Monitoring urban resilience using fine‑resolution satellite nighttime light imagery

Urban resilience—the capacity of cities and their constituent sectors to absorb, adapt to, and recover from disruptive shocks—has become a central concern for sustainable urban development, particularly in response to the COVID-19 pandemic and the unprecedented lockdown measures implemented worldwide. Despite growing interest, empirical assessment of urban resilience remains challenging due to the complexity of urban systems, the heterogeneity of sectoral responses, and the limited availability of high-frequency, spatially detailed socioeconomic data suitable for global comparative analysis. This PhD thesis advances a satellite-based framework for assessing sector-level urban resilience using time-series nighttime light (NTL) imagery as a proxy for human and economic activity. Although NTL data provide consistent global coverage and high temporal frequency, their application to fine-scale urban analysis is constrained by coarse spatial resolution and blooming effects. To address these limitations, the thesis adapts and extends a hybrid geostatistical downscaling approach—Random Forest Area-to-Point Kriging (RFATPK)—to nighttime light data, explicitly accounting for spatial heterogeneity and the point spread function to enhance spatial detail and mitigate blooming. Validation against independent high-resolution nighttime imagery demonstrates that the downscaled NTL products outperform conventional coarse-resolution data, yielding on average a 17% improvement in socioeconomic representation and substantially enhanced spatial coherence. The refined NTL data exhibit stronger associations with independent indicators, including Gross National Income and development indices. Building on these advances, the thesis introduces a structured resilience framework for assessing urban resilience based on sector-specific NTL trajectories before, during, and after COVID-19 lockdowns, applied to 98 city–sector combinations across 48 global cities. The contribution of this research is both methodological and conceptual. Methodologically, it introduces a consistent and globally scalable approach for measuring sectoral urban economic resilience using satellite nighttime lights, overcoming key limitations of conventional economic statistics in capturing dynamic, intra-urban recovery processes. Conceptually, it demonstrates that economic essentiality acts as a primary organizing dimension of resilience outcomes: sectors linked to essential provisioning disproportionately exhibit resilient or fully recovered trajectories, whereas non-essential commercial and retail activities experience sustained decline, particularly in European cities. These findings reframe urban resilience as an outcome of economic structure and functional organization, rather than an inherent or uniformly distributed property of cities, with clear implications for urban policy and resilience planning.