Comparing Boundary layer description from ICON model with ground based observations: a detailed look at warm rain formation

Claudia Acquistapace

University of Cologne, Germany

Friday, May 26, 2017, 11:00 am
DSRC Room 2A305


Abstract

The prediction of the way boundary layer clouds respond to climate change is affected by large uncertainties mainly due to the way low-level clouds are parameterized in climate models.

Climate models parameterize turbulent motions in the planetary boundary layer (PBL) by applying so-called PBL-schemes, which can be situation dependent. In order to better understand how low-level clouds are affected by the PBL scheme at a given time, a classification of different boundary layer types is used, associated with specific parameterizations of entrainment, mixing and shallow cumulus or non local schemes. Since the choice of one PBL scheme can dramatically affect the model output, it is essential to evaluate the parametrized PBL schemes with observations.

We exploit a new PBL classification based on wind lidar, ceilometer and tower measurements applied to observations of the JOYCE supersite in Germany to statistically evaluate the PBL representation in the high-resolution (150 - 300 m) Icosahedral non-hydrostatic general circulation model (ICON).

ICON is developed by the Max Planck Institute for Meteorology (MPI-M) and the German Weather Service (DWD). A PBL classification analogous to the one used for the observations has been developed for the model output to properly compare PBL regimes identified in model and observations. Also, different cloudy regimes described on the basis of the PBL classification are compared, with a specific focus on the process of drizzle formation.

In models, this process is parametrized by the autoconversion. Several parameterizations have been proposed for numerical models but the evaluation of such schemes remains difficult due to the lack of direct observations.

Here, we focus on a novel statistical criterion to detect drizzle onset within clouds based on higher Doppler spectra moments, as opposed to the commonly used "standard" moments reflectivity, mean Doppler velocity and Doppler spectrum width.

The new method has been tested on individual cases at JOYCE and areas of drizzle formation within the cloud have been retrieved. We propose that this new method can provide additional observational constraints for autoconversion parametrization in numerical models. In order to fulfill this purpose, the methodology is currently being implemented as an extension of the Cloudnet target categorization algorithm, which is currently operating at many different sites across the world and has often been used for model evaluation.

Visitors

You must provide an accepted form of identification at the Visitor Center to obtain a vistor badge. Security personnel also inspect vehicles prior to entrance of the site. Please allow extra time for these procedures.

After receiving a badge, you must arrive at the DSRC Lobby at least 5 minutes before the seminar starts to meet your security escort. If you arrive after that time, you will not be allowed entry.

Foreign Nationals: Please email the seminar contact at least 48 hours prior to the seminar to provide additional information required for security purposes.

Seminar Contact: shannon.kelly@noaa.gov