MODIS

1
Aqua Summary
Terra Summary
MODIS Geolocation Status

• Geolocation
• Accuracy
• Goal
• MODIS science team has developed remote sensing
  algorithms to derive global time-series data
  products on various terrestrial geophysical
  parameters
• Highly accurate geolocation are required to
  support MODIS change detection and retrieval of
  biophysical parameters over heterogeneous land
  surfaces
• Accuracy Goal is 0.05
• of a 1 km Pixel (1s)

• Good results C4 RMS error is better than goal
  in track direction but does not quite meet goal
  in scan direction (but much better than
  specification 150 m)
• Early post-launch coordinate system issue
  resolved before C4
• Definitive ephemeris is used for best results
  causes up to 24 hr processing delay

• Excellent results Collection 4 (C4) Root Mean
  Square (RMS) error in nadir equivalent units is
  better than accuracy goal
• Small remaining northern/southern hemisphere
  difference
• Large errors occur after orbit maneuvers (about 6
  per year)
• accuracy in following orbit suspect

Robert E. Wolfe Raytheon ITSS _at_ NASA Goddard
Space Flight Center, Code 922Greenbelt, MD, USA
Mash Nishihama Raytheon ITSS _at_NASA Goddard Space
Flight Center, Code 922, Greenbelt, MD, USA
James A. Kuyper SAIC/GSC _at_NASA Goddard Space
Flight Center, Code 922, Greenbelt, MD, USA
Terra Daily Ground Control Point (GCP) Residuals
for C4 (nadir equivalent units)
Terra Long-Term Trend (LTT)
daily mean (dots) and standard deviation (bars)

• Long-term linear and annual cyclic trend removed
  in C4
• Original coefficients derived in late 2002 were
  updated in early 2004 to reflect overall trend
  (updated coefs. only in forward processing)

Terra Hemispheric Differences

• Different trends found in Northern and Southern
  hemisphere likely temperature related
• Global trend primarily follows Northern
  hemisphere because it contains 80 of the GCPs

Aqua C4 Results
Aqua Daily GCP Residuals (C4)
Terra 16-day GCP Residuals (C4)
16-day mean

• Overall scan residuals are 30 larger than Terra
• AMSR-E jitter contributes to some of this error
• some improvements expected with new mirror
  interpolation approach
• Early mission data needs special attention
• first 30 days is out of family
• larger than expected track error in first 150
  days may be due to less accurate ephemeris
• Removal of remaining star tracker anomaly (125
  msec readout time error in one star tracker)
  would help reach goal for future data

daily mean (dots) and standard deviation (bars)
Aqua LTT
Aqua 16-day GCP Residuals (C4)

• Track direction
• Linear trend smaller than Terra (10 vs. 15
  m/year)
• Small cyclic variation (10 m amplitude in
  Southern hemisphere)
• Scan direction
• Linear trend larger than Terra (20 vs. 4 m/year)
• Cyclic variation primarily in Southern hemisphere
  (opposite hemisphere from Terra) and larger than
  Terra (20 vs. 11 m amplitude)

16-day mean
Collection 5 Changes Summary
Acknowledgement
This work funded under NASA contract NAS5-32373
in the Terrestrial Information Systems Branch
(Code 922) at the Goddard Space Flight Center
Laboratory for Terrestrial Physics. July 13, 2004

• Ancillary data New BU Land/sea mask and Shuttle
  Radar Terrain Mission (SRTM) terrain data both
  under evaluation
• Improved mirror motion model (approved for
  remainder of C4)
• New GeolocationAccuracySuspect flag based on
  list of maneuver times to help exclude data
  near/after maneuvers from science production
• Update Terra and Aqua LTT and implement ability
  to use solar elevation angle as surrogate for
  temperature to remove within-orbit variation
  (North/South hemispheric differences)
• Refresh GCP library to remove some points and to
  obtain a more global distribution using Landsat-7
  GeoCover data set
• Consider reprocessing early mission definitive
  ephemeris data