I was looking into GNSS derived TEC data and most of the data were missing and repeatedly occuring. What is the most precise way to handle those missing data? I guess interpolation is not good enough to perform for a large number of missing data.
There are a lot of websites with GNSS data,
It is better to select periods with good continuous data sets
sincerely
Christine
What data are you working with? Generally, data from GNSS receivers are largely continuous with few significant interruptions (you might have short interruptions due to scintillation or multipath), once satellites come into view (any satellite will only remain within view for 2-5 hours, depending on your location and the constellation of choice). Are you interested in processed TEC data or processing TEC from GNSS data yourself (say from RINEX files)? If the former, the Madrigal database maintains a decent assortment of processed data for a fairly large network of receivers (http://cedar.openmadrigal.org/). This includes both processed data from individual receivers and binned TEC map data. If you want to process GNSS data from your own receiver or some RINEX files you have handy, there are a number of good papers in the community that go over that task including many codes available in an assortment of different coding languages. Feel free to take a look through one of my old papers that gives a decent overview of the processing steps involved (Article Determining Receiver Biases in GPS-derived Total Electron Co...
).Cheers,
David
Thank you, Christine Amory-Mazaudier, David R. Themens, Jens Kleb, and Yenca, for your answer. It really helped a lot. I hope to get similar suggestions in a future discussion.
Hourly data is enough to examine the TEC variation. If hourly is also missing, you can use cubic spline interpolation
Use any of machine learning technique:
https://towardsdatascience.com/7-ways-to-handle-missing-values-in-machine-learning-1a6326adf79e
I would recommend Global Ionosphere maps (https://cddis.nasa.gov/archive/gnss/products/ionex/). There are almost no data gaps. If you need time resolution of 15 min then use UQRG prouduct from UPC.
You may check here:
https://arplsrv.ictp.it
https://cddis.nasa.gov/archive/gnss/products/ionex/
Hi Ashok Silwal
Please find a very useful link related to your interesting question. You can download the data of GNSS-TEC from the following website.
1. https://github.com/gnss-lab/gnss-tec
Best Regards,
SHA
- Badges
- Science topic
- Similar topics
- Engineering
- Control Theory
- Control Systems
- Control Systems Engineering