Spatiotemporal dynamics of nighttime-light responses to successive policy shocks in a data-scarce coastal city: Multi-node VIIRS time-series analysis of Sihanoukville, Cambodia (2014–2026)

This study presents a multi-node nighttime-light (NTL) time-series analysis to monitor spatially heterogeneous urban economic responses to successive policy shocks in Sihanoukville, Cambodia, from January 2014 to March 2026. Three spatially fixed regions of interest (ROIs) are defined: the Sihanoukville Special Economic Zone (SSEZ) as a formal industrial node (9.30 km2), the Golden Lions (GL) roundabout area as a dense commercial service node (2.94 km2), and Preah Sihanouk Province as the macro-scale administrative baseline (1,605 km2). Monthly VIIRS/DNB VCMSLCFG composites (500 m, n = 147 months) are processed using Seasonal–Trend decomposition via Loess (STL, period = 12, robust = True) to isolate luminosity trends from monsoon-driven seasonality. Normalised indices (2017 = 100) and year-on-year (YoY) growth rates are compared across four policy phases. Sentinel-2-derived NDBI and NDVI cross-checks at GL validate that luminosity changes are more consistent with activity intensification than horizontal built-up expansion. Results reveal strong spatial heterogeneity: the GL index reached 1,185.9 by March 2026 (approximately twelve times the 2017 baseline), the SSEZ index reached 265.9 (approximately 2.7×), and the provincial index reached 418.2 (approximately 4.2×). Province-level aggregation conceals node-level divergence of nearly 900 index points in the post-sanctions window. The framework demonstrates how fixed-ROI VIIRS monitoring combined with STL decomposition can support near-real-time spatial diagnosis of urban economic change where conventional statistics are delayed or spatially coarse.