Data Analytics Applications for City Resilience Under Climate-Induced Hazards

Increased severity and frequency of climate-induced disasters (CID) is affecting the resilience of cities worldwide. Canadian insurers are facing natural disaster-induced claims of approximately $1 billion annually, whereas it was closer to $400 million in the previous decade. Moreover, annual national liabilities of the Disaster Financial Assistance Arrangements have increased from $100 million annually in late 1990s to $500 million in 2009–2010 and reached approximately $2 billion in 2013–2014. To maintain their basic functions, Canadian cities have to maximize the resilience of their critical infrastructure systems under CID. In their previous work, the authors proposed a framework that focused on historical power outage data to quantify city resilience. The framework showed that CID are the main cause of power outages in North America. Subsequently, the aim of this work is to assess previous CID in an attempt to enhance the resilience of the power system as one of the most critical infrastructure systems in Ontario. The first part of this work involves employing descriptive data analysis to derive meaningful information from a CID database. Following that, an unsupervised machine learning technique will be employed to assess CID probability in Ontario on the spatial level. Finally, the resilience of hydroelectric power generators across Ontario will be deliberated based on the spatial analysis performed. This work is considered a step in CID assessment and prediction, based on historical hazard data, global climate models, and climate change measures, in an attempt to maximize city resilience and mitigate CID-induced risks on cities.