---

The goal of this page is to present results demonstrating the performance of CLAVR-x cloud mask over the diurnal cycle.  The original CLAVR-1 used  single value thresholds for many cloud tests that were derived for the AVHRR's in the afternoon orbits.  Consequently, the CLAVR-1 cloud mask was found to not perform well on the AVHRR data from the morning orbits and this data was excluded from PATMOS.  The explicit inclusion of solar zenith angle dependence in many of cloud tests in CLAVR-x (RGCT, C3AT, TMFT) should mitigate the sensitivity of the cloud mask performance to particular AVHRR orbit characteristics.  Inclusion of solar zenith angle dependence should also reduce the dependence on solar zenith angle observed in PATMOS as the orbits of the afternoon satellites moved to later equator crossing times.  

The results presented in this page address the apparent sensitivity to the diurnal cycle of the cloud mask, based on the derived cloud mask and clear radiances.  The data used are the global GAC AVHRR data from July 1995 from NOAA-14 and NOAA-12.  In addition, the ISCCP daily-averaged cloud amount for the same period is presented.  The resolution of all data presented here is 2.5 degrees equal-area (the ISCCP grid). The CLAVR-x data is processing into ascending and descending fields for each satellite - giving four fields spaced throughout the diurnal cycle. Each global field therefore is look at the earth at roughly the same local time. The local times of these fields is as follows.



Images of Global Fields

	NOAA-14 Des	NOAA-12 Asc	NOAA-14 Asc	NOAA-12 Des
SST	view	view	view	view
SST Std. Dev.	view	view	view	view
Total Cloud Amount	view	view	view	view
Water Cloud Amount	view	view	view	view
Ice Cloud Amount	view	view	view	view

Zonal Profiles
Clear ch4 (11 µm)  radiance over the ocean (view)
Mean Standard Deviation (within individual grid-cells) of the Clear Ch4 Radiance over the Ocean (view)

ISCCP Cloud Amount Comparison with CLAVR-x Diurnal Cycle
Mean July 1995 Daily Total Cloud Amount from ISCCP (view)
Mean July 1995 Daily Total Cloud Amount from all four CLAVR-x fields (view)
Difference between ISCCP and CLAVR-x Daily Fields (view)
Zonal Comparison  (view)


Conclusions from these Results

• the images of the fields from the different local times show no obvious erroneous dependence on the diurnal cycle
• the zonal profiles of clear ocean ch4 (11 µm) radiances show no significant dependence on diurnal cycle - as would expected for the ocean.  The curves group more with the particular satellite than with equator crossing time - hinting that the calibration biases are larger the cloud contamination differences.
• Assuming the standard deviation of the clear ch4 radiance within a grid-cell is indicative of cloud contamination, there appears to generally be no sensitivity of the cloud contamination of the clear ch4 radiance over the ocean to the diurnal cycle.  Only for NOAA-14 ASC at southern latitudes and NOAA-14 DES at northern latitudes (where the terminator coincides with cold oceans) does there appear to be higher standard deviations - this is expected.
• The global cloud distribution is somewhat stable - whether the inferred diurnal cycle agrees with other measurements is to be determined.
• The zonal profiles all show features similar to the ISCCP zonal profile, but none of the zonal profiles from CLAVR-x exceeds the ISCCP cloud amount.  This indicates that the differences between the ISCCP and CLAVR-x cloud amounts are not driven by differences in diurnal sampling.
  

---

NESDIS/Satellite Applications and Research Home | CIMSS Home | SSEC Home