Breakout Summary Report

ARM/ASR User and PI Meeting

19 - 23 March 2018

Session Title: Routine LES: LASSO and Beyond
Session Date: 22 March 2018
Session Time: 10:45 AM - 12:45 PM
Number of Attendees: 55
Summary Authors: William Gustafson, Andrew Vogelmann, & James Mather

Breakout Description

The Large-Eddy Simulation (LES) Atmospheric Radiation Measurement (ARM) Symbiotic Simulation and Observation (LASSO) workflow provides high-resolution simulations of shallow convection at ARM’s Southern Great Plains (SGP) observatory to complement the large suite of observations. The project to develop LASSO has completed its pilot phase and recommendations have been made for formal implementation of the workflow with details available in the technical report at http://www.arm.gov/publications/programdocs/doe-sc-arm-15-039.pdf. At this point, LASSO is also formally beginning to consider expansion to other cloud types and/or locations beyond the current initial implementation.

This breakout session presents the current status of LASSO and implementation plans that are underway. Current LASSO usage is highlighted and success stories shared. Then, time is devoted to discussing LASSO expansion scenarios.

The goal of the expansion discussion is to outline scenarios LASSO should consider beyond the current implementation. Discussion focuses on scientific drivers for the scenarios, high-level descriptions of what the scenarios would encompass, and potential significant difficulties.

Main Discussion

The LASSO breakout session was organized into three components.

**Component 1: LASSO review: current state and implementation plans**
The first component was a review of LASSO and an update of its status. A basic introduction was given to orient attendees unfamiliar with LASSO, of which there were several. Information was also shared regarding the transition from the pilot phase that ended last summer and the current work to design and implement LASSO software to simplify future generation of LASSO data bundles for shallow convection at the Southern Great Plains (SGP) observatory.

Some questions that were asked during this component include whether LASSO forcings will be generated for all days even if the LES is not run for the given day, and if any of the days selected for simulation in 2017 overlap with the LAFE field campaign. There was also interest in having “best” simulations indicated although there was recognition that “best” depends on the user’s application.

**Component 2: Science applications**
Short science-based talks were presented by four users of LASSO data. These talks were aimed at showing various ways LASSO can aid research. Examples include using LASSO as a proxy for understanding observation strategies (Mariko Oue, SBU), using LASSO to compliment field campaign studies such as HI-SCALE (Heng Xiao, PNNL), understanding simulations of cloud populations in terms of different cloud size measurements (Thijs Heus, Cleveland State U.), and differentiating naturally occurring cloud variability due to the diurnal cycle from aerosol-cloud effects (Ian Glenn/Graham Feingold, NOAA). It was encouraging to see various approaches to using the LASSO data bundles ranging from statistical model comparisons with observations to using LASSO as a starting point for more complex simulation studies.

**Component 3: Candidate expansion scenarios**
The attendees were led through a guided discussion of possible expansion scenarios for what LASSO could be used for beyond shallow convection at SGP. It was noted that it is important to keep the shallow convection simulations running at SGP and for that effort to not be substituted by the new one. Notes taken as part of the discussion are included on the slides associated with this breakout session, and readers should refer to those notes for more detailed information.

The first scenario was continental deep convection at SGP. A wide range of science drivers were suggested to motivate choosing this scenario, including many that are common research areas for the Atmospheric Systems Research Program. Examples include the diurnal cycle of convection, cloud organization, cold pools, surface temperature biases, shallow-to-deep convective transition, and precipitation microphysics. If this scenario is chosen, complications that will need to be overcome include choosing an appropriate compromise of domain size and resolution and the life cycle of convection occurring over a large area, much of which is not encompassed by ARM measurements.

The second scenario was marine boundary layer clouds at the East North Atlantic observatory. Science drivers suggested for this scenario include drizzle properties and evaporation, aerosol-cloud interactions, stratocumulus evolution and entrainment, and simulations that could contrast with the current continental shallow-convection configuration. Decisions for this scenario would include whether to simulate a pure-ocean domain or attempt to include the island(s), identifying the most robust forcing options, and the lack of data in the surrounding region, particularly over the ocean.

The third scenario was Arctic clouds either at the North Slope of Alaska (NSA) observatory or for the MOSAiC field campaign. Science drivers for simulations in the Arctic include transition of cloud state, understanding cloud and aerosol layering, and sensitivity of stratus clouds to higher level longwave cloud forcing. Simulations for MOSAiC were framed as a stepping stone for future simulations at NSA given that some of the complications at NSA will not exist over the open-ice region where MOSAiC will occur, e.g., land-sea contrasts, and that MOSAiC will have a network of surface flux measurements. A comment was made that it is important to get good statistics on common phenomena for testing E3SM. A complication with simulating the Arctic is the perception that ice microphysics in current-generation models are unable to capture all the relevant processes. However, this was also discussed as a driver for pursuing Arctic LASSO simulations given the need to more fully evaluate when the parameterizations behave well and poorly. In fact, there was discussion as to whether the ensemble used for LASSO might change with location, where the SGP focused on a forcing ensemble but the Arctic application might want to focus more on a microphysics scheme ensemble. Similar to ENA, a dearth of regional observations will make it difficult to produce reliable forcings at ENA, but modeling groups have successfully simulated the region, so options exist that could be pursued.

Additional scenarios were suggested but time did not allow detailed discussion. These scenarios include multiple convection types and regimes at SGP, LASSO intensive observation periods with mobile facility deployments, interactive surface for land and sea ice to capture surface gradients appropriate to the location(s), Tropical deep convection, MARCUS in the Southern Ocean, the MAGIC region of the Pacific Ocean, and katabatic winds over Greenland/Iceland/AWARE and the impact on net snowfall.

Key Findings

A growing community of users are experimenting with LASSO simulations and applying them to research applications.

There are many scientists excited to see LASSO expand to additional scenarios beyond shallow convection at SGP. All of the options described above had active participation from the audience during discussion and there were proponents for all of the options.

Action Items

Coordinate with the LAFE team to identify LASSO days during the LAFE campaign and determine possible synergistic observation and modeling efforts.

Over the next months, the LASSO team will produce simulations for the 2017 cases. A decision for the next LASSO scenario will be made over the next year.