Combined CloudSat spaceborne radar and CALIPSO spaceborne lidar cloud fraction dataset (3S-GEOPROF-COMB)

The following was contributed by Leah Bertrand and Jennifer Kay, July, 2023:

The 3S-GEOPROF-COMB product is a globally-gridded (level 3S) community data product summarizing geometrical profiles (-GEOPROF) of hydrometeor occurrence from combined (-COMB) CloudSat and CALIPSO observations (Bertrand et al. 2023). CloudSat (a 94 GHz radar) and CALIPSO (a 532/1064 nm depolarization lidar) provide a unique global view of vertical cloud fraction and spatial distributions.  When combined, they provide the best available global estimate of vertical cloud fraction.  We have used the radar cloud mask (2B-GEOPROF R05, Marchand 2008) and the co-located lidar cloud mask (2B-GEOPROF-LIDAR P2 R05, Mace and Zhang 2014) datasets to produce cloud fraction climatologies.

An example of the synergy of CloudSat and CALIPSO to detect clouds along a single orbit is shown in Figure 1. The lidar detects thin cloud at cloud top and low cloud that the radar misses (blue), while the radar detects most mid-level cloud that the lidar misses due to attenuation (orange). In all orbits (Figure 2), the lidar alone detects the majority of clouds above 13 km (blue), while the radar alone detects the most clouds from 1 to 8 km (orange).

We combine data from multiple orbits to produce gridded cloud fraction climatologies in netCDF format. For example, Figure 3b shows the zonal annual mean vertical cloud fraction from the combined CloudSat and CALIPSO product. CloudSat and CALIPSO provided the first  look at the vertical structure of clouds resolved in this way. Figure 3a shows the annual mean total cloud cover map from CloudSat and CALIPSO.

CloudSat and CALIPSO provide insight into the vertical distribution of clouds at all latitudes where they are available (82 S to 82 N). Data files are distributed at monthly and seasonal frequency, at 2.5° x 2.5°, 5° x 5°, and 10° x 10° degree horizontal resolution, and 240 m vertical resolution. Because of their limited sampling, some data (e.g. 2.5° monthly) may not be significant in all grid cells. Users should look at the number of overpasses ("n_overpasses") and the diurnal cycle ("localhour_*") to quantify the sampling in their regions of interest (see e.g. Kotarba and Solecki 2021).

Users should note a change in CloudSat data collection patterns after April 2011, when the satellite switched to Daylight-Only Operations (DO-Op) mode. The greatest reductions are in the southern hemisphere during austral winter (JJA) and the smallest reductions are in the Arctic, with a global-mean 44% reduction in sampling. We provide users with the option to subsample the pre-2011 data to the post-2011 collection patterns using the "doop" coordinate, which estimates the effect of the reduced sampling on the data.

Though the climatologies are global, CloudSat and CALIPSO are especially useful in the polar regions (e.g. Kay et al. 2008, Kay and Gettelman 2009, Bromwich et al. 2012, Morrison et al. 2018).  Unlike many satellite-based estimates of cloud amount, CloudSat and CALIPSO are "active" instruments that send radiation out and measure the return signal.  As such, CloudSat and CALIPSO do not rely on thermal or albedo contrast to detect clouds and thus unlike many satellite retrieval algorithms, their cloud fraction masks are not confused by the atmospheric inversions or bright surfaces that are common in polar regions.

For further information, including evaluation of the dataset against comparable spaceborne and ground-based products, please see the data description paper Bertrand et al. 2023.##