Comparison of upper tropospheric water vapor observations from the Microwave Limb Sounder and Atmospheric Infrared Sounder

We compare matched retrievals of upper tropospheric water vapor (UTWV) mixing ratios from the Microwave Limb Sounder (MLS) instrument on the Aura satellite, and the Atmospheric Infrared Sounder (AIRS) instrument on the Aqua satellite. Because each instrument's sampling is affected by tropical conditions, about half of mutually observed scenes in the tropics yield simultaneous successful retrievals from both systems. The fraction of mutually retrieved scenes drops to 30% at higher latitudes where clouds significantly inhibit AIRS sounding. Essentially all scenes observed by MLS in extratropical and polar regions yield successful retrievals. At 250 hPa in the tropics, measurements from the two instruments are highly correlated, the differences of their means ((Delta) over bar (q)) are smaller than 10%, and the standard deviations of their differences (sigma(q)) are 30% or less. At 300 hPa, MLS means are drier by 10-15%, and sigma(q) is 40-60%, indicating that responses of MLS and AIRS to UTWV perturbations are not one-to-one. Root mean square agreement is also poorer over the poles at 300 hPa and at 200 and 150 hPa at lower latitudes. In these regions, vertical bar(Delta) over bar (q)vertical bar = 10% or more, and sigma(q) = 40-70%. Correlations between the two data sets are 0.7-0.9 at 300 and 250 hPa globally and at 200 hPa in the tropics. This high correlation indicates that sigma(q) of 50% or greater comes mainly from systematic differences in sensitivity of the two instruments, especially for small and large UTWV amounts; larger values of sigma(q) are generally not due to large random errors from either instrument. An AIRS low-end sensitivity threshold of 15-20 ppmv leads to poorer agreement under the driest conditions. Disagreement at 300 hPa likely comes from overestimation by MLS for the wettest conditions of > 400 ppmv. While MLS is biased slightly dry overall at 300 hPa, it is biased wet in the wettest regions, particularly those associated with deep convection. These sensitivity differences explain nonunity slopes of linear fits to the two data sets. MLS everywhere has a greater dynamic range than AIRS, with larger maxima and smaller minima. Good agreement at 250 hPa suggests AIRS uncertainties of 25% up to the reported 250-200 hPa layer in the tropics and extratropics, consistent with previous comparisons with balloon- and aircraft-borne instruments. The agreement at 250 hPa also indicates that MLS is reliable from its reported 215-hPa level upward in altitude.