I followed method suggested by G. Ma(2003) to estimate GPS receiver bias and calculated VTEC by removing Receiver Bias (Standard Deviation Method), it performs best but I get few values with negative VTEC at 30o elevation cut off.
Any suggestions relating to this will be helpful. Thanks.
I suspect your problem may be due to an underlying weakness in the Ma (2003) method, the assumption that the observed TEC values can be mapped onto a screen at some fixed altitude. This assumption works at mid-latitudes, but will break down as the equatorial ionosphere is approached. The source of this problem is the plasmasphere (or protonosphere), which adds an increasingly important component to TEC as the ionospheric equatorial anomalies (located nominally at +/-18 deg dip latitude) are approached. This will lead to errors in the bias estimated, which can lead to negative TEC values.
Ma (2003) suggests that this can be mitigated by first determining a receiver bias value by operating the receiver at a mid-latitude location, and then using this receiver bias as a fixed value after the receiver is relocated. I do not expect this method to work all that well. Studies have found that the bias at the receiver end of the link (which includes everything from the antenna through the cabling to the receiver) can be sensitive to what appear to be only slight changes in the setup (particularly in the cabling) and to the ambient temperature.
You might want to look into work done by Andrew Mazella on a method he calls SCORPION (I don't have a recent citation for this). It is similar to Ma's method, but it tries to take into account problems due to plasmaspheric contributions to TEC.
Respected James Secan,
Thank you for your valuable and insightful answer.
As you have mentioned in your answer, our observational site is located under the EIA crest region (geographic 29oN) which shows dynamic behavior during the course of the day. This will result in improper estimation of Receiver Bias (~80 TECU). Also thanks for suggestion of SCORPION method, I will go through it and try to resolve the problem.
Actually the problem arises after upgrading firmware on GSV4004B which is not there before.
Consider McCaffrey et al.(2017), "GPS receiver CODE bias estimation: A comparison of two methods" and the references within:
https://www.researchgate.net/publication/313592244_GPS_receiver_CODE_bias_estimation_A_comparison_of_two_methods
Article GPS receiver CODE bias estimation: A comparison of two methods
Mr. Alex Clay Cushley,
Thanks for your useful suggestion.
As mentioned in the McCaffrey et al.(2017) that 'The MSD technique is based on the assumption of a horizontally homogeneous ionosphere' which is not the case with our region. The day time TEC variations are dynamic which would be the possible reason for incorrect Bias estimation using MSD method.
The actual problem arises after upgrading the firmware, the TEC values were got shifted around +80 TECU. So we are trying to obtain exact bias value which can be removed in post processing.
Thank you.
Hi. Try to compute the bias using a period of day with more homogeneos ionosphere. Luke from 7 AM tô 5 PM local time.
Mr. Galera Monico,
It is a EIA crest region so it is difficult to find homogeneous ionosphere. For convenience you can see that from plot below of VTEC in local time 24h without any elevation threshold. So I think this is the reason that MSD method is giving overestimated bias values.
Did this 10 years ago but what I remember is that the receiver bias added to satellite bias was normally accurate. Would look if your satellite biases are good and constant over time and maybe use a separate region to calculate those satellite biases. Could also get those biases online back then. Did use a large set of receivers though and you may only be using 1 receiver?
- Badges
- Science topic
- Similar topics
- Geoscience
- Atmospheric Sciences
- Atmosphere