SATELLITE
INTERPRETATION
DISCUSSION
NOAA/NESDIS
 Regional and Mesoscale Meteorology Team
Cooperative Institute for Research in the Atmosphere (CIRA)
Colorado State University   Fort Collins, Colorado

February 9, 2000
GOES Visible Imagery: Sensor Degradation over Time

click on image to enlarge
When GOES-9/West was launched in May 1995, the performance of the GOES-9 visible sensor was found to be significantly better than the performance of the GOES-8/East visible sensor, launched in April 1994.  Initial study indicated that the sensitivity of the GOES-9 visible sensor was 15% greater (Knapp and Vonder Haar 2000, J. Atm. Ocean. Tech, accepted).   Later studies showed that the sensitivity of the GOES-8 visible sensor was decreasing at an annual rate ranging from 4.5 to 7.6% (ibid).  With the failure of GOES-9 and the subsequent use of GOES-10/West, launched in April 1997, GOES-10 visible imagery appears noticeably brighter than GOES-8 visible imagery over the same location.  This brightness difference is examined below by comparing GOES-8 and -10 scaled radiances at solar noon along the 105W meridian, equidistant between the sub-points of the GOES-8 and -10 satellites (75W and 135W, respectively).  The use of imagery from solar noon minimizes differences in sun-earth-satellite geometry. Note that scaled radiances are linearly related to satellite-measured counts (reference), and may be considered approximate visible albedos (ranging between 0 and 100%) without corrections for solar zenith angle and Raleigh scattering.

 

click on image to start animation

 xxx This animation sequence shows four 1-km visible images centered on (40N, 105W).  GOES-8 [GOES-10] images are from 1901 [1900] UTC on February 6, 2000.  The numbers 1-5 plotted in images 3 and 4 indicate the locations, ranging between (42N, 105W) and (38N, 105W), where albedos were measured over a 31 km x 31 km area.  Because images 3 and 4 are referenced to a common point at sea level (well below the earth's surface over Colorado), a shift is observed between surface and cloud features in frames 3 and 4.  This effect is called parallax.  Due to parallax, albedos from GOES-8 and -10 may not be exactly collocated, which may increase or decrease the difference between GOES-8 and -10 measurements, but is not considered further.

Visible imagery from GOES-8 appears noticeably darker.  GOES-8 albedos are 17 and 34% less than GOES-10 albedos (calculated (G-10 - G-8) / G-10; see table below).  Albedo differences of this magnitude were expected due to the decrease in visible sensor sensitivity over time, combined with the fact that GOES-8 is three years older than GOES-10. 

Research and operational meteorologists should be aware of the differing performance between the GOES-8 and -10 visible sensors.  Research meteorologists should correct for sensor degradation when using GOES-8 and -10 visible imagery in a quantitative manner.  Operational meteorologists may increase the brightness of GOES-8 visible imagery by either modifying enhancement tables or stretching the data. 


ALBEDO COMPARISON BETWEEN GOES-8 AND GOES-10 
 
Location Latitude/Longitude G-8 Albedo G-10 Albedo % Difference
1 42N,105W 9.2% 12.5% 26%
2 41N,105W 9.3% 13.2% 30%
3 40N,105W 15.4% 23.4% 34%
4 39N,105W 8.9% 12.4% 28%
5 38N,105W 7.8% 9.4% 17%

Back to top

We welcome your comments and discussion at ramsdis@comet.ucar.edu

Previous RAMMT Satellite Interpretation Discussions
 
 

Information Contact: Lewis Grasso
CIRA/RAMM WebMaster: Roger Phillips
Author: Bard Zajac
Last Updated: February 9, 2000