Regional Ionospheric Corrections for High Accuracy GNSS Positioning

Round 1

Reviewer 1 Report

This manuscript studies the relationship between the density of the GNSS receivers and the accuracy of linear interpolation of ionospheric delay in Australia. This is useful for high precision positioning. But the manuscript needs some minor modifications before it is accepted. The major comments are listed as follows:

• In section 2.2, equations are needed to explain the processing. For example, how to get the slant ionospheric delay corrections, how to remove the receiver hardware bias, how to define the accuracy of the linear interpolation model, and so on.
• How to get Equations (3) and (4) from Equations (1) and (2)?
• The receiver and satellite clock errors are not associated with the ionosphere. What do the ionosphere-free clock errors mean? Please note what are and  in the manuscript.
• In figure 3, the vertical axis is the accuracy of ionospheric delay. Is it defined as the differences between the interpolated ionospheric corrections and measured values at the evaluation CORS station? Is the mean accuracy equal to the mean deviation? Does “mean accuracy” in Figure 3 have the same meaning to “mean of accuracy” in Figure 4?
• In section 4, Kp is used to indicate the ionosoheric quiet day or ionospheric disturbed day. But in fact, Kp is an index of geomagnetic activity. So the results need to be checked.
• The centimeter level accuracy of linear interpolation can be achieved only under some conditions. The conditions should be added in conclusion.

Author Response

Response to Reviewer 1 Comments

We highly appreciate the referee for supplying valuable comments and suggestions to improve our paper. All your comments have been carefully examined, answered, and updated in the revised version.

Point 1:

• In section 2.2, equations are needed to explain the processing. For example, how to get the slant ionospheric delay corrections, how to remove the receiver hardware bias, how to define the accuracy of the linear interpolation model, and so on.

Response 1:

This is useful. Thank you.

We added more detailed explanations regarding the way to compute slant ionospheric delays, remove the receiver hardware bias, define the accuracy of the linear interpolation model, and others in session 2.2.

Point 2:

• How to get Equations (3) and (4) from Equations (1) and (2)?

Response 2:

We presented some more equations to get the ionosphere-free combination of pseudoranges and carrier phase measurements from Equations (1) and (2). The additional part is in the 2.1 section. Equations (3) and (4) now are noted as equations (7) and (8).

Point 3:

• The receiver and satellite clock errors are not associated with the ionosphere. What do the ionosphere-free clock errors mean? Please note what are and in the manuscript.

Response 3:

Thanks for your comments. The receiver and satellite clock errors are not associated with the ionosphere. In the revised manuscript, we revised those terms as receiver and satellite clock offsets.

Point 4:

• In figure 3, the vertical axis is the accuracy of ionospheric delay. Is it defined as the differences between the interpolated ionospheric corrections and measured values at the evaluation CORS station? Is the mean accuracy equal to the mean deviation? Does “mean accuracy” in Figure 3 have the same meaning to “mean of accuracy” in Figure 4?

Response 4:

Yes, the accuracy of ionospheric delay (the vertical axis in figure 3) is defined as the differences between the interpolated ionospheric corrections and measured values at the evaluation CORS station. It is not exactly as the mean deviation. The mean accuracy in this study is an average of all the accuracy retrieved on a whole day in the test station. The mean accuracy in Figure 3 and Figure 4 are the same. For consistency, we have corrected and used the “mean accuracy” for all the related texts in the manuscript.

 Point 5:

• In section 4, Kp is used to indicate the ionosoheric quiet day or ionospheric disturbed day. But in fact, Kp is an index of geomagnetic activity. So, the results need to be checked.

Response 5:

Thanks for your comment. The Kp is the global geomagnetic activity index and is just a part of the criteria for selecting the quiet and disturbed days. The evaluation period in this study included a geomagnetic storm. Therefore, to avoid misleading, we have revised some sentences in part 4 and added a reference for selecting the quiet and disturbed days. 

Point 6:

• The centimeter level accuracy of linear interpolation can be achieved only under some conditions. The conditions should be added in conclusion.

Response 6:

In the study, we found that centimetre level accurate ionospheric corrections can be achieved if there are sufficiently dense (i.e., nominal spacing of about 200 km) GNSS CORS networks in the region of interest. In the low latitude regions, the ionospheric correction around midnight or low elevation GNSS satellites below 20 deg during daytime can be high. In the revised version, we have added this sentence to the regional map in the discussion section for clarification.

Reviewer 2 Report

The paper assesses the achievable accuracy of ionospheric correction for PPP, as a function of the density of GNSS CORS networks in Australia. The paper can be published after minor revision. The revision must aim at making the text readable for a wider audience. The specific questions follow.

L. 126. Geometric position ?^?_?, what exactly is it? The position with respect to what? According to Eqs. (1,2), it is a scalar quantity, although the position should be vector. Explain in more details.

L. 155. Provide a reference on the Melbourne-Wübbena combination.

L. 206-225. More details are needed here. As follows from the previous text, the slant ionospheric delays are derived. These are geometry-dependent, and the geometry is different for different events. How exactly are they used?
What exactly is the linear interpolation? Linear in which coordinate?

I don't quite understand Figure 4. It shows three scenarios, where the station density changes significantly. The lower panels show the accuracy as a function of the average distance. First, the authors never define the average distance. If we assume that it just the average distance between one selected station and all the other stations, then it does not look a good predictor for the expected accuracy. And this can be seen in the lower panels. And, in such a definition, the average distance is not directly related to the density, as can be seen from the comparison of panels (g) and (i), where the density changes significantly, while the average distance does not. The authors must find a different predictor. It should be something more related to the station density.

Author Response

Please see the attachment. Thank you.

Author Response File: Author Response.pdf