Human sensor networks for natural disasters
File(s)
Author(s)
Lever, Jacob
Type
Thesis
Abstract
In the face of escalating climate-driven disasters, traditional forecasting and emergency response frameworks increasingly struggle to keep pace with the speed, complexity & societal impact of natural hazards such as wildfires. This thesis proposes a novel framework of Human Sensor Networks (HSNs) as a means of leveraging human-generated data from social media platforms to develop and augment models of disaster detection, forecasting, and response. The central premise is that during natural disasters, individuals act as noisy, distributed, multimodal sensors, whose subjective observations, expressed online, can be interpreted as real-time indicators of localised conditions. By treating the digital social environment as a noisy sensing network, this work abstracts human behaviour into a formalised remote sensing model and couples with machine learning methods to model, predict, & respond to disasters more efficiently.
Version
Open Access
Date Issued
2025-08-08
Date Awarded
2026-03-01
Copyright Statement
Attribution 4.0 International Licence (CC BY)
License URL
Advisor
Arcucci, Rossella
Sponsor
Leverhulme Trust
Grant Number
RC-2018-023
Publisher Department
Department of Earth Science & Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)