Tropical Cyclones-related Impact on Human Displacement: Evaluating the Forecast Skills and Uncertainty of a Globally Consistent Impact Forecast System
OPEN ACCESS
Author / Producer
Date
2023-05-02
Publication Type
Master Thesis
ETH Bibliography
yes
Citations
Altmetric
OPEN ACCESS
Data
Rights / License
Abstract
Tropical cyclones (TCs) pose a significant threat to global populations, causing substan- tial damage and displacement each year. However, most disaster risk assessments only consider economic losses, ignoring displacement risk. Thus, it is crucial to develop a fore- casting system for TC-related displacement to help governments allocate resources for risk reduction and response. This thesis aimed to verify a displacement impact forecasting method on a global scale by forecasting displacement from past TC events (2017-2020) and comparing them to recorded data. A sensitivity analysis was conducted to identify input parameters that contribute most to output uncertainty. The study found that the system has a high level of accuracy, with a success rate of over 74% across all lead times considered, and can provide valuable information for decision-makers, particularly for severe events with high reported displacement. However, the system’s performance is limited by missed forecasts, which can underestimate potential displacement impact and result in inadequate preparation and response efforts. Wind-based impact functions were identified as the main cause of missed forecasts, highlighting the need for additional im- pact functions for each sub-peril involved in causing displacement, such as storm surges, rain-induced floods, and landslides. The results can help guide decision-making and improve preparedness and response measures for TC-related displacement. Further re- search and data collection efforts are necessary to develop impact functions for sub-perils and improve forecast performance. Future research may also analyse the uncertainty distribution obtained from the UNcertainty and SEnsitity QUAntification (UNSEQUA) module to gain a more thorough understanding of model output uncertainties.
Permanent link
Publication status
published
External links
Editor
Contributors
Book title
Journal / series
Volume
Pages / Article No.
Publisher
ETH Zurich
Event
Edition / version
Methods
Software
Geographic location
Date collected
Date created
Subject
Organisational unit
09576 - Bresch, David Niklaus / Bresch, David Niklaus