Human Mobility Networks Manifest Dissimilar Resilience Characteristics at Macroscopic, Substructure, and Microscopic Scales

Abstract

Human mobility networks can reveal insights into resilience phenomena, such as population response to, impacts on, and recovery from crises. The majority of human mobility network resilience characterizations, however, focus mainly on macroscopic network properties; little is known about variation in measured resilience characteristics (i.e., the extent of impact and recovery duration) across macroscopic, substructure (motif), and microscopic mobility scales. To address this gap, in this study, we examine the human mobility network in eight parishes in Louisiana (USA) impacted by the 2021 Hurricane Ida. We constructed human mobility networks using location-based data and examined three sets of measures: (1) macroscopic measures, such as network density, giant component size, and modularity; (2) substructure measures, such motif distribution; and (3) microscopic mobility measures, such as the radius of gyration and average travel distance. To determine the extent of impact and duration of recovery, for each measure, we established the baseline values and examined the fluctuation of measures during the perturbation caused by Hurricane Ida. The results reveal the variation of impact extent and recovery duration obtained from different sets of measures at different scales. Macroscopic measures, such as giant components, tend to recover more quickly than substructure and microscopic measures. In fact, microscopic measures tend to recover more slowly than measures in other scales. These findings suggest that resilience characteristics in human mobility networks are scale-variant, and thus, a single measure at a particular scale may not be representative of the perturbation impacts and recovery duration in the network as a whole. These results spotlight the need to use measures at different scales to properly characterize resilience in human mobility networks.

Beyond Resilience Triangle: Dissecting Resilience Curve Archetypes and Properties in Human Systems Facing Weather Hazards

“…This data is gathered when users voluntarily opt in to location services provided by partner apps. Spectus captures data from nearly 20% of the U.S. population, representing around one in (Hsu, Ho, & Mostafavi, 2022;Hsu, Liu, et al, 2022) to be representative in terms of capturing human mobility and travel mode detection due to its high spatiotemporal resolution. To safeguard user privacy, Spectus de-identifies the collected data and applies additional privacy measures, including obscuring home locations at the census-block group level and removing sensitive points of interest.…”

Unraveling Extreme Weather Impacts on Air Transportation and Passenger Delays using Location-based Data

“…Such datasets can provide high spatiotemporal resolution insights regarding population activity and mobility that can be used to understand how people move, act, and respond in hazard scenarios (Wang et al, 2020;Liu et al, 2022;. By evaluating the changes in population activity and mobility patterns, the extent of impacts of hazard-induced disruptions can be effectively evaluated (Hsu et al, 2022;Wang and Taylor, 2016;Liu et al, 2019). While location-based data has been used for evaluating hazard-induced perturbations in different infrastructures , the use of these high-resolution data has been rather limited in examining the vulnerability of air transportation systems to extreme weather events.…”