Wikimpacts: A new database of climate impacts inferred from Wikipedia mining with generative AI - Ni Li

Extreme weather events such as floods, heatwaves, wildfires, and hurricanes pose a substantial risk to many societies and frequently lead to enormous economic damages. As climate change continues, impacts of climate extremes escalate and are already widely observed across a range of societal and natural systems all over the world. A key aim of climate science is to understand (i) the meteorological drivers behind extreme events, and (ii) the link between extreme weather events and climate change. Databases on observed climate extremes and their impacts provide highly valuable information to study these links. The state of the art in this context is the ‘emergency events database’, which is a global natural hazards-relative impacts database containing essential data of over 22,000 disasters from 1900 to the present. However, this and other databases are incomplete and have large sampling biases in space and time, which hampers robust analyses of climate change impacts. In this project, we will construct a new database through mining data from Wikipedia and other online sources using a range of natural language processing algorithms and machine learning techniques. This database will include extreme weather events such as storms, droughts, extreme temperatures, wildfires, and floods. For each event, a date range, location, and set of socioeconomic impacts will be included. At minimum, this database will offer new insights on global climate impacts and provide a new data source for extreme events analysis. Ideally, analysis of the resulting data base will enable an unprecedented assessment of the degree to which recent anthropogenic climate change has fuelled enhanced economic damage and loss of life across the world.