Testing convection onset conditions in convective parameterization schemes using ARM observations at SGP and TWP

Description

Convection initiation or onset is very important to climate, from diurnal cycle and intraseasonal variability to precipitation frequency and intensity climatology. Many global climate models (GCMs) do not simulate the diurnal cycle correctly, have weak intraseasonal variability, and produce too frequent convection with too weak intensity. All these biases are in one way or another related to the improper onset of convection in the models. In this work, we make extensive use of the ARM observations from intensive operational periods (IOPs) at the Southern Great Plains (SGP) and Tropical Western Pacific (TWP) sites together with the long-term single-column model forcing data from Atmospheric System Research (ASR) principal investigators to examine many convection trigger functions commonly used in GCMs. For each trigger function examined, the performance is stratified into three groups: correct prediction, over-prediction, and under-prediction of convection onset. A statistical skill score is used to quantify the performance of each trigger function. To further identify the critical factors responsible for the performance of a trigger function, for each group we further examine the composite profiles of moist static energy, temperature, moisture, and large-scale vertical velocity. It is found that there are systematic differences in these parameters among the three groups. These results will provide useful guidance to improving the convection trigger function and the simulation of convection onset in GCMs. More details will be presented at the meeting.