This combined Aqua and Terra MODIS Cloud Properties product is especially geared to facilitate climate scientists who deal with both models and observations. Parsing this product’s recursive “COSP” acronym in its short-name reveals much about its nature. COSP stands for Cloud Feedback Model Intercomparison Project (CFMIP) Observation Simulator Package. “The primary objective of CFMIP is to inform future assessments of cloud feedbacks through improved understanding of cloud-climate feedback mechanisms and better evaluation of cloud processes and cloud feedbacks in climate models.”¹ CFMIP work is also being used to focus our understanding of other facets of climate change that include circulation, regional-scale precipitation, and non-linear changes. MODIS observations offer important information to help us better understand the error factor within climate models. Particle size retrievals, in particular, offer a unique view of cloud-aerosol interactions. The COSP package contains a MODIS simulator component that helps us to compare models and observations through synthetic MODIS observations derived from climate model integrations. The MODIS simulation-derived output helps feed the current generation (CMIP6) of the climate modelling framework called the Coupled Model Intercomparison Project (CMIP).

A code bug was discovered in mid 2021 in the COSP software, which caused a significant problem in the Low, Middle, and High aggregations of Cloud Fraction (from Cloud Mask). This problem was corrected via a new reprocessing cycle during the 2022 calendar year. Therefore, if any user had acquired the v6.1 collection MCD06COSP data files with production dates in 2021 or earlier, they should obtain the new, corrected v6.2 collection MCD06COSP data files, with “2022” (or later) production dates, from LAADS. The corrected v6.2 collection files will have “2022” as part of the filename’s production date-field, which immediately follows the collection’s version number (062). Please note that this new v6.2 collection is still derived from v6.1 collection’s 06_L2 input files. The v6.1 MCD06COSP_M3_MODIS collection will remain available until November 30, 2022, and marked for decommissioning following that date.

MCD06COSP_M3_MODIS represents the monthly product’s short-name with specific dataset definitions with which to compare against climate model output. The L3 monthly product is derived by aggregating the daily-produced Aqua+Terra/MODIS D3 Cloud Properties product (MCD06COSP_D3_MODIS). Provided in netCDF4 format, it contains 23 aggregated science data sets (SDS/parameters) that include the following:

1. Cloud Mask Fraction
2. Cloud Mask Fraction High
3. Cloud Mask Fraction Low
4. Cloud Mask Fraction Mid
5. Cloud Optical Thickness Ice
6. Cloud Optical Thickness Liquid
7. Cloud Optical Thickness Log10 Ice
8. Cloud Optical Thickness Log10 Liquid
9. Cloud Optical Thickness Log10 Total
10. Cloud Optical Thickness PCL Total
11. Cloud Optical Thickness Total
12. Cloud Particle Size Ice
13. Cloud Particle Size Liquid
14. Cloud Retrieval Fraction Ice
15. Cloud Retrieval Fraction Liquid
16. Cloud Retrieval Fraction Total
17. Cloud Top Pressure
18. Cloud Water Path Ice
19. Cloud Water Path Liquid
20. Sensor Azimuth
21. Sensor Zenith
22. Solar Azimuth
23. Solar Zenith

Each science parameter includes a combination of the following statistical variables:

• Mean, Standard Deviation, Pixel Count, Sum, and Sum Squares • Joint histogram versus Cloud Particle Size • Joint histogram versus Cloud Top Pressure

Consult the MCD06COSP User Guide for details regarding how the L3 monthly statistics are computed, and to learn more about the gridding and sampling protocols specific to this product and a number of other topics germane to the user community. Each calendar year includes 12 granules. The version 6.2 collection of this product starts from July 4, 2002 (Day-of-year 185) and currently is produced through September 2, 2021 (Day-of-year 245).

1. Webb et al., (2017) The Cloud Feedback Model Intercomparison Project (CFMIP) contribution to CMIP6. Geoscientific Model Development, Vol. 10 (1), 359 – 384.