Rural Korean Rice Paddy Study Tracks Ozone Chemistry Without Direct Radical Measurements

At a rural monitoring station surrounded by rice paddies on the west coast of South Korea, ozone levels stayed remarkably steady over three years even as nitrogen dioxide declined. That mismatch between falling precursors and stable oxidants is exactly the kind of puzzle that makes surface ozone so difficult to manage, and a new open-access study in the Asian Journal of Atmospheric Environment tackles it with an unusual toolkit: a proxy framework built entirely from routinely measured gases and meteorology, because the site had no direct measurements of volatile organic compounds or atmospheric radicals.

The research team, led by Min-Wook Kim and Min-Suk Bae, analyzed continuous hourly observations from March 1, 2023, to December 31, 2025, at Buan, a rural agricultural site located at 35°46′21″ N, 126°40′05″ E. The station sits amid croplands with no major residential or industrial emission sources within a kilometer, the nearest residential area roughly 1.5 kilometers away and the western coastline about 5 kilometers distant. The authors characterize it as a rural receptor shaped primarily by regional atmospheric transport rather than direct urban or traffic emissions, which makes it a useful natural laboratory for separating local photochemistry from background influences.

The core measurement problem the study addresses is well known in atmospheric chemistry. Ozone is a secondary pollutant formed when peroxy radicals produced during the oxidation of VOCs and carbon monoxide convert nitric oxide to nitrogen dioxide, which then photolyzes to regenerate ozone. But the reaction of ozone with nitric oxide destroys ozone while creating nitrogen dioxide, so raw ozone readings are an incomplete indicator of the instantaneous photochemical oxidant burden. To get around this, the researchers relied on the odd-oxygen indicator Ox, defined as the sum of O3 and NO2, which has long been used to reduce the confounding effect of rapid nitric oxide titration.

Because direct radical measurements were unavailable, the team constructed a steady-state proxy framework. An OH proxy was scaled to ozone concentration, a solar-zenith-angle-based photolysis proxy, and water vapor estimated from temperature and relative humidity using the Tetens formulation. Peroxy radical production was represented by a bulk term partitioned into RO2 and HO2 fractions, and steady-state proxy concentrations of both were computed from their reactions with nitric oxide. From these, an instantaneous proxy ozone production rate, p(O3), was derived from the HO2-plus-NO and RO2-plus-NO reaction rates. A VOC oxidation proxy was defined as the product of the OH proxy and an effective VOC-attributed OH-reactivity term.

That effective reactivity term, kOHVOC, could not be calculated from species-specific VOC observations, so the authors treated it as an empirical monthly parameter constrained by the data itself. They defined a daily daytime oxidant enhancement, ΔOx, as the difference between the maximum and minimum observed Ox values within an 11:00 to 15:00 local-time window, then iteratively adjusted twelve monthly kOHVOC constants to minimize the root-mean-square error between the cumulative daily proxy p(O3) and the observed daily ΔOx. The authors are careful throughout to stress what this does and does not deliver: the resulting quantities are diagnostic indicators of relative photochemical activity, not absolute radical concentrations or ozone production rates.

The three-year record showed a clear divergence between precursors and oxidants. Annual mean nitrogen dioxide fell from 6.54 ± 4.76 ppb in 2023 to 5.02 ± 4.00 ppb in 2025, while annual mean ozone stayed within a narrow band, from 32.7 ± 18.3 ppb in 2024 to 35.9 ± 18.4 ppb in 2023. Mean nitric oxide remained low but variable, ranging from 1.52 ± 3.06 ppb to 1.56 ± 3.16 ppb, and Ox varied only modestly, from 38.4 ± 19.2 ppb in 2024 to 42.4 ± 19.2 ppb in 2023. In other words, oxidant levels were broadly maintained even as the nitrogen oxide burden declined.

Seasonality dominated the signal. Ozone and Ox showed pronounced cycles with elevated levels from late spring into summer and lower values in winter, while nitrogen dioxide peaked in colder periods, consistent with weaker photochemical removal and accumulation under shallower boundary layers. During March through June, mean ozone rose to 45.3 ± 17.7, 44.8 ± 16.9, and 46.2 ± 17.9 ppb in 2023, 2024, and 2025 respectively, with seasonal Ox near 48 to 50 ppb and nitrogen dioxide below its annual mean. The proxy indicators told the same seasonal story: the OH and VOC oxidation proxies, the HO2 and RO2 proxies, and proxy p(O3) all showed warm-season maxima and winter minima, pointing to stronger daytime photochemical activity under greater solar forcing and higher temperatures.

On the interannual scale, however, the photochemical signal was flat to slightly declining. The annual mean integrated daytime proxy p(O3) decreased from 7.06 × 10⁻⁷ ± 5.19 × 10⁻⁷ in 2023 to 6.01 × 10⁻⁷ ± 4.37 × 10⁻⁷ in 2025, and observed daytime ΔOx declined modestly from 9.92 ± 6.65 ppb to 9.09 ± 6.82 ppb over the same period. The authors interpret this to mean that interannual ozone variability at Buan likely reflected the combined influence of local daytime photochemistry, background oxidant levels, and seasonal meteorological variability rather than a strengthening of local production. A seasonal ratio analysis reinforced the point: although Ox was elevated in May and June, the normalized proxy p(O3)-to-Ox ratio was comparatively lower than in late summer, suggesting that late-spring oxidant enhancement drew on regional background oxidants and long-range transport as well as local production.

The most striking results came from conditional regime filtering. The team classified days by daytime mean temperature, relative humidity, and nitric oxide concentration over the 11:00 to 15:00 window, deliberately selecting regimes with simultaneously high or low proxy p(O3) and ΔOx rather than maximizing correlation. A high-photochemistry regime of 70 days, defined by temperatures of 30 °C or above, relative humidity at or below 60 percent, and nitric oxide at or below 1.6 ppb, contrasted sharply with a low-photochemistry regime of 51 days at or below 5 °C, at or above 50 percent humidity, and at or above 1.0 ppb nitric oxide. Under the warm, dry, low-NO conditions, mean integrated proxy p(O3) reached 1.34 × 10⁻⁶ ± 4.12 × 10⁻⁷ and mean ΔOx reached 12.58 ± 6.54 ppb, whereas the contrasting regime yielded only 1.09 × 10⁻⁷ ± 6.92 × 10⁻⁸ and 4.11 ± 2.27 ppb respectively.

Those contrasts indicate that local daytime oxidant enhancement at Buan was most clearly expressed when conditions favored radical production and efficient nitric-oxide-to-nitrogen-dioxide conversion, and the authors note that part of the regime separation may inherently reflect the framework-defined favorability index used to identify it. The broader takeaway is methodological as much as atmospheric: in rural environments where direct VOC and radical measurements are rarely available, an Ox-constrained proxy framework built from routine hourly monitoring can separate periods of enhanced and suppressed local photochemical activity and help explain why ozone can hold steady even as its precursors decline. The proxy-derived variables remain relative indicators rather than absolute concentrations or rates, but they offer a practical diagnostic lens on the nonlinear chemistry that governs oxidant variability in agricultural landscapes.

Subject of Research: A proxy-based analysis of ozone formation and oxidant variability at a rural rice paddy site in Buan, South Korea, from 2023 to 2025.

Article Title: Proxy-based analysis of ozone formation and oxidant variability under conditional regimes in a rural rice paddy region

Article References: Kim, M.-W., Jeon, H., Ko, D.-H., Kim, W., Park, C., Choe, S., Song, M., Kim, J.-H., & Bae, M.-S. (2026). Proxy-based analysis of ozone formation and oxidant variability under conditional regimes in a rural rice paddy region. Asian Journal of Atmospheric Environment, 20(1), Article 13. https://doi.org/10.1007/s44273-026-00092-5

Image Credits: AI Generated

DOI: 10.1007/s44273-026-00092-5

Keywords: ozone, oxidant variability, proxy framework, rice paddy, South Korea, photochemistry, NOx, VOC oxidation, regime filtering, rural monitoring, Ox indicator, radical proxies