Cumulus Moistening, The Diurnal Cycle, and Large-Scale Tropical Dynamics

May 08, 2015

James Ruppert

Committee: Dick Johnson (advisor), Eric Maloney, Sue van den Heever, Chandra Venkatachalam

Abstract

Observations and modeling techniques are employed to diagnose the importance of the diurnal cycle in large-scale tropical climate. In the first part of the study, soundings, radar, and surface flux measurements collected in the Indian Ocean DYNAMO experiment (Dynamics of the Madden-Julian Oscillation, or MJO) are employed to study MJO convective onset. According to these observations, MJO onset takes place as follows: moistening of the low-midtroposphere is accomplished by cumuliform clouds that deepen as the drying by large-scale subsidence and horizontal advection simultaneously wane. This relaxing of subsidence is tied to decreasing column radiative cooling, which links back to the evolving cloud population. A new finding from these observations is the high degree to which the diurnal cycle linked to air-sea and radiative fluxes invigorates clouds and drives column moistening each day. This diurnally modulated cloud field exhibits pronounced mesoscale organization in the form of open cells and horizontal convective rolls. Based on these findings, it is hypothesized that the diurnal cycle and mesoscale cloud organization represent two manners in which local convective processes promote more vigorous day-to-day tropospheric moistening than would otherwise occur.

A suite of model tests are carried out in the second part of the study to 1) test the hypothesis that the diurnal cycle is an important moistening agent on longer timescales, and 2) better understand the relative roles of SST and radiation in the diurnal cycle of convection. Moist convection is explicitly represented in the model, the diurnal cycle of SST is prescribed, and cloud-interactive radiation is simulated with a diurnal cycle in shortwave heating. The large-scale dynamics are parameterized using the spectral weak temperature gradient (WTG) technique recently introduced by Herman and Raymond, wherein external (i.e., "large-scale") vertical motion is diagnosed based on the internal diabatic heating in the model, which is then used to advect model temperature and humidity. This vertical motion opposes domain-averaged temperature anomalies via adiabatic warming and cooling, and strikes a feedback between the model diabatic heating and the large-scale column moisture source associated with large-scale vertical motion. With a control simulation that successfully replicates a regime of shallow convection similar to nature, it is found through sensitivity tests that the diurnal cycle in radiative heating is the dominant driver of both diurnal column moisture variations and nocturnal rainfall in this regime, the latter of which agrees with previous findings by Randall et al. The diurnal cycle in SST and surface fluxes, in turn, is the main driver of the daytime convective regime, which is distinct from the nocturnal regime by its rooting in the boundary layer.

A simulation in which the diurnal cycle is stretched to 48 h amplifies an important nonlinear feedback at work in the diurnal cycle, which owes to the high-amplitude diurnal cycle in column relative humidity RH. This diurnal cycle in RH limits the amount of evaporation, and hence evaporative cooling, that takes place in the cloud layer. By throttling down the diabatic cooling, the diurnal cycle throttles down the daily-mean moisture sink driven by large-scale subsidence, such that the environment drifts toward a more moist state, all else being equal. When the diurnal cycle is not present, this nonlinear moisture source is weaker, and the environment drier. This feedback rectifies diurnal moistening onto longer timescales, thereby linking the diurnal cycle to longer timescales. These findings suggest that 1) the diurnal cycle of moist convection, as observed in DYNAMO, cannot be ruled out as a moisture source important to MJO initiation, and 2) proper representation of the diurnal cycle is prerequisite to accurate representation of large-scale climate, at least within the regime studied herein.