ABSTRACT
The vertical distribution of relative humidity with respect to ice (RHI) in the Boreal wintertime Tropical Tropopause Layer (TTL, ≃14-18 km) over the Pacific is examined with the extensive dataset of measurements from the NASA Airborne Tropical TRopopause EXperiment (ATTREX). Multiple deployments of the Global Hawk during ATTREX provided hundreds of vertical profiles spanning the longitudinal extent of the Pacific with accurate measurements of temperature, pressure, water vapor concentration, ozone concentration, and cloud properties. We also compare the measured RHI distributions with results from a transport and microphysical model driven by meteorological analysis fields. Notable features in the distribution of RHI versus temperature and longitude include (1) the common occurrence of RHI values near ice saturation over the western Pacific in the lower-middle TTL (temperatures greater than 195 K); (2) low RHI values in the lower TTL over the central and eastern Pacific; (3) common occurrence of RHI values following a constant mixing ratio in the middle-to-upper TTL (temperatures between about 190 and 200 K), particularly for samples with ozone greater than about 50-100 ppbv indicating mixtures of tropospheric and stratospheric air; (4) RHI values typically near ice saturation in the coldest airmasses sampled (temperatures less than about 190 K); and (5) common occurrence of RHI values near 100% across the TTL temperature range in air parcels with low ozone mixing ratio (O3 < 50 ppbv) indicative of recent uplift by deep convection. We suggest that the typically saturated air in the lower TTL over the western Pacific is likely driven by a combination of the frequent occurrence of deep convection and the predominance of radiative heating (rising motion) in this region. The low relative humidities in the central/eastern Pacific lower TTL result from the lack of convective influence, the predominance of subsidence, and the relatively warm temperatures in the region. The nearly-constant water vapor mixing ratios in the middle-to-upper TTL likely result from the combination of slow ascent (resulting in long residence times) and wave driven temperature variability on a range of time scales (resulting in most air parcels having experienced low temperature and dehydration). The numerical simulations generally reproduce the observed RHI distribution features and sensitivity tests further emphasize the strong sensitivities of TTL relative humidity to convective input and vertical motions.
