CHECK OF OZONE DATA QUALITY

LA/CNRS carefully controls the ozone measurement quality : (1) periodical absolute calibration against a reference laboratory instrument before and after each flight period (about one year) ; (2) automatic in-flight calibrations at 3 levels ; and (3) check of inlet and line cleanness to ensure that ozone is not destroyed before entering the analyzer.

Flights comparisons between 2 MOZAIC aircraft are possible when aircraft are flying on the same routes or near the same airports, within less than 3 hour interval. They confirm the ozone data precision of ± (2 ppbv + 2%).

(click on image for enlargement)

Systematic inter-aircraft comparisons have been made on routes from Europe to North America, Asia and sometimes South America:

Most inter-comparisons showed an excellent agreement if aircraft are sufficiently close in time. Below is presented a case where 3 aircraft have followed the same route during nearly 5 hours and within 1.5 hour interval. Ozone concentrations measured by the 3 aircraft agreed within 2%.

(click on image for enlargement)

However, some inter-comparisons revealed the existence of small discrepancies which were not detected during the calibration procedure or by the inlet checks. This was therefore the case for the aircraft MSN 053 (Lufthansa) which was found systematically 10% lower than the other aircraft over the period 20/03/1996 to 25/05/1998. These 2 dates correspond to yearly maintenance operations on the system and the explanation proposed is the existence of a small leak in the Teflon connections, replaced on 20/03/1996, which has been supressed after the maintenance on 20/05/1998. Except this period, this aircraft always showed an excellent agreement with the other ones.

A 5 ppbv offset was also detected on aircraft MSN 049 (Air France) over the period 20/12/1996 to 05/02/1998. Here again, these dates corresponded to a yearly maintenance, but in this case no definitive explanation has been found (calibration, leak, loss ...).

In addition, an improved data filtration was applied in order to better remove sparse erratic ozone values (some tens of individual 4 seconds data per month). Lastly, in very few cases a wrong relationship was observed between barometric altitude and external pressure, although they were both recorded from the aircraft data computer.

The MOZAIC data base was thus updated, taking into account the corrections mentionned above.