Impact of changes in temporal resolution and convective parameterization on the simulation of tropical climate in NCAR CAM3 GCM

Abstract

In this thesis, the simulation of tropical climate by the Community Atmospheric Model (CAM) developed at National Center for Atmospheric Research (NCAR) has been discussed. Although the model successfully simulates many aspects of the observations, several shortcomings remain. For example, the rainfall over the south of the equator between 60°E to 120°W is underestimated during the northern winter, and overestimated over tropical Africa, northern Australia, and north of the equator over the western Pacific. Over the western and eastern equatorial central Pacific, and the eastern and head of the Bay of Bengal, the model underestimates the rainfall during northern summer and overestimates it over Saudi Arabia, the western Indian Ocean, the western Arabian Sea, and the western part of the South Pacific. The intra-seasonal oscillation is too weak and propagates too fast. The simulation of the spatial pattern and temporal variability of Indian Summer Monsoon Rainfall (ISMR) is poor in the model.

The rectification of some of the shortcomings of the model is the main objective of this work. For this purpose, several numerical experiments were carried out to understand the dependence of rainfall simulation on certain model components. First, the dependence of the model simulation on time step was examined. Secondly, the influence of the moist physics on rainfall simulation was explored. Both aqua-planet as well as actual-planet frameworks were used in this investigation.

When the model time step was increased from 5 minutes to 60 minutes, the rainfall in the Intertropical Convergence Zone (ITCZ) decreased substantially. The increase in ITCZ rainfall at higher temporal resolution was primarily due to the increase in large-scale precipitation. This increase in rainfall was due to positive feedback between surface evaporation, latent heating, and surface wind speed. Similar results were obtained when the semi-Lagrangian dynamical core was replaced by the Eulerian dynamical core. When the surface evaporation was specified, changes in rainfall were largely insensitive to temporal resolution. The impact of time step on rainfall was more sensitive to the latitudinal gradient of sea surface temperature (SST) than to the magnitude of SST. With a smaller time step size, the vertical heating profile becomes top-heavy, which in turn reduces the speed of Kelvin waves. Change in time step has a significant impact on the simulation of the monsoon. The simulation capability of interannual variability (IAV) of ISMR is better with a smaller time step.

From a multiple parameter sensitivity study, it is shown that convective relaxation time scale (TAU) is the most influential parameter and has a strong impact on model simulation. By varying TAU over a wide range, it has been shown that it affects the relative proportion of different rainfall components. An increase in TAU decreases the fraction of deep convective rainfall (DRF), but increases the fraction of shallow convective rainfall (SRF) and large-scale rainfall (LRF). This affects the vertical heating profile by lowering the level of peak heating at higher TAU. It is shown that when the value of TAU is prescribed as 8 hours, the simulation of rainfall is more realistic. The improvement in the simulation of the pattern of climatological mean rainfall, hierarchy of intra-seasonal variability (ISV), Kelvin waves, and ISMR is demonstrated.

Finally, a novel concept for TAU, i.e., dynamic TAU, is proposed, which is assumed to be a function of Convective Available Potential Energy (CAPE) and hence varies in space and time. Two dynamic TAUs are proposed. In the first case, TAU decreases linearly with an increase of CAPE, while in the second case, TAU decreases exponentially with an increase of CAPE. It is shown that both dynamic TAUs perform better than the control configuration of the model. Comparatively, TAU (linear) performs better than TAU (exponential). Moreover, in some respects, the performance of TAU (linear) is found to be better than that achieved with TAU (8 hr), hence TAU (linear) is recommended in place of the current static TAU