ASSESSING THE CORRELATION BETWEEN SATELLITE-DERIVED NIGHTTIME LIGHTS (NTL) AND RDNA4 INDICATORS OF DAMAGES, LOSSES, AND RECOVERY NEEDS ACROSS REGIONS OF UKRAINE

Abstract

In the context of the full-scale war in Ukraine, identifying trends in economic development becomes especially important when official statistics are unavailable or incomplete. The purpose of this article is to assess whether satellite-derived nighttime lights (NTL) can serve as an approximate indicator of regional economic disruptions by comparing NTL dynamics with oblast-level aggregate indicators reported in the World Bank’s Fourth Rapid Damage and Needs Assessment (RDNA4). The study’s methodology is based on using a non-conventional data source—the NASA VIIRS Black Marble product—to construct a monthly panel of nighttime radiance, followed by correlation analysis.

Within the study, a pre-war reference level is defined for each oblast (January 2021–February 2022). The post–full-scale invasion NTL gap index is defined as the average deviation starting from March 2022 (positive values are interpreted as a reduction in luminosity relative to the pre-war level, while negative values indicate an increase).

To quantify the relationship between the NTL-based index and RDNA4 indicators, Pearson and Spearman correlation coefficients are computed. RDNA4 distinguishes three categories of indicators: damages (physical destruction of assets), losses (foregone or under-received production, income, and service provision), and needs (costed requirements for recovery and reconstruction, incorporating the “build back better” principle).

The results indicate that the NTL index is most closely associated with RDNA4 losses (Pearson r≈0.71; Spearman ρ≈0.51), whereas correlations with damages (r≈0.23; ρ≈0.27) and needs (r≈0.31; ρ≈0.37) are weak. The divergence between Pearson and Spearman estimates, together with sensitivity analysis, suggests that the linear relationship for losses is partly shaped by specific regions (notably Kyivska and Dnipropetrovska oblasts), while several oblasts (e.g., Zaporizka, Khersonska, Luhanska) reduce concordance. This pattern may reflect heterogeneity in electricity outages, the effects of curfew restrictions, differences in sectoral structure, and spatial shifts of activity under wartime conditions. The practical significance of the study lies in substantiating the feasibility of using NTL as a timely complement to regional monitoring in data-scarce contexts; at the same time, the need for cautious interpretation is emphasized due to potential confounding factors and the imperfect correspondence among luminosity, economic activity, and recovery needs.