This dataset contains about 5 years of analysed observations regarding the degree of convective aggregation, or clumping, across the tropics - these are averaged onto a large-scale grid. There are also additional variables which represent environmental fields (e.g. sea surface temperature from satellite data, or humidity profiles averaged from reanalysis data) averaged onto the same large-scale grid. The main aggregation index is the Simple Convective Aggregation Index (SCAI) originally defined in Tobin et al. 2012, Journal of Climate. The data were created during the main years of CloudSat and Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) satellite data so that they could be compared with vertical cloud profiles from this satellite data, and the results of this analysis appear in Stein et al. 2017, Journal of Climate.

Each file is one year of data (although the year may not be complete).

Each variable is an array: var(nlon, nlat, [nlev], ntime)

longitude, latitude, pressure, time are variables in each file

units are attributes of each variable (except non-dimensional ones)

missing_value is 3.0E20 and is an attribute of each variable

Time is in days since 19790101:00Z and is every 3hours at 00z, 03z, ...

The actual temporal frequency of the data is described for each variable below.

The data is for each 10deg X 10deg lat/lon box, 30S-30N (at outer edges of box domain), with each box defined by its centre coordinates and with boxes overlapping each other by 5deg in each direction.

In general, each variable is a spatial average over each box, with the value set to missing if more than 15% of the box is missing data.

Exceptions to this are given below.

The most important exception is for the brightness temperature data, used in aggregation statistics, which is filled in using neighborhood averaging if no more than 5% of the pixels are missing, but otherwise is considered to be all missing data. The percentage of missing pixels is recorded in 'bt_miss_frac'.