Structural Differences in Temperature and Water Vapor among Convective, Near-Neutral, and Stable Boundary Layers and between High-Vertical-Resolution Radiosonde and the ERA5 Reanalysis Data

Abstract Earth’s surface conditions influence the boundary layer and then influence the free atmosphere, which are measured by densely distributed manual and automatic surface stations over land. This study represents the first step toward increasing the role of surface-sensitive observations in high-resolution mesoscale storm prediction. Comparisons are made for planetary boundary layer (PBL) heights and autocovariances and cross covariances of temperature and specific humidity among convective, near-neutral, and stable boundary layers and between high-vertical-resolution radiosonde data (∼5–6 m) and the ERA5 reanalysis data at 119 Chinese radiosonde stations in the summers of 2021–23. The PBL heights in the radiosonde profiles differ greatly from those in the ERA5 reanalysis data, with 137 vertical levels. The PBL heights calculated from the radiosonde profiles at the thinned resolution, which is the same as that of ERA5 but adds a vertical level in the middle of each ERA5 layer, approximate reasonably well to those determined from the original higher vertical resolution profiles. The autocovariances and cross-covariances of temperature and specific humidity in the low troposphere calculated from data with convective PBLs are broader and extend into levels higher than those of near-neutral and stable PBLs. The covariance matrices of the ERA5 are generally consistent in pattern but weak in magnitude with those from radiosonde observations. This study reveals the importance of having a sufficiently high vertical resolution in the low troposphere to simulate PBL heights accurately and to obtain three separate covariance matrices that describe temperature and specific humidity in the different PBLs.