When a disaster emerges,timely acquisition of information is crucial for a rapid situation assessment.Although automation in the stan-dard satellite-based emergency mapping workflow has been advanced,delays still occu...When a disaster emerges,timely acquisition of information is crucial for a rapid situation assessment.Although automation in the stan-dard satellite-based emergency mapping workflow has been advanced,delays still occur at crucial steps.In order to speed up the provision of satellite-based crisis products to emergency man-agers,this paper proposes a geo-social media-based approach that detects disaster events based on the spatio-temporal analysis of georeferenced,disaster-related Tweets.The proposed methodol-ogy is validated on the basis of two use cases:wildfires in Chile and British Columbia.The results show the general ability of Twitter to forecast events several days in advance,at least for the Chile use case.However,there are large spatial differences,as there is a correlation between population density and the reliability of Twitter data.Consequently,only few meaningful alerts could be generated for British Columbia,an area with very low population numbers.展开更多
基金funding from the European Commission-European Union under HORIZON EUROPE(HORIZON Research and Innovation Actions)under grant agreement 101093003(HORIZON-CL4-2022-DATA-01-01)。
文摘When a disaster emerges,timely acquisition of information is crucial for a rapid situation assessment.Although automation in the stan-dard satellite-based emergency mapping workflow has been advanced,delays still occur at crucial steps.In order to speed up the provision of satellite-based crisis products to emergency man-agers,this paper proposes a geo-social media-based approach that detects disaster events based on the spatio-temporal analysis of georeferenced,disaster-related Tweets.The proposed methodol-ogy is validated on the basis of two use cases:wildfires in Chile and British Columbia.The results show the general ability of Twitter to forecast events several days in advance,at least for the Chile use case.However,there are large spatial differences,as there is a correlation between population density and the reliability of Twitter data.Consequently,only few meaningful alerts could be generated for British Columbia,an area with very low population numbers.
