 15 cm. Acquisition of the LiDAR data was performed in December 2013, following an extreme precipitation period that resulted in numerous landslides. Topographic derivative datasets for interpreting landslide morphology were computed from the LiDAR DTM: hillshade maps, degree of slope, contour lines, curvature, and surface roughness. Visual interpretation of LiDAR DTM derivatives was taken for the pilot area of 21 km2, which represents 12% of the hilly area in Zagreb City. This resulted in a landslide inventory map, indicating the contours of 676 landslides. Seventy-five percent of the landslide bodies showed a size between 159 and 2018 m2. The area of the smallest identified landslide in the test area is 43 m2. The majority of mapped landslides are located in a forested area. Each mapped landslide was assigned a level of confidence based on the LiDAR characteristics and it could be concluded that the LiDAR-based terrain model is a valuable tool for preparation of landslide inventories in heavily vegetated regions such as the hilly area of Medvednica Mt. The analysis of morphological properties of the landslides will be implemented to perform a semi-automated landslide mapping in the entire hilly area of the City of Zagreb (total area is 180 km2).

Landslides in the Dinarides and Pannonian Basin — from the largest historical and recent landslides in Croatia to catastrophic landslides caused by Cyclone Tamara (2014) in Bosnia and Herzegovina

Two neighboring countries in the southeastern region of Europe, Croatia, and Bosnia and Herzegovina (BIH), belong to the same geotectonic units of the Dinarides and to the Pannonian Basin, which influence relief types, lithology, and types of slope movements, i.e., landslides. The Dinarides are a mountain chain with a northwest-southeast direction that span from Slovenia through Croatia, BIH and Montenegro to Albania. The Pannonian Basin is situated within the Alpine, Carpathian, and Dinaric mountain belts at the boundary between Central and Southeastern Europe. The paper describes the general geological and geomorphological conditions in the Dinarides and the European Pannonian Basin in Croatia and BIH that are preparatory causal factors for landslides in the following environments: the hills of the Istrian Peninsula and Rječina River Valley; hills, low- and mid-altitude mountains in the Dinarides; and lowlands and hills in the Pannonian Basin. Landslide types, dimensions, and activities in the described areas are related to natural conditions primarily influenced by tectonic evolution and by recent anthropogenic processes, e.g., urbanization. More detailed descriptions are provided for the following selected phenomena, which are interesting because of the associated damage and potential risk: translational block landslide Brus and erosion phenomena on the Istrian Penninsula; relict and historical large, deep-seated landslides in the Rječina River Valley, including the recently reactivated Grohovo and Valići Landslides; catastrophic landslides triggered by precipitation during Cyclone Tamara in BIH (2014) (the Mačkovac-Šerići Landslide, Mjestova Ravan Landslide, Kosova Landslide, and Lukavica Landslide); and a large, deep-seated landslide in urban area of Zagreb, the Kostanjek Landslide. Recent rainfall triggering conditions of landslides in Croatia (2013) and BIH (2014) are also specified to emphasize the landslide risk and necessity of landslide risk management. The conclusions of the paper also note historical and potential damage due to landslide reactivations together with the spatial distribution of landslide-prone areas, which requires landslide mapping in the form of landslide inventory, susceptibility, hazard, and risk maps.

Prediction of the Kostanjek Landslide Movements Based on Monitoring Results Using Random Forests Technique

Prediction of landslide movements with practical application for landslide risk mitigation is a challenge for scientists. This study presents a methodology for prediction of landslide movements using random forests, a machine learning algorithm based on regression trees. The prediction method was established based on a time series data gathered by two years of monitoring on landslide movement, groundwater level and precipitation by the Kostanjek landslide monitoring system and nearby meteorological stations in Zagreb (Croatia). Because of complex relations between precipitations and groundwater levels, the process of landslide movement prediction is divided into two separate models: (1) model for prediction of groundwater levels from precipitation data; and (2) model for prediction of landslide movements from groundwater level data. In a groundwater level prediction model, 75 parameters were used as predictors, calculated from precipitation and evapotranspiration data. In the landslide movement prediction model, 10 parameters calculated from groundwater level data were used as predictors. Model validation was performed through the prediction of groundwater levels and prediction of landslide movements for the periods from 10 to 90 days. The validation results show the capability of the model to predict the evolution of daily displacements, from predicted variations of groundwater levels, for the period up to 30 days.

TXT-tool 2.385-1.1 A Comprehensive Landslide Monitoring System: The Kostanjek Landslide, Croatia

In this paper the general design of the integrated monitoring system of the Kostanjek landslide (Zagreb, Croatia) will be briefly presented. A comprehensive automated real-time monitoring system has been installed as a part of the research activities in the Croatian–Japanese SATREPS FY2008 scientific project on landslides in 2011–2013. The monitoring system consist of sensors for geodetic, hydrological and geotechnical monitoring to monitor landslide movement and landslide causal factors. Equipment for landslide monitoring at the surface and in the underground include 15 precise GNSS stations, long- and short-span extensometers, pore pressure gauges in boreholes, water level gauges in wells, rain gauge, weather station and accelerometers. All monitoring equipment is connected in the system entitled the Kostanjek Landslide Observatory. Monitoring of this deep and large landslide during multiple extreme hydro-meteorological events from January 2013 to January 2015 enabled collection of data for the analysis of landslide movement both on and below the surface as well as analysis of the relationship between landslide reactivations and their triggers, i.e., changes in groundwater level and precipitation. Monitoring results have proven established monitoring system as a sound tool for long-term monitoring as well as for landslide characterization. Preliminary application of landslide characterization (i.e., determination of landslide geometry) was presented in the form of numerical modeling to access changes on the landslide surface and run-out length for hazard and risk mapping purposes.

TXT-tool 4.385-1.1: Method for Prediction of Landslide Movements Based on Random Forests

Prediction of landslide movements with practical application for landslide risk mitigation is a challenge for scientists. This study presents a methodology for prediction of landslide movements using random forests, a machine learning algorithm based on regression trees. The prediction method was established based on a time series consisting of 2 years of data on landslide movement, groundwater level and precipitation gathered from the Kostanjek landslide monitoring system and nearby meteorological stations in Zagreb (Croatia). Because of complex relations between precipitations and groundwater levels, the process of landslide movement prediction is divided into two separate models: (1) model for prediction of groundwater levels from precipitation data; and (2) model for prediction of landslide movements from groundwater level data. In a groundwater level prediction model, 75 parameters were used as predictors, calculated from precipitation and evapotranspiration data. In the landslide movement prediction model, 10 parameters calculated from groundwater level data were used as predictors. Model validation was performed through the prediction of groundwater levels and prediction of landslide movements for the periods from 10 to 90 days. The validation results show the capability of the model to predict the evolution of daily displacements, from predicted variations of groundwater levels, for the period up to 30 days. Practical contributions of the developed method include the possibility of automated predictions, updated and improved on daily basis, which would be an important source of information for decisions related to crisis management in the case of risky landslide movements.

The Croatian-Japanese SATREPS Joint Research Project on Landslides (IPL-161)

The Croatian-Japanese joint research Science and Technology Research Partnership for Sustainable Development (SATREPS) project ‘Risk Identification and Land-Use Planning for Disaster Mitigation of Landslides and Floods in Croatia’ was performed from 2009 to 2014. Key objectives of the project were landslides and floods hazard analysis and the development of guidelines for use in urban planning. This project is also designated as on-going IPL project 161. The aims of the working groups dealing with landslides were to establish a methodology of comprehensive real time monitoring at two most important landslides in Croatia based on the results of previous investigations and new in situ and laboratory testing and behavior analysis; laboratory soil testing and numerical modelling of static and dynamic landslide behavior; development of landslide inventories using direct sensing and remote sensing techniques followed by the development of methodologies of landslide hazard analysis and zonation in three pilot areas in Croatia. In this paper we will present the most important achievements of working groups related to landslide studies at the project pilot areas: two in Primorsko-Goranska County (the Rjecina River Basin and the Dubracina River Basin) and one in the City of Zagreb (a hilly area of Medvednica Mt.). The identification and mapping of existing landslides in the hilly area of Medvednica Mt., Dubracina River Basin and Rjecina River Basin so as establishment and results of the monitoring systems installed on the Grohovo Landslide and the Kostanjek Landslide will be described.

Landslide Risk Reduction in Croatia: Scientific research in the framework of the WCoE 2014- 2017, IPL 173, IPL 184, ICL ABN

In this paper scientific activities of the Croatian Landslide Group (CLG), World Centre of Excellence on Landslide Risk Reduction (WCoE) of the International Consortium on Landslide (ICL) for the period 2014–2017, are shortly described. The results of scientific research are presented through the fields of landslide science: landslide identification and mapping, landslide investigation and testing, landslide monitoring, landslide modelling and landslide stabilization and remediation. It is concluded that the resulting landslide inventory maps, regional empirical rainfall intensity-duration thresholds, kinematic landslide models and soil strength parameters, landslide movement prediction models, numerical models and simulations and behavior of geotechnical construction for landslide stabilization provide necessary information for landslide risk management in Croatia. Besides applied scientific research, the general objectives of ICL WCoE are achieved in the framework of two Croatian IPL Projects and regional ICL Adriatic-Balkan Network.

Identification of Rock Fall Prone Areas on the Steep Slopes Above the Town of Omiš, Croatia

The aim of this paper was identification of rock fall prone areas above the historical town of Omis, located at the Adriatic coast in Croatia. Unstable areas were identified by kinematic analysis performed based on relative orientations of discontinuities and slope face. Input data was extracted from the surface model created from the high-resolution point cloud. The town of Omis is threatened by rock falls, because of its specific location just at the toe of Mt. Omiska Dinara. Rock fall risk is even higher due to rich cultural and historical heritage of the town. Collection of spatial data was performed by Time of Flight and phase-shift terrestrial laser scanners in order to derivate high resolution point cloud necessary for derivation of surface model. Split-FX software was used to extract discontinuity surfaces were semi-automatically from the point cloud data. Spatial kinematic analysis was performed for each triangle of TIN surface model of the investigated slopes to identify locations of possible instability mechanism. From the results of the spatial kinematic analysis, the most critical parts of the slope have identified for planar and wedge failure and flexural and block toppling. Verification of identified rock fall areas was performed by visual inspection of hazardous blocks at the surface model. Identified rock fall prone areas, unstable blocks and probable instability mechanisms on the steep slopes above the town Omis, present the input data for risk reduction by efficient design of countermeasures.