GEWEX WORKING GROUP: CLOUDS

WGDMA TITLE

NASA Goddard Institute for Space Studies

NASA Goddard Space Flight Center

GEWEX DATA PRODUCT ASSESSMENT: CLOUDS

MEMBERSHIP

B. Baum (co-chair, NASA Langley Research Center), C. Stubenrauch (co-chair, Laboratoire de Meteorologie Dynamique), G. Campbell (Colorado State U.), A. Heidinger (NOAA NESDIS), T. Nakajima (Tokai U.), W. Rossow (ISCCP representative, NASA GISS)

PLAN OF WORK

MEETING CALENDAR

1st Clouds Assessment Workshop, 5-7 April, 2005	Space Science Engineering Center, U. Wisconsin, Madison, Wisconsin, USA (See report.)
2nd Clouds Assessment Workshop, 6-7 July, 2006	Space Science Engineering Center, U. Wisconsin, Madison, Wisconsin, USA See the GEWEX Cloud Climatology Intercomparison Workshop Web page.

DATASETS AND RESULTS

	MEAN ANNUAL TOTAL CLOUD AMOUNTS	MEAN SEASONAL CYCLE ANOMALY	DESEASONALIZED ANOMALY
ISCCP D2
SOBS
HIRS Wisconsin
TOVS Path-B
HIRS Pathfinder TOVS-A
TOVSB-3I
	MEAN ANNUAL TOTAL CLOUD AMOUNTS	MEAN SEASONAL CYCLE ANOMALY	DESEASONALIZED ANOMALY

Note: Cloud amount climatologies based on the PATMOS-X data set (Andy Heidinger) will be posted in this Web page.

DATA COMPARISONS

Global Mean and Seasonal Amplitude Differences

Deseasonalized Slopes Difference and Correlation Coefficients

Whole Record Mean Slopes Difference and Correlation Coefficients

Regional Means and Seasonal Amplitudes

ISCCP D2

The International Satellite Cloud Climatology Project (ISCCP) cloud dataset is produced from an analysis of imaging radiometers on an international constellation of weather satellites. The image pixels are sampled at 30 km and 3 hr intervals. Cloud detection is performed in three steps: (1) analysis of the space and time variations of infrared and visible wavelength radiances, separately, obtains several different estimates of the clear sky radiances for each location and time, (2) comparison of these estimates produces a final best estimate, and (3) comparison of the original radiances to the clear sky estimates, separately for infrared and visible wavelengths, indicates the presence of clouds if the difference exceeds regionally dependent threshold values. Cloud amount in a single pixel is zero or 100%; cloud amount in each 280 km equal-area map grid is the fraction of total pixels that are declared to be cloudy. Clear radiances are used to retrieve surface temperature and visible reflectance and cloudy radiances are used to retrieve cloud top temperature/pressure and visible optical thickness. Cloud types are determined by ranges of these cloud properties. The preferred results are based on the infrared-visible analysis; a parallel infrared-only analysis is also conducted. The difference between these two analyses during daytime are used to interpolate the daytime results over nighttime. The final products are produced by combining the cloud analysis results with the atmospheric and surface ancillary information used; global results are available for the period July 1983 – December 2004 at intervals of 30 km, 3 hr (DX data), 280 km, 3 hr (D1 data) and 280 km, monthly (D2 data presented here). The following links provide access to obtain the full documentation for ISCCP datasets and to download the complete D2 dataset.

Selected References

Selected Cloud Amounts from the ISCCP D2 dataset are available here, along with climatological results: mean annual map, mean seasonal cycle, deseasonalized global monthly anomaly. Your selection can be viewed as GIF images or downloaded to your local disk (see browse data format description and sample FORTRAN read subroutines). Data presented here can also be obtained via anonymous ftp through our ftp site.

sample img

ISCCP-D2 Anomaly plots

Global Monthly/Seasonal Mean Maps

Select a variable:

Select a time period.
Enter Monthly Means (YYMM) from Jul. 1983 to Dec. 2004.

Now you may a GIF image,
Or select a format and

SOBS

The Surface Observations (SOBS) cloud dataset is produced from an analysis of the visual observation made by surface weather observers at international collection of land surface weather stations and ships. These results have been edited to remove nighttime observations lacking sufficient lunar illumination. Observations are made every 3 hr. Total sky cover fraction by all clouds and the types and amounts of standard morphological clouds are reported. Cloud base height above local topography for low-level clouds is also estimated. Total cloud amount for an observation, representing an area about 40-80 km across, is reported in octas, each octa being 12.5%. Observers are instructed to report one octa if any cloud is seen, no matter how small an amount, and to report seven octas if any clear sky is seen, no matter how small an amount. Cloud type amounts have been corrected for obscuration effects. The final products are produced by mapping the observations into a common grid, with an interval of 2.5° over land and 10° over ocean and combining some rarer cloud types to reduce the number of categories. Results are generally not available for the southern midlatitude oceans, especially in wintertime and over some sparsely populated land areas in Africa and South America. "Global" results are available for the period 1971 - 1996 at monthly intervals. The following links provide access to obtain the full documentation for SOBS datasets and to download the complete SOBS monthly dataset.

Selected References

Selected Cloud Amounts from the SOBS dataset are available here, along with climatological results: mean annual map, mean seasonal cycle, deseasonalized global monthly anomaly. Your selection can be viewed as GIF images or downloaded to your local disk (see browse data format description and sample FORTRAN read subroutines). Data presented here can also be obtained via anonymous ftp through our ftp site.

sample img

SOBS Anomaly plots

Global Monthly/Seasonal Mean Maps

HIRS Wisconsin

The HIRS-Wisconsin cloud dataset is produced from an analysis of infrared sounding radiometers on the NOAA polar orbiting weather satellites. Measurements are made at about 17 km intervals, twice daily from each operating polar orbiter (usually there are two). Cloud detection is performed in two steps: (1) using an estimate of surface air temperatures taken from operational weather forecast model analysis, some pixels are identified as clear by a threshold test on the “window” channel at about 10 ?m wavelength, (2) radiance ratios among the five sounding channels from available clear sky pixels (if none are available, they are interpolated from nearby pixels) are compared with other observed radiance ratios to detect the presence of clouds by the deviation of the observed values from the clear sky estimates (the CO2 slicing method). Cloud amount in a single pixel is zero or 100%; cloud amount in each 2.5? map grid is the fraction of total pixels that are declared to be cloudy. The radiance ratios also provide an “effective cloud amount”, physically the product of cloud cover fraction and infrared emissivity and the top pressure for a single cloud layer in each map grid. Cloud types are determined by ranges of the cloud top pressures. The final products are available for the period 1979 – 2003. The following links provide access to obtain the full documentation for HIRS-Wisconsin dataset and to download the complete HIRS-Wisconsin dataset.

Selected References

Selected Cloud Amounts from the HIRS-Wisconsin dataset are available here, along with climatological results: mean annual map, mean seasonal cycle, deseasonalized global monthly anomaly. Your selection can be viewed as GIF images or downloaded to your local disk (see browse data format description and sample FORTRAN read subroutines). Data presented here can also be obtained via anonymous ftp through our ftp site.

sample img

HIRS Wisconsin Anomaly plots

Global Monthly/Seasonal Mean Maps

Select a variable:

Select a time period.
Enter Monthly Means (YYMM) from Jan. 1979 to Dec. 2001.

Now you may a GIF image,
Or select a format and

TOVS Path-B

The TOVS Path-B cloud dataset is produced from an analysis of infrared and microwave sounding radiometers on the NOAA polar orbiting weather satellites. The infrared radiances are sampled at 17 km intervals and the microwave radiances are sampled at about 100 km intervals, twice daily from each operating satellite (usually there are two). Cloud detection is performed in three steps: (1) the microwave radiances (after a simple check to eliminate any effects of heavy precipitation) are used in a fast radiative transfer code to predict the clear sky infrared radiance values, (2) deviations of the observed infrared radiances from these predictions by a sufficient amount are used to identify cloudy pixels, (3) additional spectral threshold tests are performed on the infrared radiances, including a kind of spatial coherence test to enhance detection of low-level clouds, are performed to detect cloudy pixels. Cloud amount in a single pixel is zero or 100%; cloud amount in each 1.0 map grid is the fraction of total pixels that are declared to be cloudy. The average clear radiances are used to retrieve surface temperature and atmospheric temperature/humidity profiles. Average cloudy radiances are used to retrieve cloud top temperature/pressure and infrared emssivity. The product of cloud amount and emissivity is also reported as an "effective" cloud amount. Cloud types are determined by ranges of these cloud properties. The final product is available for the period 1987 - 1995 at twice daily to monthly intervals. The following links provide access to obtain the full documentation for TOVS Path-B dataset and to download monthly mean cloud products from TOVSB dataset. (Click on HFTP and login as anonymous and click on directory claudia and then on TOVSB.)

Selected References

Selected Cloud Amounts from the TOVS Path-B dataset are available here, along with climatological results. Your selection can be viewed as GIF images or downloaded to your local disk (see browse data format description and sample FORTRAN read subroutines). Data presented here can also be obtained via anonymous ftp through our ftp site.

sample img

TOVS Path-B Anomaly plots

Global Monthly/Seasonal Mean Maps

HIRS Pathfinder

The HIRS-Pathfinder cloud dataset is produced from an analysis of infrared sounding radiometers on the NOAA polar orbiting weather satellites. Infrared measurements are made at about 17 km intervals, twice daily from each operating polar orbiter (usually there are two). Cloud detection is performed in two steps: (1) the two "warmest" pixels (largest "window" radiances) are selected within a 2.5° region and the ratios of the five sounding channels used to estimate clear radiances, (2) deviations of the infrared sounding channel radiances from the clear sky values are used to detect the presence of clouds. An "effective cloud amount", physically the product of cloud cover fraction and infrared emissivity, for each 2.5° map grid is determined from the difference of the average radiances and the clear sky values. The final products are available for the period 1979 - 2002. The following links provide access to obtain the full documentation for HIRS-Pathfinder dataset and to download the complete HIRS-Pathfinder dataset.

Selected References

Selected Cloud Amounts from the HIRS-Pathfinder dataset are available here, along with climatological results: mean annual map, mean seasonal cycle, deseasonalized global monthly anomaly. Your selection can be viewed as GIF images or downloaded to your local disk (see browse data format description and sample FORTRAN read subroutines). Data presented here can also be obtained via anonymous ftp through our ftp site.

sample img

HIRS Pathfinder (TOVS-A) Anomaly plots

Global Monthly/Seasonal Mean Maps

TOVSB-3I

This dataset contains the TIROS Operational Vertical Sounder (TOVS) level 3 geophysical parameters derived using the classification method of Chedin and Scott (1985) and processed by the Atmopheric Radiation Analysis (ARA) group of the Laboratoire de Meteorologie Dynamique (LMD) du CNRS (France). This method, which is a priori data-dependent but model-independent, was designated as the Path B scheme by the TOVS Pathfinder Science Working Group. Two of the three sounding instruments aboard the NOAA-xx series of Polar Orbiting Satellites are used to produce global, 1 degree by 1 degree fields of the 3-dimensional temperature-moisture structure of the atmosphere, namely, the 20 channel High resolution Infrared Radiation Sounder 2 (HIRS2) and the 4 channel Microwave Sounding Unit (MSU). In addition to profiles of temperature and moisture, the HIRS2/MSU radiances are used by the Path B methodology to derive surface skin temperature, surface microwave emissivity, cloud fraction, and cloudtop height and temperature. The basic premise behind the Path B technique involves deriving geophysical parameters from TOVS radiance data in a manner which is independent of the output from any hydrodynamic model forecast or analysis. It is based upon the Improved Initialization Inversion (3I) algorithm of Chedin et al.(1985), which relies on a pattern recognition approach to relate cloud-corrected radiances to geophysical parameters. The following links provide access to obtain the full documentation for TOVS Path-B dataset and to download HDF monthly mean products from TOVSB-3I dataset.

Selected References

Selected Cloud Amounts from the TOVSB-3I dataset for the period Jan. 1981 to Jun. 1986 are available here, along with climatological results. Your selection can be viewed as GIF images or downloaded to your local disk (see browse data format description and sample FORTRAN read subroutines). Data presented here can also be obtained via anonymous ftp through our ftp site.

sample img

TOVSB-3I Anomaly plots

Global Monthly/Seasonal Mean Maps

Select a variable:

Select a time period.
Enter Monthly Means (YYMM) from Jan. 1987 to Jun. 1995.

Now you may a GIF image,
Or select a format and

Back to GEWEX and Other Global, Long-Term Datasets

WGDMA Webmaster: WGDMA Webmaster