Case selection for LASSO-ENA went through a careful selection process. An early scientific consideration was on sampling from the wide variety of the ENA cloud regimes, enabling LASSO-ENA simulations to enhance studies on the mesoscale organization and cellular structure of marine boundary clouds. The need to simulate many events for those science drivers also needed to be balanced by the requirements for an efficient use of ARM computing resources, including considerations for representative ensembles useful for detailed cloud process studies at relevant model domain size, grid scale and microphysical complexity.
The LASSO-ENA scenario (64 cases, with a core of 24-hour 6-member ensemble simulations) was divided across three marine boundary layer stratocumulus organizational categories found over the site: 1) closed cell stratocumulus organization (20 cases), 2) open cellular cloud organization (23 cases), and 3) a transitional or hybrid organizational mode (21 cases) that exhibited frequent clouds over the ENA site, but less well-defined spatial or continuous open/closed behaviors. Datasets were collected at the ENA site starting from its origin in late 2014 and considered cases until late 2022. Of the 64 cases, there were 51 unique “events” (i.e., 14 represented sequential days).
Figure 2 Illustration of the ENA LES small and large model domains relative to the main site location and the 5 degree model forcing scene.
The case selection process at ENA consisted of a few key steps:
First, a list of candidate dates was obtained from a variety of sources, including community recommendations, expert- or ENA literature identified case lists, and objective cloud regime classifications developed within ARM [Tian et al., 2025]. Our primary reference when narrowing case selection was ARM-profile and satellite imagery analysis, including cloud type definitions determined by column-radar time-height cloud property checks using ARM ARSCL products as found in Rémillard et al. [2012] and Sookdar et al. [2025], or spatial cloud coverage checks (visible imagery from SEVIRI or polar orbiting Aqua and Terra satellites) similar to Jensen et al. [2021]. Several cases exhibited “closed-cellular” conditions for a column-type perspective over the ENA site however lacked contiguous cloud coverage over larger O(100–400) km domains. For LASSO-ENA, these events were typically defined under “transitional/hybrid” as compared to characteristic “closed cellular” cases. In agreement with recent studies including Naud et al. [2016] and Ghate et al. [2020], the common LASSO-ENA “open cellular” events were identified under colder-season and post-frontal conditions, having cold-air outbreak characteristics. This LASSO-ENA selection did not identify a summertime (JJA) open cellular event. Only a small number of summertime closed cellular events were identified in the final scenario cases. Late shoulder season events found in the final record were often classified as “transitional” organization, sometimes reflecting a mixture of open and closed cellular behaviors within a given day or sequence of days.
A second, interwoven selection criteria was based on the quality of the ARM ENA datasets associated with the identified cases. Case selection priority was given by the LASSO team to days during ENA field campaigns [e.g, ACE-ENA Wang et al., 2022] or relevant sitewide activities. These times typically offered additional aircraft and/or radiosonde launches, advanced scanning radar and profiling capabilities, and other complementary products that could enrich anticipated studies. Events having additional radiosonde launches in common with the ACE-ENA field campaign, NPOESS validation efforts, and/or the GNSS ground station campaign during the ENA record were typically included, however this rarely was a deciding factor over suitability of cloud regime type and properties. Similarly, timeframes that had available scanning radar assets (SACR, XSAPR) were given priority, however the highest-quality scanning radar availability was often also associated with ACE-ENA campaign windows. Accordingly, suitable events would have been favored as having ARM AAF Gulfstream-1 aircraft and/or more frequent radiosonde (4/day) launches.
A third consideration tied to community recommendations was to place emphasis on prioritizing events having boundary layer winds directed from a northerly direction (approximately 315 to 90 degrees). This suggestion is found in multiple observational studies at ENA because this site sits at the northwest edge of Graciosa Island. The northerly wind directions have been suggested to partially mitigate the potential role that island terrain (Graciosa and surrounding) may lead to enhanced/suppressed ENA cloud properties in site observations [e.g., Ghate et al., 2021, Giangrande et al., 2019]. Overall, LASSO-ENA prioritized case selection for “oceanic” wind direction first, however the final library includes several events having southerly and/or westerly flows. Recent longer-term ENA studies as in Sookdar et al. [2025] have suggested the uncertainty of key observations LASSO uses under these flows (i.e., cloud fraction, LWP, COD) often falls within the uncertainty expected for instrument ability to sample those fields.
With respect to aerosol conditions and their role on case selection, the LASSO-ENA ensembles include 3 different aerosol loading configurations (low, mean, high). These levels were not specified for each case individually, but determined using the statistical aerosol behaviors for Aitken and Accumulation mode peaks across a multi-year ENA record. Additional information on the aerosol distribution properties and their estimation be found in the Modeling Methodology Section. Breakdowns for aerosol properties according to the event type and larger-scale synoptic conditions were performed, with the selected levels determined as suitable to capture the spread of possible site conditions.
