This dataset comprises of model output from 25 runs (5 case studies, with 5 runs in each case study) of the Met Office Unified Model (MetUM) in realistic limited-area one-way nesting mode. The output data include values for model fields (e.g. temperature, humidity, winds, pressure) at model grid points over regularly spaced time intervals. These runs were used in a paper on convective aggregation: Holloway (2017, Journal of Advances in Modeling Earth Systems).

All runs use the ""New Dynamics"" dynamical core, MetUM version 7.5, as described in Holloway (2017). The simulations are run with 4-km horizontal grid spacing. They all have a horizontal domain size of 20 degrees latitude X 20 degrees longitude (or 574 X 574 grid points, although the grid points in the outer 8 points on all sides, the ""rim"", should be discarded before analysis), with 70 vertical levels. All runs are initialised from operational analyses from the European Centre for Medium-Range Weather Forecasting (ECMWF) taken from actual cases. Lateral boundary conditions are comprised of 6-hourly ECMWF analyses, and the model is relaxed to these conditions in and near the outer rim as described in Holloway (2017). Sea surface temperatures (SST) are taken from the initial ECMWF analysis and are held constant in time for the 15 days (but are not constant in space). There are small land regions in four of the case studies which include an interactive land surface model.

Each simulation was run for 15 days. The model output includes hourly model-level prognostic variables (temperature, specific humidity, pressure, wind components, liquid water, ice water) as well as some model-level increments to temperature and specific humidity. There are also many fields containing surface variables and fluxes (averaged over each hour or every 15 minutes). Note that the ""control"" simulations have slightly more available data than the other four runs in each of the five case studies.

The five case studies are centred on the equator and occur between 2008 and 2010. See Holloway (2017) for further details:

http://onlinelibrary.wiley.com/doi/10.1002/2017MS000980/full

For each case, there are five runs:

1) control (interactive radiation, interactive surface fluxes)

2) constant radiative cooling run (radiative cooling over sea points is prescribed from domain-time mean of control run)

3) constant surface flux run (surface latent and sensible heat fluxes over sea points are prescribed from domain-time mean of control run)

4) constant radiative cooling and constant surface flux run (combination of 2 and 3 above)

5) no rain evaporation run (rain is prevented from evaporating in the atmosphere)"