I noticed that I get negative TEC values on some IGS stations when I convert the pseudoranges into slant TEC using simple formula sTEC = 2.852*(P2-C1)/speed_of_light. The problem appears on those IGS sites which use Trimble receivers. For example, I looked at the two IGS stations closely located, somewhere in Kazakhstan. One (called chum) uses Trimble receiver and another one (called pol2) uses Ashtech receiver. Both stations give similar-shaped TEC curves but the TEC from Trimble station is offset down by about 150 TEC units thus giving hugely negative values. Does anyone know why? Is this particular to Trimble receivers? Is there a correction for that?
- 13 Sep 2010
If you answer a question - or have a question you asked answered by someone - please remember to edit the page and set the status to answered. The status is in a drop-down list below the edit box.
The simple formula that you are using does not account for the interfrequency biases in the transmitter and receiver. These biases arise from a relative time delay between the signals as they travel through the transmitter or receiver (from computer to antenna). It sounds like the Trimble and Ashtech receivers that you are working with have very different receiver biases. The specifics of the system architectures cause these differences. A 150 TECu offset between the receivers corresponds to ~53nsec relative delay between the Ashtech and the Trimble. This is large, but not horrible. The fact that the Trimble receivers are usually producing negative TECs is not a problem, just a feature of this Trimble. In the past, we have not noticed a tendency for negative biases in any particular receiver brand, but then we have never looked. The general assumption in the ionospheric community is that many receivers will have a negative bias. Ideally, if we can't have perfect equipment (a bias of 0 nsec, which is impossible), then we would like for half of the receivers to have negative biases.
As far as dealing with the receiver biases, the solution depends upon your needs. If you trying to create a global ionospheric reconstruction, then it is best to solve for the satellite and receiver biases with the ionosphere in some form of Kalman filter or similar solver. If you are only interested in the local phenomena, work with the relative TEC (basically you subtract off the TEC at the highest elevation so that the observed TEC runs from 0 TECu and up). You can also look for another nearby ionospheric data source (I believe there is an ISR in or near Kazakhstan) and force your vertical TEC to agree with other data source. This will still not give you the true absolute TEC since you are calibrating the receiver's TEC which includes all of the plasma to 22000 km to an ionospheric TEC measurement that is at best good to about 1500 km. If you have control of the receivers there are techniques for calibrating out the bias, but this calibration needs to be done at multiple times per days (thermal expansion of the hardware is a known contributor to receiver biases).
- Negative TEC measurements are caused by the receiver biases.
- These biases are normal and expected. A ~53nsec difference between two nearby receiversis large, but not unreasonable.
- The receiver biases can be solved for in Kalman filters or removed by working with relative TEC.
Please let me know if this answer is confusing and/or incomplete.
- 13 Sep 2010
Dear Trevor, Thank you for the response. I thought that the interfrequency bias should not exceed 10-20 TECu. The bias of over 100 TEC units really surprised me. Do you know how consistent the interfrequency biases are? For example, if I calibrate one of Trimble NetR5?
receivers using an ISR or smth -- could I expect other Trimble NetR5?
receivers to have the same bias? Also, is there any information available on the biases of various IGS stations? If yes, I could probably calibrate some of the Trimble receivers using nearby IGS stations. Best regards, Dimitry
- 14 Sep 2010