On Satellite-Borne GPS Data Quality and Reduced-Dynamic Precise Orbit Determination of HY-2C: A Case of Orbit Validation with Onboard DORIS Data

Round 1

Reviewer 1 Report

The paper presents a new method for validation of reduced dynamics orbit determination, which is important for determination of  accuracy of the satellite position.
The article is readable, but the pictures resp. tables quite often do not agree with the text - Tab 5, Fig 6 (? were they changed at the last minute?)
Other comments on how to improve the text:
line 31: is ... radiopositioning in combination with satellite ... to be official ... radiopositioning integrated by satellite ...
end of line 52: I propose to add ... determination [3] by SLR (Satellite Laser Ranging).
line 114: add for clarity: ... DORIS data (at frequency 2036.25 MHz) ...
line 250: is ... by the Czech National Institute of Geodesy, Topography and Cartography; ... 
is supposed to be ... by the Research Institute of Geodesy, Topography and Cartography in the Czech Republic; ... (it is not a goverment institution)
Fig 2 description: please add ... effects of HY-2C and G09 satellites on DOY 348 in 2020. a) MP1, b) MP2
lines 290-291: the values ​​listed do not match what is in the referenced Table 5 - correct
line 367: suddenly you write about the components in the directions R, T, N. Fig 1 introduce + X, + Y, + Z. 
It is necessary to add a definition of which directions R, T, N.
line 369: the reference (Flavien et al. 2021) should be by number - probably [39].
Table 6 and the following: what is the difference between STD and RMS - please define both. In Tables 7, 10, 11, the values ​​in both columns are identical.
line 416: difference in R direction is worse than in the N direction only in DOY 353 (based on Figure 6) - no in DOY 248 - please correct.
line 456: from Figure 8: the residual flucutates in the R direction within + -0,0075 m, no + -0,075 m
line 457: again from Figure 8: in N direcion the residuals are between -0,0125 m and -0,0025 m (no zero)
line 510: is ..., except station 7838, the ... is supposed to be (from Figure 11) ..., except stations 7838 and 7841, the ...
line 542: what does it mean that "at most 52 beacon stations and at least 49 stations"? All DORIS stations are the beacon stations. 
The number of stations vary between days? Incomprehensible, please describe more or explain.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Dear authors,

you present quality checks of tracked GPS data of the HY-2 REceiver (carried out with the Czech G-Nut Software) and later on some statistics about orbit comparisons. One special feature is your try to use DORIS observations for orbit quality control.

The topic of the manuscript could be interesting to readers. However I have to note a large number of flaws in the text which urges me to reject the paper in the present form.

Below is a long list of technical questions flaggeg with 'T'. Aside there are numberous grammar issues ( I merge this grammar hints with the technical questions-however the grammar list by far not complete)

my commenst are ordered according to the line number at the right side of the manuscript

1) l17

'.. which was independently developed..'

2) l 19

The performance of the HY-2_Receiver...'

3) L23

',, models of the GPS antenna...'

4) T l26

here you note 'from 0.007 m to 0.008m ..'

It makes no sense to provide a range of 1mm for phase residuals-its sufficient to note

'.. is in the range of +/-8mm and the orbit overlap validation...'

5) l34 T

here you provide the number 0.0063m/s for an orbit accuracy?

I assume you meant the 'range-rate accuracy'

6) l54

'.. of data is small [4]. Therefore, SLR...'

7) l68, grammar +T

'The orbit accuracy employing the RD method is also higher compared to the dynamic and kinematic modelling.'

First of all note to which reference you compare here the orbit models? e.g

external ranging data ?

Moreover the esential advantage of the RD model is the lower number of required parameters, especially compared to the kinematic approach, Unfortunately your paper does not state in any line a comparison of required parameters by these methods !

8) l79

'.. satellites, which leads to an improved orbit accuracy...'

9) l99 T

why do refer here to the old ITRF2008 and not to the newer ITRF2014- DORIS of course also contributes to the ITRF2014

10) l101,102

'The CNES/CLS analysis ... '     

'In 1990 the SPOT-2 satellite, first ...'

11) l109 grammar +T

 here you note that a format is similar to a GPS measurement?

12) l110

'For a detailed ...'

13) l115

'.. is at the millimeter level. Converting the phase data into a format of average range rate, the nominal...'

14) l118,119

The sentence starting with 'The space-borne..' is not understandable and should be re-phrased

15) l122

Skip sentence 'That is...'

16) l124

'the performance of the HY2-Receiver...'

17) l152 T

I wonder why you note here a number of antennas to be 'perpendicular' to the z-axis; according to the plot most of them are parallel to the Z-axis.

In your own text you note in l158 that at least the LRA antenna points towards the positive Z-axis !!!

18) l155 T

'The HY2_Receiver ?provides? some parameters...'

the receiver provides GPS observations BUT NO parameters. These offsets are either detremined from analizing observations or much better from a pre-launch calibration

19) l163,164

'The DORIS system operates at two frequencies, ...'

20) l171

'.. (CODE) [36] with an sampling interval of 15 min.'

21) l182-184

The sentence about the DORIS format 3.0 is exactly also in the Introduction.

Please prevent such duplications. Might be that the sentence is better located here and should be skipped in the introduction

22)l188

Instead of 'All sources...'

'A summary of the data sources is provided in Table 3.'

23) l192-193 T

In the first sentence of chapter 2.2.1 you describe somehow dubious how Bernese handels orbits. Usually the bernese SW does nor establish an 'apriori kinematic orbit'. The apriori orbits are based either on available SP3 files or on broadcast messages from RINEX navigation files. Based on these apriori orbits the clock synchronsation is done CODSPP and later on also the pre-processing of phase data.

24) T l 196

'.. a new ambiguity parameter is set in the POD process?

new ambiguities are set in the parameter estimation process and not in the

POD process; POD parameters are just a subset of estimated parameters.

25) l200 T

here you write 'simulated dynamic satellite parameters'; why 'simulated'

Do you base your calculations and RD orbits on simulations?

In general the text lacks here from a clear description how you established your RD orbit. I assume you used the Bernese SW. In the POD you have set up the parameters decribed in lines 201-203? Also a clear statement is missing about the arc length of your RD orbits- did you set up 7 24 hours orbits? or longer, or shorter?

26) T, l203

here you note '.. 6 minutes as a constraint...'

why the wording 'constraint', its a pulse modelled as 3 parameters in directions R,T,N and these 3 pulses act on the positions within the next 6 minutes; Maybe the wording 'constraint' is not perfect

27)T

here I miss a discussion what happens if you further increase the update rate of pulses down to 30sec- is there still a remaining difference to a kinematic orbit ?

28) l205

'The models and parameters used together with satellite-borne GPS observations...'

29) l209

'Suppose that the DORIS phase observations...and f2 at time..'

30) l211

'.. phase value at f1...'

31) T

From the text in sub chapter 2.2.3 it becomes not clear if you have calculated range-rates from raw DORIS observations or if you have utilized a given HY-2 orbit calculated by e.g. CNES from DORIS observations. Do you further derive your DORIS range rates from Eq (2) or from Eq (3)?

32) l215-217

'The time system used by DORIS is International Atomic Time (TAI) while the RD orbit relates to GPS Time (GPST). Therefore the time-offset between both scales has to be considered before validation is performed.'

33) l221,222 T

here I assume you mix up 2 different vectors denoted by the delta-rho acronyms. On the one hand the satellite movement is the vector between

2 subsequent positions of the HY2 satellite. On the other hand if you calculate rho from the pythagoras (as done in line 223) the difference 

rhok-rhoj is the range difference which should not to be mixed up in your text.

34) l232

the wording 'accidental' error is strange

35) sentence in line 236 has to be re-phrased for a better understanding

36) l247

',, the performance of the HY2...'

37)l252

'... and supports the ..'

38) T text below equations (5),(6)

the description of the constituents in formulae (5) and (6) is not correct

 MP1,MP2 is the code multipath on freq 1 and 2

 M1, M2 are the code ranges on both frequencies

mi is the multipath phase noise

39) l267

enter the time scale , is it 20:25 to 21:05 UTC ?

40) l 267

'.. is selected. The Code multipath as well as the elevation angle of G09 as observed from HY-2C are calculated and shown in Fig.2.'

41) l275

'.. elevation angle is larger than...'

42) l276

'.. angle is above 30...'

43) T l 278

you deduce from Figure 2 that MP1 is larger than for MP2;

this is slightly different from experience - usually the MP is larger for

signals with worse Signal/noise ratio . So here you should also focus

on the S/N ratio of the investigated HY-2 data, check it; is P2 S/N really

better than for P1

44) l288

',, the performance of the HY-2C satellite-borne GPS recveiver data.'

45) T , Figure 3

why do the plots show positive and negative ionospheric delays for various satellites? the changing sign of the rate of change might be due to

rising or sinking satellites. But the ionospheric delay is always positive for code observations and negative for phase observations.

Explain?

46) l318

The sentence 'Since the time window..' should be re-phrased

47) l327

what are 'abnormal values' ?

48) T l347,348

here you note '.. of PCV must come after that of PCO...'

this is slightly dubious. PCO and PCV are correlated-therefore they can be solely estimated together- or you got pre-launch calibrations of at least PCO. IN that case PCO is fixed and PCV is estimated on top.

49) T, l358

explain in one more sentence what's the meaing of 'a direct method'?

50) add to legend of Figure 4 that the graphic shows the PCV model of

HY-2 and not for an GPS antenna; the GPS PCO.PCV values can be extracted from the IGS14.atx file

51) T, table 6

a) note in the legend over which period this comparison is done

 one day? one week? or revolution periods?

b) note once more the origin of your PSO orbit - from CNES?

c) you denote coordinate differences as 'minimum' and maximum'

which are in reality the largest negative and the largest positive difference

d) think about the necessity of addidng columns for std and RMS

as most of the means shown here and in subsequent tables are close to zero

you may skip column std as thats always the same as rms (for a zero mean)

e) to provide position differences in units of 0.1 mm is not useful or reasonable. To give numbers up to the mm-level would be sufficient

PLEASE NOTE-my comments to columns c) d) e) also relate to most of the

subsequent comparison tables!!!

52) T, can you explain why introducing PCO,PCV models affect mostly the accuracy in T direction and not in R?

53) T, Figure 5 displays a quite daily pattern of orbit residuals in N direction but none in R and T, do you have an explanation?

Do you show here 7 concatenated 24 orbit residuals or?

54) T ; l 433

in table 7 you provide statistics about phase residuals

again my comments 51) are valid- especially the because the residuals are obviously output of an least squares proces with [residuals]= zero -> see you column mean ; so mean, and std are useless

55) T chapter 3.2.3 -overlaps

usually overlaps are checked on the boundaries of the individual arcs;

again the question raises, did you calculate 24 h arcs (would be reasounalble as your GPS orbits are 24h arcs) or 1 long arc of 7 days? or just arcs of the revolution period? So its slightly unusual to show orbit overlaps between 12-18h. You should comment on the patterns of the graphs. do they reflect the revolution period of the satellite?

56) T ,l456

I assume you want to write 0.0075 instead of 0.075 ?

57) l 460

'.. Fig.9 shows the RMS...'

58) T , Table 8

are the statistics in table 8 calculated as the mean over 7 individual day slots between 12-18h ?

59) l470

'.. difference between the overlapping orbits...'

60) l484,485

'.. accuracy of HY-2C. According to [60] the accuracy of the SLR orbits is in the range of 0.01-0.02m. '

61) T l486

check if there is also a SLRF FRame 2014 available?

62) T l501-503

make more clear in the text whats checked here

a) range residuals between the SLR NP data and the range differences between your RD orbit with respect to the ground station coordinates ?

b) range differences between a SLR orbit (calculated by e.g. CNES or ILRS)

and your RD orbit?

63) T

I really wonder why you do not discuss the aspect that the LRS is not located in the CoM of the satellite (to which your RD refers). Did enter somewhere the information about the LRA vector to Com as well as information about the satellite attitude?

64) line 515

to which loaction refers station number '1873' ?

65) l525 and 530

duplication; the sentence 'A new validation..' is duplicated here

66) l550

' .. the accuracy of the RD orbit of HY-2C gained from the DORIS validation.'

67) Tables 10 and 11)->see my comments 51) and 54)

68) In table 11 you show comparisons of range rates between your RD orbit and the PSO orbit with respect to DORIS range rates. the RMS of about 0.007 m/s is quite goord but there are still differences up to 0.12m. From the table I learn that PSO and RD orbits are quite close but there are some large differences to the DORIS range rates; Is it still reasonable to rate the DORIS observation a reasonable tool for checking orbit accuracy?

69) lines 577-580 grammar plus T

the text here has definitely to be re-phrased.

maybe you should add a number of state-of.the art GPS POD accuracy. Moreover its quite clear that modelling a LEO orbit is more demanding than for a MEO. You should add also in the begin of your article a table with rough keplerian elements of HY-2C.

70) chapter 4.discussion

is this chapter necessary or helpful- should be shifted to an appendix

71) l 605-607, T

again refer to my comments above on your MP analyses. moreover the L1 or L2 band are not susceptible to elevation changes but maybe to multipath effects which are of course closely linked to elevation of the incoming signals

72) T, l612

here you write 'The range of iono-free phase residuals varies from 0.007 to 0.008 m, But in Figure 7 these residuals vary in the +/- 3cm band??

best regards

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

see the attached file

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Dear authors

your paper was clearly improved during the revision process.

Almost all of my raised issues were settled. See below

a short list of remaining (mostly grammar) deficiencies

1) Introduction ,lines 59,60

  'DORIS POD relies solely on the dynamic method. In contrast GPS POD can be based on the dynamic approach as well as reduced-dynamic (RD) and kinematic methods. '

2) l 170

  '.. the force acting on the satellite is specified in a satellite orbit coordinate...

 .. RTN). This RTN system is well suited for an easy analyses of the POD quality. '

3) l 197

  '... in RINEX DORIS 3.0 format. RINEX has been ...'

4) l 199

  'The RINEX format can also be easily adapted to contain data ...'

5) l 212

'.. is applied, an a priori dynamic orbit ...'

6) l 289

'.. the errors of the satellite position in three directions in an Earth-Centered..'

7) in line 292 you again use the somehow misleading word 'accidential'

8) l302

'... HY2_Receiver is assessed...'

9) below new equation (1o)

I wonder if the equation numbering is still correct?

check all equ numbers in text

l370) really Eq (8) ?

l379 really Eq(7) and (8) ?

10) Figures 3 ,

I accept your explanation of the + and - signs of the ionopheric residuals

but this should be also noted in the legend to Figure 3

make clear that in figures (a) are ionospheric residuals shown according to new Eq (10)

11) l382

' A 24-hours window will...'

12) l416

'.. the PCV values are estimated at each grid point. This method is not ...'

13) l427

'.. antenna of the HY2_Receiver...'

14)  inline 434 you write that you have not enough observation in elevations between 80-90 deg. Is this caused by the overall small data set or just by geometry between HY2 and the GPS orbit constellation?

15) l437

'determine' - not 'determinate'

16) l457

'.. that the RMS values...'

17) l464

'Compared to Scheme 1, the accuracy...'

18) l481

'.. extreme values show up because the pseudo-stochastic pulse parameters fail to capture the influence of errorsr, such...'

19) l511,512

'.. stable changes. Table 7 suggests that the average RMS value of the 7-day carrier-phase residuals is about 0.008 m.'

20) lines 581,582

  '.. RD orbit and the ground station coordinates. The SLR range validation...'

best regards

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments for Figure 4: It is not strict to say the values in 0-30 are small, as we can see some dark red spots; empty values between 80-90 is not reflected in the figure. 'Figure 5' is wrongly used in the text.

Please revise and read through your whole manuscript to remove typos.

Author Response

Please see the attachment.

Author Response File: Author Response.docx