The modern B199 Brewer was installed in 2021 to supplement the Dobson spectrophotometer, whose Reykjavik record began in 1957. Across 745 days with usable direct-sun observations, ERA5 daily mean total column ozone agreed closely with B199: the root mean square error was 7.7 Dobson units, or 2.1%, and the Spearman correlation was 0.98. Almost 50% of comparisons fell within the Brewer instrument’s stated accuracy of plus or minus 1%, while approximately 90% were within plus or minus 3.5%. Comparisons with Brewer measurements at five other Arctic and subarctic sites produced correlations of at least 0.97 and root mean square errors of 2.1–3.6%, supporting the reanalysis for the study’s spatial analysis.

Agreement was not perfect. ERA5 generally preserved the temporal ranking of the Brewer observations but compressed their range, with the standard deviation of daily values about 4% lower than that measured by B199. The reanalysis tended to overestimate low ozone totals and underestimate high ones, consistent with smoothing of the gridded ozone field. At Reykjavik, the correlation between relative differences and measured ozone was −0.32, whereas its relationship with solar zenith angle was negligible at 0.02. Other Arctic Brewer records showed the same general tendency toward flatter extremes, although the effect on the relative ordering of observations was small. During polar night, comparisons with Focused Moon retrievals at Alert and Resolute also showed reasonable agreement, with correlations of at least 0.87 and root mean square errors no greater than 5.6%, though winter validation remained more difficult.

After the predictable annual and semiannual ozone cycles were removed with a harmonic decomposition, the researchers tested whether Reykjavik tracked the broader North Atlantic. Its deseasonalized series captured prominent features such as the exceptionally high spring ozone in 2024 and the lower spring values in 2025. Its correlation with the area-wide mean north of 40°N nevertheless varied substantially by month, ranging from 0.51 in April to below 0.15 in March, September and November. Local and regional fluctuations could therefore move Reykjavik far from the average: the absolute difference between the city and the regional mean reached plus or minus 155 Dobson units during the six-year comparison. The correlation was strongest in late spring and summer and weakest in autumn and early spring.

Correlation maps revealed a broadly circular zone of coherence centered on Iceland, covering much of the northeastern Atlantic between Greenland and Scandinavia. At individual ERA5 grid points, correlations above 0.90 occurred about 300–500 kilometers from Reykjavik, depending on the month. The radius of the broader area with correlations above 0.50 ranged from less than 900 kilometers in June to more than 1,500 kilometers in August. Across the 11 stations evaluated, high-correlation areas were generally largest from late winter into spring, when large-scale atmospheric dynamics exert stronger control on ozone distribution, and smallest during summer and parts of autumn. February showed especially broad coherence among the more northerly stations, while March also produced broad areas around Scandinavia.

Comparing Reykjavik with area-weighted means showed that its signal extended across a larger region than the individual-grid-point analysis alone implied. At a radius of 1,000 kilometers, correlations between Reykjavik and the regional ozone mean remained at or above 0.88 in every month except June and March. At 2,000 kilometers, they ranged from 0.63 to 0.86. Even at 2,500 kilometers, the correlation exceeded 0.50 in every month except March. The distance at which regional and local residual variability became equal ranged from about 1,350 kilometers in June to 2,350 kilometers in April. On this basis, the researchers judged the Reykjavik series representative of individual grid points over roughly 1,000 kilometers and capable of expressing major spatial patterns across an area with a radius of approximately 1,500 to more than 2,000 kilometers, depending on the month.

Empirical orthogonal function analysis provided a second test of the station’s regional value. The leading mode accounted for an average of 84% of ozone variance within 500 kilometers, 62% within 1,000 kilometers, 47% within 1,500 kilometers and 38% within 2,000 kilometers. Reykjavik’s series remained correlated with the leading mode beyond 2,500 kilometers in all months except March and November, when a sharp decline indicated some decoupling from the broader regional pattern. The similar behavior of direct area averages and the leading spatial mode indicated that the area-weighted ozone field was a strong proxy for regional variability. The analysis nevertheless distinguished this regional signal from the station’s own local signature, which could deviate substantially from surrounding values in particular months.

Comparisons with ten selected monitoring sites near or north of 64°N showed a consistent relationship between ozone variability and spatial representativeness. Across stations and months, the area with correlations above 0.90 was strongly related to the standard deviation of local ozone, with a correlation of 0.80. The corresponding relationships were 0.82 for areas above 0.70 and 0.79 for those above 0.50. These associations remained strong after the effect of interannual variability was reduced: the correlations were 0.74, 0.71 and 0.65, respectively. Locations and seasons with larger, more spatially coherent ozone variations therefore generally possessed wider zones of high correlation. The analysis associated this coherence with periods when large-scale dynamical processes, including planetary-wave activity, are important in controlling the distribution of total column ozone.

Ground-based comparisons supported the spatial results. Across 55 pairs of deseasonalized time series, correlations calculated from ERA5 agreed closely with those calculated directly from Brewer observations: the correlation between the two sets of coefficients was 0.98, and the ERA5-based value was higher by an average of only 0.02. The study has several important limits. It covers only six years, linear trends were not estimated because extreme endpoint years strongly influenced them, and influences such as stratospheric temperature, geopotential height and atmospheric circulation were not separately extracted. Validation during polar night was also limited because high solar zenith angles reduce direct-sun observations, so the representativeness results for those periods require caution. The study used a selected set of stations rather than an exhaustive inventory of Arctic monitoring sites.

Within those boundaries, the analysis identifies Reykjavik as a strategically important ozone-monitoring location. Its coverage fills a geographic gap between stations in Greenland and Scandinavia and provides an independent, high-quality ground-based series for studying northeastern North Atlantic ozone variability. The modern Brewer adds automated measurements, precision and stability to the older Dobson observations, while the long Reykjavik record provides continuity in a region where ground measurements remain sparse. The results do not mean that every local ozone anomaly can be inferred from the Reykjavik series, nor do they establish a recovery trend. Instead, they show that one well-maintained Arctic station can serve as a strong indicator of regional ozone patterns when local departures and seasonal differences are explicitly considered.

Subject of Research: Spatial representativeness of ground-based total column ozone monitoring in Reykjavik and the North Atlantic

Article Title: Spatial representativeness of total column ozone monitoring in Reykjavik, Iceland, and its regional significance in the North Atlantic region

Article References: Čížková, K., Láska, K., Tichopád, D., Staněk, M., von Löwis, S., Metelka, L., & Stráník, M. (2026). Spatial representativeness of total column ozone monitoring in Reykjavik, Iceland, and its regional significance in the North Atlantic region. Theoretical and Applied Climatology, 157(11), 745. https://doi.org/10.1007/s00704-026-06683-0

Image Credits: AI Generated

DOI: 10.1007/s00704-026-06683-0

Keywords: total column ozone, Reykjavik, North Atlantic, Brewer spectrophotometer, Dobson spectrophotometer, ERA5 reanalysis, Arctic atmosphere, spatial correlation, ozone monitoring, Iceland, Spatial, representativeness