Breakout Summary Report
ARM/ASR User and PI Meeting
19 - 23 March 2018
Warm marine low cloud processes and modeling
19 March 2018
1:30 PM - 3:30 PM
100
Robert Wood and Xue Zheng
19 March 2018
1:30 PM - 3:30 PM
100
Robert Wood and Xue Zheng
Breakout Description
Conveners: Rob Wood and Xue ZhengDescription: The goal of this breakout is to bring together observational, modeling, and model evaluation with focus on processes controlling warm marine low clouds. Of particular interest are (a) observational analyses from recent sites/campaigns to study warm marine low clouds (e.g., MAGIC, CAP-MBL, ENA, LASIC); (b) LES and other process modeling designed to advance understanding warm marine low clouds and cloud processes involving interactions among aerosol, cloud, drizzle, turbulence and atmospheric circulations; (c) assessment and improvements in the treatment of warm marine clouds in GCMs and how observations and process modeling can be used to achieve improved cloud representation in large-scale Models.
Main Discussion
The session consisted of science presentations from speakers. Broadly, these fell into the following categories: (a) microphysical and dynamical variability of clouds; (b) model representation and evaluation; (c) precipitation processes; (d) aerosol variability; (e) new retrieval approaches.Key Findings
MICROPHYSICAL AND DYNAMICAL VARIABILITY OF CLOUDSVirendra Ghate [ANL] examined a large data set consisting of closed and open cells from the ENA site (>200 hours of each cell type). Specific focus on the work is to examine impacts of precipitation on MBL turbulence structure, as function of the radiative driving of the systems.
Retrieved drizzle microphysical properties were used, along with turbulence information in and below cloud using radar and lidar Doppler. These are the first observations clearly showing that drizzle evaporation weakens subcloud layer turbulence.
Katia Lamer [Penn State] examined low clouds from the ENA site and showed that, in contrast to spaceborne radar, ground-based ARM remote sensors possess the sensitivity and vertical resolution necessary to observe the full spectrum from drizzle to light rain. Parsivel disdrometer observations confirmed the reliability of rain rates retrieved from ground-based backscattering measurements. CloudSat only observes 5% of the drizzle events detected by ground-based sensors, because most observed drizzle events comprise virga with a precipitation rate of 10-4 mm hr-1, a rate below the sensitivity of CloudSat. Stratiform cloud precipitation frequency increases with LWP and stabilizes at about 50% when LWP reaches 300 g m-2.
Ewan O’Connor [U Helsinki] described a new data set from the Doppler lidar at ENA that can be used to assess turbulent MBL dynamics. The importance of accurately quantifying uncertainties for estimating turbulence properties was highlighted. Ewan also showed how Doppler lidar can be used to examine island impacts. The data set will be made available in May 2018.
Jim Hudson [DRI] showed observations from the MASE aircraft campaign (NE Pacific coastal Sc) and found a strong association between the CCN spectral shape (unimodal or monomodal) and the production of drizzle in marine Sc. Clouds associated with bimodal CCN had greater droplet concentrations with broader size distributions that promoted up to two orders of magnitude more drizzle than clouds associated with unimodal CCN. Low vertical winds of stratus confined nucleation to the accumulation mode of bimodal spectra. But for unimodal CCN, nucleation occurred on a wider variety of critical supersaturation (Sc) Aitken particles that led to broader droplet spectra than the narrower Sc range of the accumulation mode of bimodal CCN.
MODEL REPRESENTATION AND EVALUATION
Takanobu Yamaguchi [NOAA ESRL] introduced a new SciDAC project to improve representation of low clouds in E3SM, a collaborative project with P. Bogenschutz (LLNL), G. Feingold (NOAA), and D. Martin (LBNL). The project aims to improve representation of low clouds in E3SM by using higher vertical resolution for selected processes. The framework will be refined with use of E3SM, E3SM-single column model, and SAM (as a regional model); for example, resolution will be adjusted depending on atmospheric state (adaptive vertical mesh refinement).
Peng Wu [U. Arizona] used ENA remote-sensing data to evaluate autoconversion and accretion enhancement factors due to subgrid variability in GCM warm-rain parameterizations.
Compared to observations, GCMs produce precipitation that is too frequent and too light, due in part to the misrepresentation of subgrid-scale variance and covariance of cloud and precipitation microphysical properties in the autoconversion enhancement factor. In general, it was shown that to account for subgrid-scale variabilities, GCMs should decrease autoconversion rate and increase accretion rate to produce similar precipitation rates as observation. Adjustment of the two process rates depends on the boundary-layer regimes (stable versus unstable) and also upon the grid size.
Hugh Morrison [NCAR] used a simple Eulerian model of condensational growth during ascent with a vertical resolution comparable to those in LES/CRM models. A number of different bin microphysical approaches were used and compared with an exact Lagrangian parcel solution. He showed that all the bin schemes develop spurious broadening of modeled cloud droplet spectra, suggesting that major problems remain in the use of bin microphysical schemes in Eulerian models, and suggested that Lagrangian droplet models should be developed to accurately represent condensational growth on discrete grids.
WARM RAIN PRECIPITATION PROCESSES
Pavlos Kollias [Stonybrook University] showed examples of the ENA radar capability for characterizing clouds and drizzle at the ENA site, focusing on the ACE-ENA periods. A combination of vertically pointing 35-GHz radar and ceilometer backscattering observations is used to estimate light precipitation mass flux [10-4 to 3-5 mm hr-1] below cloud using the approach of O’Connor. These best estimates are then applied to the scanning radar low-elevation surveillance observations, which are converted to domain rain rate using reflectivity-rain rate (Z-R) relationships. The techniques together provide complementary views on shallow precipitation. Vertically pointing sensors collect information on weaker rain events (due to their higher sensitivity) at higher temporal resolution. On the other hand, while less sensitive and of coarser temporal resolution, scanning sensors document a larger number of heavier drizzle events and provide domain-representative estimates of shallow precipitation.
Fan Yang [BNL] derived an analytical relationship between drizzle virga depth and cloud thickness. This relationship was tested using ENA observations and shows good agreement with the analytical solution. Surprisingly, the relationship did not appear to require information on cloud droplet size.
AEROSOL VARIABILITY
Guanjie Zheng [BNL] used the long record of aerosol observations at the ENA site to examine seasonal variation in surface aerosol and the controlling processes. Seasonal variations of different size modes showed different seasonal cycles. Sea spray aerosol was shown to contribute only minimally to the CCN population during all seasons. Aitken and accumulation mode concentrations scaled well with BC entrained from free troposphere (but not wind speed), suggesting a significant influence of long-range transport on the aerosol population at the ENA site.
NEW RETRIEVAL APPROACHES
Zeen Zhu [SUNY] applied a dual wavelength ratio (DWR) technique to the zenith measurements from the scanning ARM Cloud Radars (SACRs) at the ENA site to retrieve vertical profiles of liquid water content of clouds. Two variants were used: optimal estimation theory and a DWR fitting function. The DWR technique makes no assumptions about the vertical structure of the LWC profile; it is not affected by the presence of drizzle and provides LWP estimates. The performance of LWC retrievals showed considerable promise and compared well with MWR LWP estimates. Zhu hopes to process LWC data for the community based on this method.