Controls on tropical upper tropospheric humidity

Abstract

Water vapor is the dominant gaseous absorber of infrared radiation in Earth. Since small changes in the dry regions of the subtropical upper troposphere have a large impact in the Earth's energy budget, it is important to know the distribution of humidity and the processes controlling the distribution over this region. Here the variability of the tropical and subtropical upper tropospheric humidity and its control mechanisms are examined using observations, statistical models, and trajectory-based water vapor simulations.

First, we examine the controls on subtropical upper tropospheric humidity (UTH) using measurements from the Atmospheric Infrared Sounder (AIRS) satellite instrument together with meteorological analyses. There are significant zonal variations that are related to the regional variability in the processes that determine subtropical UTH. This analysis shows that Rossby wave breaking events that bring high potential vorticity air into the subtropics are the dominant cause in variability of relative humidity (RH) over the eastern Pacific and the Atlantic Ocean. In contrast, over the Indian Ocean and western Pacific, the variability of RH is closely linked to the location and strength of subtropical anticyclones associated with the Madden Julian Oscillation (MJO).

We also form a two-parameter statistical model of the distribution of tropical tropospheric RH. This model fits the observed probability density functions (PDFs) of the RH well, which vary between regions and with altitude. The two model parameters concisely characterize the variations in the PDFs and provide information on the processes controlling the RH distributions. The parameters from the model fits to the observations indicate that there is rapid, frequent moistening in the tropical convective region, and that there is slower, more regular moistening in the tropical non-convective region.

Finally, the controls on subtropical UTH are quantified using trajectory-based water vapor simulations. Comparisons of these simulations with observations shows good agreement for mean values at different altitudes, but are in less agreement with the RH PDFs, especially in the tropical convective region. Clustering analysis shows that the dominant trajectories patterns differ between convective and non-convective regions. Furthermore, the variability of RH can also be partially explained by the mean RH of the trajectory clusters. Coherent variations of clusters with longitude are also consistent with the transient intraseasonal convection.