CloudSat/Calipso data and plot calculation details

Jen Kay (jenkay at ucar dot edu), last updated 3-30-2011

 

CloudSat (a 94 GHz radar) and CALIOP (a 532/1064 nm depolarization lidar) provide a unique global view of cloud vertical structure.

I am working with the radar cloud mask and dBZ (2B-GEOPROF) and a co-located lidar cloud mask (2B-GEOPROF-LIDAR, termed RL). 

I use data from the R04 release, available June 2006-present.

 

More Information: CloudSat website, CloudSat data processing center, Calipso website

Example data:  radar, radar_lidar, dataqualityconcerns

 

A particularly striking example of the synergy of using CloudSat and CALIOP to detect clouds is available here.

You can see how the lidar detects thin cloud at cloud top and low cloud that the radar misses (green).

You can also see how radar detects most mid cloud, cloud that the lidar misses due to attenuation (blue).

There are some interesting low clouds that have radar-only tops (a resolution issue?) and lidar-only bottoms (radar out due to ground clutter).

 

Plots are available at the following websites:

Arctic R04 cloud fraction plots

Arctic R04 dBZ plots

Global R04 cloud fraction plots

Global R04 dBZ plots

 

Plot calculation details:

 

1. Cloud ID:

For the CloudSat only analysis, a positive cloud ID results if the 2B-GEOPROF cloud mask equals 20, 30, 40.  For more information on 2B-GEOPROF cloud mask, see the quality statement.  For identification of specific cloud types (binning by profile) or for calculation of cloud thickness or cloud top height, I require 2 cloudy bins (cloud thickness (dz) ge 580 m).   The 2B-GEOPROF-LIDAR product provides the fraction of the radar volume that is identified as cloud by the lidar.   If any lidar cloud is present in the radar volume (i.e., fraction > 1%), I count this as a positive cloud ID.

 

2. Cloud type definitions follow CAM pressure-based cloud type definitions:

low = 1200-700 hPa (0-2.75 km asl)

mid = 700-400 hPa  (2.75-7 km asl)

high = 400-50 hPa (7 - 22.5 km asl)

I converted the pressure-based definition to a height-asl definition (using log(P) altitude, H=7.5 km).

 

For fun - I also identify thick and unique clouds in the column.

thick = Require 15 cloudy bins (cdz=3.6 km) or 30 cloudy bins (cdz=7.2 km) at any height in the column.

uniq = Require less than 2 bins outside of the specified height range have a positive cloud ID.

 

3. CloudSat Surface Clutter:

Because CloudSat receives a large amount of backscatter from the surface, the range bins close the ground surface are contaminated by surface clutter.  For R03 - the net result is that CloudSat has trouble identifying clouds in the bottom km of the profile (i.e., 0-0.96 km agl).  For R04 – CloudSat has trouble identifying clouds in the bottom 720 m of the profile.  When I manually exclude bin1tobin3, the file names will have no attached to their name. Trust the plots with no bin1 to bin3 – they represent the best CloudSat view of cloudiness in the atmosphere and are valid only above 720 m agl.

 

4. Cloud top height and cloud thickness calculations (still experimenting…):

CldTopHt, CldDz, and CldTotDz calculations require that clouds have at least two bins (580 m thick) and that they are entirely contained within the specified height range.  For example, the all calculation includes clouds that are entirely contained from 0 to 20 km, whereas the low calculations include only clouds that are entirely contained from 0 to 2.88 km and the low+ calculations include only clouds that are entirely contained from 0 to 4.08 km. The individual cloud thickness calculation (CldDz) incorporates all individual contiguous clouds.  The total cloud thickness calculation (CldTotDz) is a sum of all individual clouds in a column.