AJOPSIS technical note, 2026-10-29 — written by our own engineering team.
Between 1991 and 2000 a DOAS system at the ISPESL monitoring station in central Rome measured nitrogen dioxide, sulphur dioxide, ozone, benzene and toluene across an open path about 280 m long. Ten years at one site is unusual for optical monitoring, and the record is worth reading twice: once for the pollution trends, and once for what it shows about path-averaged measurement in a dense city.
The station sits in an area of heavy traffic and domestic heating. A transmitter with a high-pressure xenon lamp projects light to a receiver, which passes it through an optical fibre to a spectrograph and then to the management computer. Inorganic gases were recorded over 240–350 nm; the aromatics in the 250–290 nm range, where oxygen, ozone and sulphur dioxide are the main interfering absorbers. The system collected roughly 100 spectra per second at better than 0.05 nm resolution over 1000 channels, and produced one averaged result every seven minutes. Concentrations were derived from the absorption using the Lambert–Beer law in two steps: division by a stored system reference spectrum, then removal of broadband attenuation with a fifth-order polynomial, leaving the differential spectrum that is fitted against laboratory-generated differential cross-sections. Cross-sensitivity for benzene and toluene was checked against ozone, sulphur dioxide and nitrogen dioxide and stayed below 2% of the interferent concentration. Radon was measured with an OPSIS SM200 atmospheric stability monitor and used as a tracer of boundary layer mixing.
Benzene fell from an annual mean of about 25 µg/m³ in 1991 to about 8 µg/m³ in 2000, and toluene from about 108 to about 54 µg/m³ — a fall that follows the introduction of catalytic converters and the fuel limits in Italian law 413/97, which also required vapour recovery at filling stations. Italian benzene emissions are reported to have dropped by roughly 60% over the same period. Toluene and benzene tracked each other closely, with a toluene-to-benzene ratio near 5, consistent with the 3–5 range reported for urban areas and consistent with a common traffic source; the correlation between the two series was R² = 0.984. Sulphur dioxide fell from about 15 to about 4 µg/m³ as fuel sulphur limits took effect under European directives 93/12/EEC and 98/70/EC, with winter peaks from domestic heating dropping from 36 µg/m³ in January 1992 to 10 µg/m³ in January 2000. Short episodes were equally instructive: on the evening of 3 February 1997 a benzene hourly value reached about 80 µg/m³ under very stable conditions, then fell below 10 µg/m³ when mixing returned a few days later.
Ozone and nitrogen dioxide showed no clear annual trend. Neither is emitted directly in the quantities that matter; both are produced and consumed in the photochemical system formed by NOx, volatile organic compounds and sunlight, so their concentrations follow atmospheric conditions more than emission trends. Summer ozone was higher, as expected from photochemical activity.
The most useful part of the study is the comparison between the open-path DOAS and an offline GC-PID that sampled at a fixed point. The two tracked the same trend, with a correlation of R² = 0.984, but the regression was y = 0.427x − 0.047, the point analyser reading higher. The authors attribute the difference to sampling geometry rather than to either instrument being wrong: the DOAS value is an average over a 280 m path that includes streets, courtyards and gaps, while the GC-PID measures the air at one point, which in a street canyon carries the traffic signal more strongly.
Three lessons survive the decade. A path average is a better description of what people breathe over an area than any single point, but it is not interchangeable with a point value and should not be dropped into a limit comparison that assumes one. Long records are what separate a trend from a weather episode — a year of data in Rome would have suggested several contradictory conclusions. And a comparison between two instruments only means something when the sampling geometry is described: here the difference between the methods was itself the finding.
This note is based on the long-term study carried out at the ISPESL station in Rome with an OPSIS DOAS system between 1991 and 2000, as summarised in our Chinese article, and on the published work it refers to. Concentrations, ratios and regression values are quoted as reported there; they are not measurements made by AJOPSIS. For the Chinese version see 城市空气污染物的十年测量.
AJOPSIS technical note, written by our own engineering team, 2026-10-29. Chinese reader? See the Chinese version.
Send the gases, ranges, duct or path dimensions and required certification — we will respond with a technical proposal.