SeaWiFS Navigation Anomaly
Posted: Mon Mar 13, 2006 12:42 pm America/New_York
We are currently experiencing a degraded navigational accuracy with SeaWiFS data as a result of the spacecraft's inability to track GPS satellites. OrbImage reported that on "Friday March 3, 2006 (day 62) at 08:46:00 UTC, the OV-2 GPS receiver went through a hard reset. Upon the reset, the GPS almanac became corrupted. The receiver is currently tracking zero GPS satellites and the almanac status is set to "bad".
OrbImage has been working the problem but unfortunately, the problem was not able to be resolved prior to the known March GPS Anomaly period which started yesterday 3/12/06 and ends 3/26/06. That means that the current situation will remain in effect until 3/27/06 at which time it is anticipated that teh spacecraft will be ableto successfully track GPS satellites again
Fred Patt, who wrote the navigation software that we use in processing SeaWiFS data has summarized the current situation as of this morning as follows:
"As of last night's midnight dump, the estimated navigation error is about -0.24 degree in latitude, or about 27 km along track.
There are two main contributors to this error: orbit propagation, and spacecraft time code. The orbit propagation error results from our model running without GPS updates for the past 10 days. My estimate of this is about 15 km, increasing monotonically. This component of the error has been steadily increasing since the GPS tracking was lost early on Day 62. It should be mostly recoverable once GPS tracking is restored.
The spacecraft time code has accumulated errors in random steps during the past 10 days. We have observed numerous time code jumps in the GAC data. It is very likely that other jumps have occurred in the back-orbit. So far we have observed short bursts of time code jumps on Days 62, 65, 69, 70, 71 and 72. The accumulated time code errors were negative immediately after the start of the GPS outage, but now appear to be positive (based on the current navigation errors). There have been several jumps since the noon passes yesterday (Day 71), resulting from GPS trying to reset itself.
The time jumps are by far the more difficult effect to deal with, since: 1) we have incomplete information about their timing and magnitude, and therefore have to estimate the size of the time code errors and when they changed; 2) they are manifested differently in different spacecraft subsubsystems; specifically, the ACS time code jumps that have been observed have not tracked those in the minor frame time code; 3) correcting these jumps is largely a manual process, and will most likely need to be iterated by reprocessing the data, re-estimating the errors, making additional corrections, etc.
We have now entered the window of vulnerability for the annual Vernal Equinox GPS anomaly. For a detailed description and analysis of this anomaly, please refer to our postlaunch TM volume 28, section 4.4.3:
http://oceancolor.gsfc.nasa.gov/SeaWiFS ... LVol28.pdf
Because of this, we may not recover GPS tracking until the start of Day 85. In the meantime, we can expect to see continued accumulation of orbit propagation errors and time code jumps."
We believe that we should be able to retrospectively correct the navigational errors once the spacecraft returns to normal operations. The only impact that this has on the current SeaWiFS data is that we will refrain from producing any refined SeaWiFS products satrting with data collected on 3 March 2006. Since there is a 14-day embargo on the most recently collected data, the impact on the availability of the science data products should be minimal.
I will keep you posted in any change to the anticipated plan as stated above.
Regards,
gene
OrbImage has been working the problem but unfortunately, the problem was not able to be resolved prior to the known March GPS Anomaly period which started yesterday 3/12/06 and ends 3/26/06. That means that the current situation will remain in effect until 3/27/06 at which time it is anticipated that teh spacecraft will be ableto successfully track GPS satellites again
Fred Patt, who wrote the navigation software that we use in processing SeaWiFS data has summarized the current situation as of this morning as follows:
"As of last night's midnight dump, the estimated navigation error is about -0.24 degree in latitude, or about 27 km along track.
There are two main contributors to this error: orbit propagation, and spacecraft time code. The orbit propagation error results from our model running without GPS updates for the past 10 days. My estimate of this is about 15 km, increasing monotonically. This component of the error has been steadily increasing since the GPS tracking was lost early on Day 62. It should be mostly recoverable once GPS tracking is restored.
The spacecraft time code has accumulated errors in random steps during the past 10 days. We have observed numerous time code jumps in the GAC data. It is very likely that other jumps have occurred in the back-orbit. So far we have observed short bursts of time code jumps on Days 62, 65, 69, 70, 71 and 72. The accumulated time code errors were negative immediately after the start of the GPS outage, but now appear to be positive (based on the current navigation errors). There have been several jumps since the noon passes yesterday (Day 71), resulting from GPS trying to reset itself.
The time jumps are by far the more difficult effect to deal with, since: 1) we have incomplete information about their timing and magnitude, and therefore have to estimate the size of the time code errors and when they changed; 2) they are manifested differently in different spacecraft subsubsystems; specifically, the ACS time code jumps that have been observed have not tracked those in the minor frame time code; 3) correcting these jumps is largely a manual process, and will most likely need to be iterated by reprocessing the data, re-estimating the errors, making additional corrections, etc.
We have now entered the window of vulnerability for the annual Vernal Equinox GPS anomaly. For a detailed description and analysis of this anomaly, please refer to our postlaunch TM volume 28, section 4.4.3:
http://oceancolor.gsfc.nasa.gov/SeaWiFS ... LVol28.pdf
Because of this, we may not recover GPS tracking until the start of Day 85. In the meantime, we can expect to see continued accumulation of orbit propagation errors and time code jumps."
We believe that we should be able to retrospectively correct the navigational errors once the spacecraft returns to normal operations. The only impact that this has on the current SeaWiFS data is that we will refrain from producing any refined SeaWiFS products satrting with data collected on 3 March 2006. Since there is a 14-day embargo on the most recently collected data, the impact on the availability of the science data products should be minimal.
I will keep you posted in any change to the anticipated plan as stated above.
Regards,
gene