Validation of ozone monitoring instrument nitrogen dioxide columns

We review the standard nitrogen dioxide (NO2) data product (Version 1.0.), which is based on measurements made in the spectral region 415-465 nm by the Ozone Monitoring Instrument (OMI) on the NASA Earth Observing System-Aura satellite. A number of ground- and aircraft-based measurements have been used to validate the data product's three principal quantities: stratospheric, tropospheric, and total NO2 column densities under nearly or completely cloud-free conditions. The validation of OMI NO2 is complicated by a number of factors, the greatest of which is that the OMI observations effectively average the NO2 over its field of view (minimum 340 km 2), while a ground- based instrument samples at a single point. The tropospheric NO2 field is often very inhomogeneous, varying significantly over tens to hundreds of meters, and ranges from <10(15) cm(-2) over remote, rural areas to >10(16) cm(-2) over urban and industrial areas. Because of OMI's areal averaging, when validation measurements are made near NO2 sources the OMI measurements are expected to underestimate the ground- based, and this is indeed seen. Further, we use several different instruments, both new and mature, which might give inconsistent NO2 amounts; the correlations between nearby instruments is 0.8-0.9. Finally, many of the validation data sets are quite small and span a very short length of time; this limits the statistical conclusions that can be drawn from them. Despite these factors, good agreement is generally seen between the OMI and ground- based measurements, with OMI stratospheric NO2 underestimated by about 14% and total and tropospheric columns underestimated by 15-30%. Typical correlations between OMI NO2 and ground- based measurements are generally >0.6.