Decoupling Noise and Revealing Temperature Thresholds to Strengthen PM2.5–NO ₂ Predictive Relationships: A Case Study in Nong Khaem District, Bangkok, Thailand

ABSTRACT Urban PM2.5–NO x relationships are often masked by regime mixing. Focusing on Nong Khaem, Bangkok, this study decouples noise via a de‐noising and stratification protocol that removes sentinel/nonphysical values and conditions on season, diurnal period, wind speed (WS), relative humidity (RH), temperature tertiles, and circular wind direction (WD). In winter nights with low wind and moist air, PM2.5 co‐varied most strongly with NO 2 (moderate, precise correlations), while PM2.5–NO linkages were weaker or negative—consistent with nocturnal nitrate formation (NO 3 /N 2 O 5 → HNO 3 ) and NH 4 NO 3 partitioning under shallow mixing and elevated aerosol liquid water (ALW). Within this constrained window, an operational temperature threshold near 26–28°C maximized PM2.5–NO 2 coupling and then plateaued, matching thermodynamics that favor particle‐phase nitrate at cooler temperatures. RH acted as a graded modulator: low RH throttled heterogeneous/aqueous pathways; high RH enhanced N 2 O 5 hydrolysis and partitioning, with wet scavenging a possible limiter at very high RH. Replication in Sai Mai and Thawi Watthana showed robustness in direction but heterogeneity in magnitude, including NO‐driven ozone titration under very humid, low‐O 3 conditions in Sai Mai. Findings support a policy‐relevant screen—winter night + low wind (≤2 m/s) + moist air (≈ ≥80%–90% RH) + a district‐specific temperature guardrail (∼26–28°C)—to prioritize NO 2 ‐based inference of PM2.5 and guide nighttime NO x /NH 3 controls where PM mass monitoring is sparse.