University of Twente Student Theses

Login

Analyzing open source, python-based, tools for multi-hazard risk assessment

Girma, Frehiwot Amha (2021) Analyzing open source, python-based, tools for multi-hazard risk assessment.

[img] PDF
10MB
Abstract:Multi-hazard risk assessment is crucial for risk reduction planning by decision makers such as emergency managers and planners. The demand for multi-hazard risk assessment information is increasing due to the expected trend of more frequent disasters, climate change, growth of (urban) population, and increased inequalities. In most cases, risk assessment is still conducted for single hazards using hazard specific models and risk assessment tools. The available tools are very data demanding, have a poor data interoperability, and lacking considering the changing risk and hazard interaction. Proprietary tools for multi-hazard risk assessment are not available for the authorities and research community, and exploration of Open-source tools is very important. The aim of this study is to compare available Open source and Python-based tools (i.e., CLIMADA and RiskChanges) and validate the loss estimation for a documented disaster event: the building losses in Dominica resulting from the 2017 hurricane Maria documented in the Post Disaster Needs assessment report (PDNA). Multi-hazard risk assessment using CLIMADA and RiskChanges requires different formats of input data. For this research, the input data collected consisted of multi-hazard data (flood, landslide, and debris flow hazard data for the 2017 hurricane Maria made through OpenLISEM modelling and: wind hazard maps from IBTrACS), building data collected from OpenStreetMap (OSM) and vulnerability functions (for flood and wind). The OSM building data were classified based on the general occupancy type, construction type, roof type, and roof shape to select representative vulnerability functions of the buildings. The replacement value of buildings was estimated using real estate prices by considering the area of the buildings. CLIMADA uses the hazard intensity within a point location to estimate the loss whereas the RiskChanges calculates the maximum hazard value per building, and also subdivides it into spatial units based on their different hazard levels. To compare the estimated loss of the CLIMADA and RiskChanges with the losses reported by PDNA of Hurricane Maria the loss was categorized into four categories. The result shows that the total building losses for the 2017 Maria Hurricane calculated were both in line with those in the PDNA report. In addition to the loss estimation, the study also compared the capacity of the tools based on five criteria (i.e., data requirement, integration of the hazard interaction, risk calculation component, decision making support capability, and ease of use). Both CLIMADA and RiskChanges have the capacity for data interoperability, but the input data should be prepared based on the data requirements of the respective tools. Both tools do not fully incorporate the uncertainty management in the loss and risk assessment. In addition to this, both tools have difficulty to objectively express the spatial probability of the hazard, and the values should be estimated by expert opinion, whereas this component has a high impact on the loss results. As can be expected, the quality of the results in the tools completely depends on the quality of the input data. The study identified the two most important features that can improve the functionality of the tools. The first one is to represent the spatial variability of the spatial probability, instead of a single value for the entire area. The second aspect is that the tools should incorporate the uncertainty of all risk components into the risk assessment. However, both components would require more detailed input data, which is often not available. Whereas RiskChanges incorporates the hazard interactions into the overall loss assessment, this is not the case in CLIMADA. RiskChanges also need to integrate valuable features of CLIMADA which is the ability in accessing open-source data and visualization capacity in the Python-based version apart from its Graphical User Interface.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:https://purl.utwente.nl/essays/88723
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page