to Track Turtles
Tracking sea turtles via satellite is a powerful tool, but it is easier said than done.
Three of the satellites (NOAA-11,14, and 15) have ARGOS instruments on board that can receive ultra-high frequency (UHF) messages sent from so-called satellite platforms on earth. As the satellite comes over the horizon and in view of a site on earth from which a message is being transmitted, say from one of our project's turtles in the Sulu Sea, it begins receiving the signal. The turtle's transmitter sends a new message about every 40 seconds. Each message contains the identification code of the transmitter and some technical data, like how deep the turtle's last dive was, how often she has dived since the last transmission, etc. The satellite continues to receive signals until it disappears over the opposite horizon. Such a satellite pass can take between 2 and 12 minutes, depending on how high above the horizon the satellite passes. The satellite stores all the information, and as soon as it comes within view of an earth station, it re-transmits or downloads it to the station. From there it reaches first the ARGOS system computers and then our project computers.
This animation of the ARGOS data transfer was produced by the BBC for another one of our projects.
(You need the VIVO plugin installed on your browser to view the movie)
But what determines
how many messages the satellite receives during a pass? Sea turtles live
in the ocean, and spend most of their time swimming, eating and resting
in the water. The transmitter's signals can only be received by the satellite
when the turtle is at the surface and the antenna is out of the water
and in the air. But turtles rarely come to the surface; they do so only
to breathe or maybe to float for a while as they rest or sun. For a good
quality location to be calculated, several things must all happen at once:
the turtle must be at the surface, the antenna must be in the air, and
the satellite must be above the horizon. All these have to take place
long enough for at least 4 messages to be received during one pass. Only
then can the computers calculate at least a blue quality location.
Tough chance! Even
when a turtle surfaces while a satellite is overhead, one small wave or
wave splash that smothers the tip of the antenna at just the wrong time
is enough to interrupt the transmission. It is not easy to get good locations
for a sea turtle!! That's why our tables with the raw data have so few
good locations. We received many more messages, but most of them had to
be thrown away because they were useless for plotting accurate positions
of the turtles.
But there are more
problems in getting signals from our turtles.
Here are some more details about the analysis of the satellite telemetry data.
Because the satellite is moving rapidly at about 26,640 km per hour, the frequency of the signals sent from the transmitter to the satellite is influenced by the Doppler shift. That is similar to what happens when a speeding train or car moves towards you and then passes by -- the sound suddenly changes from high-pitched to low-pitched. The ARGOS computers make use of the Doppler shift, by analyzing the frequency change of the incoming signals and, with a complex series of formulas, they calculate two predicted positions for the transmitter for each over pass. The position that is too far away to be possible then has to be discarded.
In general, the more messages the satellite has received during a single pass, the more accurately the computers can calculate the location of the transmitter. But even under the best of conditions, there is always some error in the prediction of the location. ARGOS calculates the predicted location along with the size of the error, called Location Class or "LC". LC 3 is the best, it means that the position that ARGOS has calculated should fall 95 out of 100 times inside a circle with a radius of 150m. For LC 2, the location should fall inside a circle with a radius of 350 m, and for LC 1, the circle will have a radius of 1000m or one kilometer.
At least four messages per pass must be received to achieve LC 3, LC 2 or LC 1 quality. On our maps we have indicated these high quality positions in red. If at least four messages are received, but for some reason the conditions are poor, the calculation may be LC 0. This means that a position is calculated, but there is no estimate for how big the error is.
If only 3 messages are received during a pass, a position can be calculated, but there will be no estimate of how big the error is: this is LC A.
When only 2 messages are received during a single pass, ARGOS can still calculate a position, but it is not possible to calculate how much error is involved: this is LC B. Finally, if only one message is received during a pass, LC Z is reported; there is no prediction of position.
On our map we coded the mediocre quality positions (LC 0) in blue; the poor quality positions are not shown. Likewise, in our analysis, we only use the red and blue locations. Only when we have no other information from the turtle at all do the poor quality positions give us at least an indication of where she might be.