Unveiling the HONO Offsetting Effect: Rethinking NO_<i>x</i> Emission Controls during Urban Ozone Pollution Episodes

Conventional ozone (O 3 ) control typically targets nitrogen oxides (NO x ) and volatile organic compounds (VOCs), yet the role of nitrous acid (HONO) is often overlooked. Here, machine learning (ML)-derived HONO–NO x reduction relationships in the real atmosphere are integrated into the process-based photochemical model (OBM–MCM) to diagnose the sensitivity of the O 3 –NO x –VOCs during high-pollution episodes in Shenzhen, China. OBM simulations constrained by observed HONO show a 95% increase in daytime net O 3 production rates [ P net (O 3 )] compared to the conventional unconstrained case through enhanced OH radical formation that accelerated VOC oxidation and HO 2 /RO 2 + NO pathways. Relative incremental reactivity (RIR) of HONO exhibits a strong anticorrelation with NO x ( R 2 = 0.86), indicating that a greater NO x -driven increase in the level of O 3 corresponds to a greater HONO-driven decrease in the level of O 3 . ML predicts that a 10% reduction in NO x synchronically results in reducing atmospheric HONO and TVOCs by ∼7.6 and ∼3%, respectively, leading to a shift in P net (O 3 ) from a maximum of 28% increase to a 14% decrease through the reshaping empirical kinetic modeling approach (EKMA), thereby demonstrating that HONO can offset the O 3 increase induced by NO x reduction. These findings challenge traditional EKMA frameworks that NO x control brings adverse effects under VOC-limited regimes, highlighting the feasibility of NO x control strategies when HONO responses are considered.