On this page: Temporal Coverage | Spatial Coverage | Levels | Update Schedule | Download/Plot Data | Analysis Tools
Restrictions | Details | Caveats | File Naming | Citation | References | Original Source | Contact

NOAA Highly Reflective Clouds

Brief Description:

  • NOAA Highly Reflective Clouds for the Global Tropics

Temporal Coverage:

  • Long term monthly statistics, derived from data for years 1971/01 - 1985/12.
  • Monthly statistics, derived from data for years 1971/01 - 1985/12.
  • Daily values for 1971/01/01 - 1988/01/31.

Spatial Coverage:

  • 1.0 degree latitude x 1.0 degree longitude global grid (51x360).
  • 25.0N - 25.0S, 1.0E - 360.0E.


  • Entire atmosphere considered as a single layer.

Update Schedule:

  • Static

Download/Plot Data: (FTP Access)

Variable Statistic Level Download File Create Plot/Subset
Highly Reflective Clouds Monthly Long Term Normalized Mean Standard Deviation Atmosphere hrc.ltmstddev.nc plot
Highly Reflective Clouds Monthly Long Term Normalized Mean Atmosphere hrc.normltm.nc plot
Highly Reflective Clouds Monthly Normalized Mean Atmosphere hrc.norm_monthlies.nc plot
Highly Reflective Clouds Monthly Missing Days Atmosphere hrc.nmissdays.nc plot
Highly Reflective Clouds Monthly Mean Atmosphere hrc.monthlies.nc plot
Highly Reflective Clouds Daily Value Atmosphere hrc.dailies.nc plot

Usage Restrictions:

  • None

Detailed Description:

  • The data presented here are the record of subjectively identified areas of large-scale organized convection over the global tropics on the daily daytime Mercator projection mosaics for a 15-year period from 1971 to 1988. These convection areas appear as highly reflective clouds in the visible-range mosaics and as white areas in the infrared (IR) range. Highly reflective cloud (HRC) is defined here as a deep, organized tropical convection system extending at least 200 km horizontally. HRCs are composed of many individual convective cells embedded within a common cirrostratus canopy. These cloud systems, commonly known as cloud clusters, have been extensively studied in recent years (e.g., Houze, 1982); they are responsible for most tropical rainfall and are important components in the general circulation of the atmosphere. These data originally appeared in the Atlas of Highly Reflective Clouds for the Global Tropics: 1971-1985.

    Until now, the most widely used indicator of large-scale convection over the global tropics has been the outgoing longwave radiation (OLR) data set (Gruber and Krueger, 1984). OLR data for 9 years were recently summarized in NOAA Atlas No. 6 (Janowiak et al., 1985). In the tropics, low values of OLR are found over areas covered by high clouds with low cloud-top temperatures. In most cases, such clouds are convective. However, low OLR values are also associated with nonconvective clouds such as cirrostratus plumes. In contrast to the OLR data, the HRC data presented here deliberately exclude nonconvective cold clouds.

    The original idea for the creation of the HRC data set came from a study conducted at the University of Hawaii by Kilonsky and Ramage (1976). The study was based on the assumption that most tropical rainfall occurs in organized convective systems (cloud clusters), which appear as HRC in the polar-orbiting satellite picture mosaics. Kilonsky and Ramage (KR) obtained monthly rainfall data for Pacific atoll stations assumed to be representative of open ocean conditions. They correlated these data with the number of days per month having HRC. The presence of HRC was noted in grid squares of the satellite picture mosaics. KR assumed a linear relationship between the number of days with HRC at a particular location and the amount of rainfall recorded there; a linear regression equation was thus derived, linking the number of days with HRC cover (the independent variable) to the observed rainfall (the dependent variable). The correlation coefficient between the two variables was 0.75. Both the regression relationship and the correlation between the two variables were found to be significant at the 1% level.

    Subsequent to the original work of Kilonsky and Ramage, Garcia (1981) used HRC data and the KR technique to produce rainfall estimates for the tropical Atlantic Ocean. The estimates proved to be reasonably close to those obtained by using more sophisticated geostationary satellite techniques and to those estimated from shipboard radar data. These results provided the impetus for extending the HRC data set to the entire tropical belt, including land areas, and for updating the HRC data continually.

    Note that the validity of using HRC data from the daytime visible and IR Mercator projection mosaics as a means of estimating monthly rainfall amounts is greatly affected by the mean diurnal cycle of convection over different regions of the tropics. There is some evidence (e.g., Griffith et al., 1980) that convection over land areas has a fairly pronounced maximum during the late evening hours, whereas convection over oceans is more evenly distributed throughout the 24-hour cycle. Thus, a rainfall estimation scheme (such as the KR technique) that samples conditions only once a day during daylight hours will tend to underestimate rainfall over land areas. In addition, changes in the time of passage of polar-orbiting satellites can significantly change cumulative amounts of convection observed in a given area.


  • None

Related File Naming & Structure Information:

File Names:

  • hrc.stat | time.nc   (In directory: /Datasets/noaa_hrc/)

Dataset Format and Size:

  • PSL standard NetCDF 444 Kbyte file for each LTM file .
  • PSL standard NetCDF 7 Mbyte file for each monthly file .
  • PSL standard NetCDF 458 Mbyte file for each daily file .

Missing Data:

  • Missing data is flagged with a value of 32767s.
  • The number of missing days in the monthly accumlations is recorded in the file hrc.nmissdays.nc.


  • Please note: If you acquire NOAA_HRC data products from PSL, we ask that you acknowledge us in your use of the data. This may be done by including text such as NOAA_HRC data provided by the NOAA/OAR/ESRL PSL, Boulder, Colorado, USA, from their Web site at / in any documents or publications using these data. We would also appreciate receiving a copy of the relevant publications. This will help PSL to justify keeping the NOAA_HRC data set freely available online in the future. Thank you!


Original Source:


  • Physical Sciences Laboratory: Data Management
    325 Broadway
    Boulder, CO 80305-3328