Aftermath of Storm Daniel in Libya: Drivers, Implications, and the Need for A Rapid Assessment System (Model and Dataset)

We provide here in all necessary data to reproduce the results of: Aftermath of Storm Daniel in Libya: Drivers, Implications, and the Need for A Rapid Assessment System Authors: Mohamed Fawzy, Essam Heggy, György Szabó, and Arpad Barsi Abstract: The dire aftermath of Storm ‘Daniel’, which hit Derna City on the North-Eastern coast of Libya in the fall of 2023, resulted in the most devastating flood in Africa over the last century, with a human toll exceeding 11,300, and a significant part of the population was internally displaced. The Derna disaster is an omen of the alarming vulnerability of arid coastal areas to intensifying hydroclimatic extremes which remains poorly quantified. Furthermore, the absence of rapid damage assessment systems for storm aftermaths compromised rescues and relief efforts in impacted areas, which are best carried out within the first 48 hours of the disaster. To address those deficiencies, we review the hydroclimatic and anthropogenic drivers that exacerbated Derna’s flood impacts and hindered rapid assessment, resulting in elevated human casualties despite Libya being one of Africa’s most arid countries. We then show how the above limitations can be overcome using a convolutional neural network system that can be run locally with a limited number of very high-resolution multispectral images, providing rapid and reliable assessments of building and road damage to prioritize rescues within the first 48 hours. Our system results indicate that 4302 buildings in Derna have been affected by the floods, and 23.5% of them are severely damaged. In addition, we identified 26.55 km of damaged roads, representing ~46% of the transportation network. We demonstrate that anthropogenic drivers override natural ones in determining the magnitudes of storm-observed impacts. The findings of our rapid assessment convolutional neural network system are validated by published reports from global agencies, demonstrating a strong match. Considering the above, we provide near- and long-term recommendations to increase the resilience of coastal cities to the growing risks of hydroclimatic extremes, given inefficient urban planning and insufficient observational resources for monitoring natural hazards in arid areas.