This data set contains UCLALES-SALSA simulation outputs used in a closure study of aerosol-cloud interactions during stratocumulus formation using in situ observations from the 2020 cloud sampling campaign at the Puijo SMEAR IV station in Kuopio, Finland (62.9092o 27.6556o, 306 m above mean sea level, 225 m above local lake level, FMISID: 101587, LPNN: 3618). Case studies occurred in September, 24th, 2020 from 07:54 to 12:49 (UTC+2) and in October, 31st, 2020 from 00:35 to 06:35. Both cases correspond to single-layer clouds topped by clear sky conditions.

Large-eddy-simulations were performed with UCLALES-SALSA (DEV version 17.08.2021) available from https://github.com/UCLALES-SALSA/UCLALES-SALSA/tree/DEV). Simulations ran in parallel using 64 processors in 2 nodes. Sub-domain outputs were merged using post-processing code included in https://github.com/UCLALES-SALSA/UCLALES-SALSA/tree/DEV/les-post-processing. A detailed description of model inputs is given in the NAMELIST file of each simulation.

The model domain contain 64 by 64 by 120 grid points with horizontal and vertical resolution of 30 m and 10 m, respectively. The final size of the horizontal model domain is 1.92 km by 1.92 km and extends up to 1.2 km in altitude. Latitude at the Puijo station was set to be 62.53 degrees. Simulations were started two hours before the beginning of the period of interest, the 1st hour was set as spin-up period and data analysis started after the end of the 2nd hour to match local time. Simulated lasted for 7 hours in total for the case of 24-09-2020, and 8 hours in total for the case of 31-10-2020. Every simulation was initialized combining meteorological data from the Savilahti and Vehmasmäki stations, 2 km southwest and 13 km southwest to the Puijo station, respectively. Data from ECMWF-ERA 5 was used when observations were not available. Simulated properties at all grid points of the model domain were taken as snapshots every minute. The size of these data files is not suitable for this data repository. Link to external sources are provided. Simulations for every case study comprise a base scenario and alternative scenarios used to assess the model sensitivity to variations in conditions used for initialization. All data included here is analyzed in the manuscript “Aerosol-stratocumulus interactions: Towards better process understanding using closures between observations and large eddy simulations", submitted to the peer-reviewed journal “Atmospheric Chemistry and Physics” (OCLC number 57509369).

As a side note, observations used in the closure study from Halo doppler lidar and cloud radar are available from the Aerosol, Clouds and Trace Gases-ACTRIS data centre: -CLU (2002) Cloud profiling products: Classification, Drizzle, Ice water content, Liquid water content, Categorize; cloud profiling measurements: Lidar, Radar; harmonie-fmi-6-11 model data. Cloudnet Collection from 2020-09-10 to 2020-12-01 at the Savilahti Station, Kuopio Finland. https://hdl.handle.net/21.12132/2.ef1a7d312c8a402d.

Observations of hydrometeors size distributions are available after request to the authors and presented in
- Tiitta, P., Leskinen, A., Kaikkonen, V., Molkoselkä, E., Mäkynen, A., Joutsensaari, J., Calderon, S., Romakkaniemi, S., & Komppula, M. (2022). Intercomparison of holographic imaging and single-particle forward light scattering in-situ measurements of liquid clouds in changing atmospheric conditions. Atmospheric Measurement Techniques Discussions, 2022, 1–20. https://doi.org/10.5194/amt-2021-423

The data set is divided in five containers , each one of them is labeled with the date of the cloud event and the identification of the simulation scenario as “YYYYMMDD_scenario”, YYYY as year, MM as month and DD as day. Each simulation is stored in one container that includes the NAMELIST file with the model settings, the sound_in file with the vertical profiles used for model initialization, the kmls.lay profile or radiation background profiles, and several NetCDF files including post-processed model outputs for 4D variables and 5D variables (binned distributions). Variables names and units are described in the metadata. The data set has been stored in the following containers: -20200924_baseL41ms, base scenario, UCLALES-SALSA in level 4 for liquid cloud, initialization with aerosol particles in a single mixing state (Size: 100 GB) - 20200924_altL42ms, alternative scenario, UCLALES-SALSA in level 4 for liquid cloud, initialization with aerosol particles in two mixing states (Size: 204.8 GB) - 20201031_baseL4, base scenario, UCLALES-SALSA in level 4 for liquid cloud, initialization with 100 % of observed aerosol loading (Size: 122.6 GB) - 20201031_altL4red, alternative scenario, UCLALES-SALSA in level 4 for liquid cloud, initialization with 60% of observed aerosol loading (40 % reduction) (Size: 122.6 GB) - 20201031_altL5red, alternative scenario, UCLALES-SALSA in level 5 for mixed-phase cloud, initialization with 60% of observed aerosol loading (40 % reduction) (Size: 204.76 GB)