Using MARCUS, MICRE, and COMBLE Data to Improve Understanding and Modeling of Cloud, Aerosol, and Boundary Layer Processes at High-Latitudes

Abstract

Because it is believed general circulation (GCM) and numerical weather prediction (NWP) models underestimate shortwave radiation over the Southern Ocean (SO) due to inadequate representations of boundary layer (BL) and cloud processes, it is critical to improve our understanding of key aerosol, cloud, precipitation, and BL processes. Over the north Atlantic Ocean (NA), cold-air outbreaks are common, yet few studies of the aerosol and environmental controls of the associated convective BL clouds exist as needed to develop and evaluate GCM representations. Although processes cannot be observed, cloud and aerosol properties can be measured in-situ or remotely retrieved, which when combined with numerical simulations enable process level understanding required to improve model representations.

Unique sets of data on SO clouds were obtained during the Department of Energy (DOE)-funded Measurements of Aerosols, Radiation and Clouds over the Southern Ocean (MARCUS) and Macquarie Island Cloud and Radiation Experiment (MICRE) projects capturing the seasonal, latitudinal and meteorological variability of aerosol, cloud and precipitation properties across the SO and over cold waters south of 60^oS where supercooled and mixed-phase BL clouds are frequent, model biases largest and past observations sparse. Observations being obtained during the DOE Cold-Air Outbreaks in the Marine Boundary Layer Experiment (COMBLE) project are characterizing source air masses of cold-air outbreaks and vertical profiles of clouds, precipitation and aerosols over the NA. Data from these projects will test the following hypotheses:

1)     The concentration of cloud condensation nuclei (N_CCN) and accumulation mode aerosols (N_a) are more correlated with the presence/amount of precipitation than with surface wind speeds regardless of the synoptic conditions or geographic location where observations were made;

2)     There are three hypotheses examining relations between aerosols, clouds and the environment;

a.      The variation in supercooled liquid water (SLW), ice mass, and ice in drizzle in shallow clouds at a given cloud-top temperature (CTT) are more strongly correlated with updraft velocities than changes in N_a;

b.     A similar relation between CTT and SLW occurrence holds in the NA and SO;

c.      Differences in the overall occurrence of SLW between the NA and SO can be explained by changes in the distribution of CTT independent of N_a.

3)     BL processes assumed in conventional dry PBL schemes are insufficient to describe processes in cloud-topped marine BL, thus these conventional schemes have significant deficiency to simulate cloud-topped BLs and the corresponding clouds over the SO.

Hypotheses will be tested by quantifying how distributions of cloud macrophysical (e.g., CTT, cloud boundaries) and microphysical (e.g., SLW, ice mass, ice in drizzle, particle size, liquid water path) properties vary with environmental (e.g., lower atmospheric stability, cold air outbreak index, convective potential energy, sea surface temperature, wind speed, air mass origin, synoptic condition, presence of sea ice, whether clouds coupled to BL, etc.) and aerosol characteristics (surface and fluorescent biological aerosol concentrations, N_a, N_CCN, chlorophyll-a concentration, ice nucleating particle concentration, etc.). The hypotheses will be tested separately for MARCUS, MICRE and COMBLE as well as using the integrated data. Statistical tests (e.g., Mann-Whitney U test, Median test) will determine the significance of correlations between cloud properties and aerosol/environmental conditions to evaluate the hypotheses.

Hypotheses will be further tested by using the WRF model to replicate the observed dependence of cloud properties on synoptic conditions, BL processes and aerosol characteristics, determine if simulations are consistent with primary hypotheses (true or false) and if the process-oriented understandings obtained from the above analyses is represented in models. Sensitivity tests with alternate microphysical and PBL schemes and tests where key processes are modified will test impacts on simulated cloud structure and properties.