Climatological Variations in the Planetary Boundary Layer over the Continental United States and Europe

Boundary layer processes are important in climate because they determine vertical fluxes of energy, water and trace gases, cloud formation, and lapse rate, all which influence climate feedback processes and therefore climate sensitivity. However, PBL climatology and variability have received little attention, partly for lack of observational datasets, and the representation of the PBL in climate models has not been evaluated in detail. We analyze the PBL over Europe and the continental U.S. during 1981-2005 using radiosonde observations, the ERA-Interim (a reanalysis that assimilates observations), and two contemporary climate models, the NCAR CAM5 and GFDL AM3 (which do not assimilate observations). A bulk Richardson number method is used to determine a PBL height metric in a consistent fashion from these datasets. Although the method yields uncertainties that can exceed 50 % for heights < 1 km, they are generally < 20 % for heights > 1 km.

Using the resulting 25-year datasets with temporal resolution of 2/day (from radiosondes), 4/day (climate models) or 6/day (reanalysis), we have analyzed the climatological diurnal and seasonal variability of the PBL over the two midlatitude spatial domains with the greatest density of radiosonde observations. Climatological heights are generally < 1 km during daytime and < 0.5 km at night over both Europe and the U.S. The amplitude and phase of the diurnal cycle is reasonably well captured by fortuitously-timed radiosonde observations over Europe, but over the U.S. the 0000 and 1200 UTC soundings do not capture the full range of diurnal variability apparent in the reanalysis. Therefore, we base all model-observation comparisons on matched temporal sampling.

During daytime, summertime heights are deeper than in wintertime, but at night winter values exceed summer values. The models all reproduce general features of observed spatial, seasonal and diurnal variability, with better agreement for daytime than nighttime. While ERA-Interim tends to produce lower heights than observed by radiosondes, both climate models tend to produce higher values than observed and have difficultly simulating shallow, stable boundary layers. These new datasets will be available to the research community for further analysis of PBL variability and model verisimilitude.