Direct-sun versus sky-scan Pandora formaldehyde retrievals: implications for satellite validation and sampling representativeness in Tropical Southeast Asia

Ground-based Pandora spectrometers are widely used for validating satellite formaldehyde (HCHO) retrievals, yet the roles of retrieval geometry and sampling representativeness remain poorly constrained in tropical environments. This study evaluates Pandora Level-2 HCHO columns from five Southeast Asian stations (2021–2024), distinguishing between direct-sun (DS) and sky-scan (SS) observations and comparing them with satellite products from Ozone Monitoring Instrument (OMI), TROPOspheric Monitoring Instrument (TROPOMI), and Geostationary Environment Monitoring Spectrometer (GEMS) following uncertainty-based quality control. DS and SS retrievals show strong internal consistency under high-quality conditions ( r  > 0.7). Besides, DS generally exhibits greater variability and, at several sites, stronger sensitivity to localized variability, whereas SS often yields lower RMSE/MAE and more spatially representative agreement with satellite observations. Satellite comparisons reveal a clear performance hierarchy: OMI shows weak correlations ( r  < 0.4) and large errors, TROPOMI improves agreement ( r  ≈ 0.3–0.5), and GEMS further enhances performance in urban environments ( r  ≈ 0.58–0.65 at Bangkok) with reduced error magnitudes due to higher temporal sampling. However, discrepancies persist even under near-synchronous conditions, indicating that improved temporal resolution – within current satellite capabilities – does not fully resolve satellite–ground differences. These residual differences are consistent with sampling scale mismatches between localized Pandora measurements and spatially averaged satellite footprints. Overall, the results demonstrate that satellite validation in tropical regions is governed by the combined effects of retrieval geometry, spatial sampling, and temporal resolution, providing a framework for interpreting satellite–ground HCHO comparisons and guiding future validation strategies.