Nighttime lights analysis reveals disparate Covid-19 impacts in Uganda's refugee settlements

The COVID-19 pandemic exposed structural inequalities worldwide, disproportionately affecting marginalized groups, including refugees. In Uganda, a nationwide lockdown from March to October 2020 restricted the movement of refugees living in managed settlements, despite national asylum policies emphasizing integration and mobility. These restrictions did not apply to host populations residing within the same settlements, raising questions about the differential impacts of COVID-19 policies. This study uses satellite remote sensing to assess how the lockdown affected refugee and host communities in Bidi Bidi and Rhino Camp, two of Uganda's largest refugee settlements. We analyze NASA Black Marble Visible Infrared Imaging Radiometer Suite (VIIRS) nighttime lights (NTL) time series data, normalized by building count, and apply Interrupted Time Series Analysis using ARIMA models to characterize shifts in nighttime radiance before, during, and after the lockdown. Median NTL density declined by 20.33% in Bidi Bidi refugee zones and 10.05% in Rhino Camp refugee zones during lockdown, with host communities experiencing similar but slightly smaller declines (15.39% and 9.55%, respectively). Refugee and host zones within the same settlement exhibited similar NTL trajectories, while dynamics diverged sharply between settlements (cross-settlement RMSE > 0.70), indicating that shared settlement context matters more than refugee versus host status in shaping lockdown impacts. Both refugee and host zones approached pre-lockdown NTL levels within 15 months, indicating recovery in aggregate nighttime radiance. These findings highlight the heterogeneous effects of public health interventions across refugee settings and demonstrate the value of building-normalized NTL analysis for evaluating crisis impacts in data-constrained humanitarian settings.