Publication Type:Journal Article
Source:Journal of Atmospheric Chemistry, Volume 6, Issue 3, p.281 - 298 (1988)
Keywords:Atmospheric Protection/Air Quality Control/Air Pollution, Climatology, effective sampling, Meteorology/Climatology, ozone lidar, Ozone network, ozonesonde, trends
An examination of typical tropospheric ozone variability on daily, monthly, annual and interannual timescales and instrumental precision indicates that the current ozonesonde network is insufficient to detect a trend in tropospheric ozone of ≤1% per year at the 2σ level even at stations with records a decade in length. From a trend prediction analysis we conclude that in order to detect a 1% per year trend in a decade or less it will be necessary to decrease the time between observations from its present value of 3–7 days to 1 day or less. The spatial distribution of the current ozonesonde stations is also inadequate for determining the global climatology of ozone. We present a quantitative theory taking into account photochemistry, surface deposition, and wind climatology to define the ‘effectively sampled region’ for an observing station which, used in conjunction with the instrumental precision and the above prediction analysis, forms the basis for defining a suitable global network for determining regional and global ozone climatology and trends. At least a doubling of the present number of stations is necessary, and the oceans, most of Asia, Africa, and South America are areas where more stations are most needed. Differential absorption lidar ozone instruments have the potential for far more frequent measurements of ozone vertical profiles and hence potentially more accurate climatology and trend determinations than feasible with ozonesondes but may produce a (fair weather) biased data set above the cloud base. A strategy for cloudy regions in which either each station utilizes both lidars and sondes or each station is in fact a ‘doublet’ comprised of a near-sea-level lidar and a proximal-mountain-top lidar could serve to minimize this bias.