The Diurnal Cycle of Clouds and Precipitation at SGP

Description

Sixteen years of Millimeter Wavelength Cloud Radar (MMCR) data from Dec. 1996 to Dec. 2010 are used to examine the annual and seasonal diurnal cycle of clouds, light precipitation and heavy precipitation, and are compared with output from the Multiscale Modeling Framework climate model. Taking advantage of the high temporal and spatial resolution of MMCR, we focus on the vertical structure of the diurnal cycle. We find a strong diurnal pattern in boundary layer cloud frequency and height. Both increase after sunrise in annual mean, as well as in the spring, summer and autumn seasons, while in winter the frequency of boundary layer cloud increases after sunrise but there is no discernable increase in cloud top height. Moreover, diurnal variations of high cloud occurrence are small except in summer during which clouds tends to appear most often in the evening and early morning. As regards drizzle and light precipitation, there is a minimum occurrence in the afternoon and a peak occurrence in the early morning in the middle atmosphere and near the surface in the annual mean. This cycle is most pronounced in summer and while not as obvious in other seasons it is nonetheless present. Heavy precipitation (large reflectivity) is mainly concentrated in the lower atmosphere in the winter, and autumn with no significant diurnal variations. In summer it tends to be most frequent in the early morning (with a strong minimum during the late morning and early afternoon). In spring, heavy precipitation frequency is largest during the late afternoon or early evening. Multiscale Modeling Frameworks (MMFs), also known as superparameterization (SP), have emerged as one approach to explicitly simulate deep convection, fractional cloudiness, cloud overlap, and the spatial distribution of precipitation intensity in climate models (Randall et al. 2003). In this study, we used MMF model outputs in combination with a radar simulator (Marchand et al. 2009) to compare directly the model diurnal cycle with MMCR radar observations. Consistent with previous studies, we find the MMF model produces far less boundary layer cloud both in the annual mean and seasonal means and has some difficulty capturing the correct diurnal variations. In part, this is because the model produces too much precipitation. The model also generates too much high clouds, with the highest cloud tending to be higher and more plentiful than in the observations. Nonetheless the model does a decen