Assessing and mitigating the radar &ndash; radar interference in the German C-band weather radar network

Frech, Michael; Hald, Cornelius; Schaper, Maximilian; Lange, Bertram; Rohrdantz, Benjamin

doi:https://doi.org/10.5194/egusphere-2022-532

Preprints

https://doi.org/10.5194/egusphere-2022-532

Preprints

29 Aug 2022

| 29 Aug 2022

Assessing and mitigating the radar – radar interference in the German C-band weather radar network

Michael Frech, Cornelius Hald, Maximilian Schaper, Bertram Lange, and Benjamin Rohrdantz

Abstract. The national German weather radar network operates in C-band between 5.6 and 5.65 GHZ. In a radar network, individual transmit frequencies have to be chosen such that radar-radar induced interferences are avoided. In an unique experiment the Hohenpeissenberg research radar and five operational systems from the radar network were used to characterize radar-radar induced interferences as a function of the radar frequency. The results allow to assess the possibility to add additional C-band radars with magnetron transmitters into the existing network. Based on the experiment, at least a 15 MHz separation of the nominal radar frequency is needed to avoid a radar-radar interference. The most efficient mitigation of radar-radar interference is achieved by the "Radar Tango". Latter refers to the synchronized scanning of all radar systems in the network. Based on those results, additional C-band radar systems can be added to the German weather radar network if a further improvement of the radar coverage is needed.

Received: 23 Jun 2022 – Discussion started: 29 Aug 2022

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.

Download & links

Preprint (PDF, 2694 KB)

Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
Preprint (2694 KB)

Download & links

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Journal article(s) based on this preprint

20 Jan 2023

Assessing and mitigating the radar–radar interference in the German C-band weather radar network

Michael Frech, Cornelius Hald, Maximilian Schaper, Bertram Lange, and Benjamin Rohrdantz

Atmos. Meas. Tech., 16, 295–309, https://doi.org/10.5194/amt-16-295-2023,https://doi.org/10.5194/amt-16-295-2023, 2023

Short summary

Michael Frech, Cornelius Hald, Maximilian Schaper, Bertram Lange, and Benjamin Rohrdantz

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2022-532', Anonymous Referee #1, 25 Oct 2022

This manuscript describes how DWD tries to minimize the effects of radar - radar interferences in the German weather radar network. In an unique experiment using five operational and one research radar the effects were assessed and evaluated. The manuscript is very comprehensive and certainly deserves publication in AMT after minor changes.

There are two points which could be elaborated in more detail:

- the definition of STD is not precise and seems to be a special product of the GAMIC signal processor. Also, the Schaper et al. 2022 preprint does not give sufficient details. In line 56 STD is defined as the standard deviation of the power using I&Q data. In Schaper et al. 2022 it is defined as a normalized standard deviation of received power. Obviously, the expected range of 0 .. 1 implies normalization. But to which value? What is “mean” STD (line 62) in this context? The authors should give more details for users not familiar with the GAMIC signal processor. I also can not follow the statement that STD app. 0.5 indicated all pulses have the same power (line 156). If in a timeseries all data have the same value, standard deviation is 0! SQI has a more common definition, it is provided by a number of signal processors, and is generally defined as the magnitude of autocorrelation lag 1 divided by autocorrelation lag 0; or simply 1 by (spectral width normalized with Nyquist interval). Assuming that Tx phase is removed from measured phase angle.

- the authors should give some more explanation why the interferences are reduced in the case where all radars operate with the same PRF (line 187).

Minor remarks:

Line 30: is 2 km in Fig.1

Line 41: explain PRF

Line 110: receive end: MHP? Transmit end: other radars?

Line 125: H_b above radar?

Line 154+155: better: The received power is independent from range.

Line 229: refer to Table A2 for the elevations used by PCP at MHP

Figure 10: to what does 1.4 MHz refer? Red numbers at lower left side

Line 283: in Table A3 sweep 5 is at 0.5° elevation

Citation: https://doi.org/10.5194/egusphere-2022-532-RC1
- AC1: 'Reply on RC1', Michael Frech, 01 Dec 2022
  
  We thank the reviewer for his helpful and constructive comments which will help to improve the paper. We respond to the issues raised in point by point reply and incorporate this response in the revised manuscript.
  Reviewer comment: "the definition of STD is not precise and seems to be a special product of the GAMIC signal processor. Also, the Schaper et al. 2022 preprint does not give sufficient details. In line 56 STD is defined as the standard deviation of the power using I&Q data. In Schaper et al. 2022 it is defined as a normalized standard deviation of received power. Obviously, the expected range of 0 .. 1 implies normalization. But to which value? What is “mean” STD (line 62) in this context? The authors should give more details for users not familiar with the GAMIC signal processor. I also can not follow the statement that STD app. 0.5 indicated all pulses have the same power (line 156). If in a timeseries all data have the same value, standard deviation is 0! SQI has a more common definition, it is provided by a number of signal processors, and is generally defined as the magnitude of autocorrelation lag 1 divided by autocorrelation lag 0; or simply 1 by (spectral width normalized with Nyquist interval). Assuming that Tx phase is removed from measured phase angle."
  Response: STD: We operate the signal processor Enigma in a DAS ("dynamic angle syncing" mode). Radar moments are computed from the batch of received pulses that are collected for the predefined angle width (1°). The number of processed pulses depends on the antenna speed and the PRF. For a batch of pules, STD is the normalized standard deviation of the I/Q magnitude in a given range bin. The standard deviation of the I/Q power in the batch is normalized with the mean I/Q power. Pulsed interferences usually are not correlated with the radar PRF, and the standard deviation of the I/Q power magnitude is increased because the interference is expected to be not present in every single pulse. The noisiness then increases, and the SQI decreases. This is the principle idea behind the usage of STD and SQI for the interference detection.
  To clarify the description, we will change "standard deviation of the power" in line 56 to "normalized standard deviation", remove the word "mean" in line 62 and add some more in depth explanation of STD in the revised version. Line 156 will be rewritten to: "In most of the scanned area, STD is at ~0.5, indicating a typical value for recorded noise. A stable weather signal usually is at STD < 0.2.".
  Reviewer comment: the authors should give some more explanation why the interferences are reduced in the case where all radars operate with the same PRF (line 187).
  Response: We have not found a clear reason for this feature. We assume that the identical PRF settings might somehow minimize the amount of time spans that the receiving MHP radar can record the interference. Possible influencing factors are the PRFs, the maximum range set by them, and the fact that our operational scans only record the first 150 km of range. We will add a sentence in the revised manuscript stating that reason for this yet unknown."
  Response to the minor comments:
  Line 30: is 2 km in Fig.1
  -Thank you. We will change the number from 3 to 2 in the revised manuscript
  Line 41: explain PRF
  -PRF (pulse repetition frequency) is already explained in line 21/22.
  Line 110: receive end: MHP? Transmit end: other radars?
  -Correct. We will add the short names of the respective transmitting and receiving radars in brackets.
  Line 125: H_b above radar?
  -H_b in this case describes the beam height over an imaginary "earth" with a starting radius of r_e (which depends on latitude, elevation and finally the radars height over ground) and the 4/3 factor. So we think "above ground" is more fitting than "above radar".
  Line 154+155: better: The received power is independent from range.
  -Agreed. We will change the text to your suggestion.
  Line 229: refer to Table A2 for the elevations used by PCP at MHP
  -Good suggestion! We will add a cross reference to the table in the revised manuscript.
  Figure 10: to what does 1.4 MHz refer? Red numbers at lower left side
  -The number of 0.4 µs was wrong in the figure and will be replaced by 0.8 µs in the revised version. The 1.4 MHz refers to the 3dB bandwidth of the transmitted pulse with a pulse duration of 0.8 µs.
  Line 283: in Table A3 sweep 5 is at 0.5° elevation
  -We will change the number in the sentence to 0.5°. For some additional info on the chosen elevations:
  
  Sweep number 5 is indeed at 0.5° elevation. The elevation steps of 1° are chosen according to the antenna beam width (roughly 1°) at low elevation (up to 5.5°) in order to obtain a good volume sampling close to the surface. Compared to the precipitation scan, a 0.5° elevation scan may contain more clutter, but at the same time also contains valuable radar information close to the surface. The 0.8° elevation of the precipitation scan has been chosen empirically in order to provide good quantitative precipitation estimates close to the surface and to avoid ground clutter.
  
  Citation: https://doi.org/10.5194/egusphere-2022-532-AC1
RC2:
'Comment on egusphere-2022-532', Anonymous Referee #2, 01 Nov 2022

The paper investigates radar-to-radar interferences on the national german weather radar network operated by DWD. Experimenting with radar located in Southern Germany by setting up dedicated scans, the authors demonstrate that interferences are strongest when close radars operate within 15 MHz. The authors also showed that the "Radar Tango" scan strategy can reduce interferences.

The work is of interest for operational applications in densely populated weather radar areas operating at the same band. The study is scientifically based and the research is quite clearly exposed. Hereafter, some remarks that can be improved clearness.

Major remarks

RFIs at C-band are nowadays widespread in Europe. The study focuses on RFIs caused by other weather radars. In the data analysis, it is not clear how, in the data analysis, wifi RFIs are distinguished from weather radar ones. The amount of wifi RFIs can be partially inferred by Figure 6, where about 200 bins are classified as disturbed.

A short description of current meteorological conditions, focusing on atmospheric refractivity, could help to better interpret the analysis. Moreover, it is unclear which PW is set up in the experimental scans.

Figure 6 and Figure 7 summarize the number and properties of disturbed bins: RhoHV shows a wide variability: it's worthing to add some comments trying to explain possible causes.

The German radar network shows large overlapping areas where 2D or even 3D winds can be retrieved. Do the authors believe that the "Radar Tango" scan strategy affects or limits such retrieval? The reviewer suggests a discussion on the possible drawbacks of "Radar Tango".

Minor remarks

Figure 1 - Add scale bar and North Arrow, please.

Figure 2 - Add scale bar and North Arrow, please. Adding distance in km at each arrow could improve clearness.

Figure 6 - change "Boxplot of the number of disturbed pixels" to "Boxplot of the number of disturbed bins", please

Figure 6 and Figure 7 should report the same range on both axes.

Citation: https://doi.org/10.5194/egusphere-2022-532-RC2
- AC2: 'Reply on RC2', Michael Frech, 01 Dec 2022
  
  We thank the reviewer for his helpful and constructive comments which will help to improve the paper. We respond to the issues raised in point by point reply and incorporate this response in the revised manuscript.
  
  Reviewer comment: RFIs at C-band are nowadays widespread in Europe. The study focuses on RFIs caused by other weather radars. In the data analysis, it is not clear how, in the data analysis, wifi RFIs are distinguished from weather radar ones. The amount of wifi RFIs can be partially inferred by Figure 6, where about 200 bins are classified as disturbed.
  Response: We did not try to eliminate the WIFI disturbances. We analyze the relative difference in the numbers of the identified classified rangebins. Any differences are then attributed to the respective radar configuration of the transmitting radar. This makes the assumption, that the WIFI disturbance is constant throughout the considered time period. For the MHP radar this is a valid assumption based on the continuous monitoring for WIFI interferences as described in Schaper at al. (2022).
  Reviewer comment: A short description of current meteorological conditions, focusing on atmospheric refractivity, could help to better interpret the analysis. Moreover, it is unclear which PW is set up in the experimental scans.
  Response First, all transmitting radars were using a PW of 0.8 µs during the experiment. We will add this information to table 3. The PWs of the receiving radar MHP is described in the appendix, tables A.1 and A.3. All evaluated sweeps use a PW of 0.8 µs as well.
  
  Second, we agree that showing the connection between disturbed bins and refractivity gradients would be helpful. We will add a new paragraph at the end of section 2 where we calculated refractivity gradients from radiosonde data and compare the frequency of occurrence of strongly negative values with the amount of recorded interferences. This will be supported by an additional figure. It will show that we have more disturbed bins when the radiosondes show lower values for the refractivity gradient.
  Reviewer: Figure 6 and Figure 7 summarize the number and properties of disturbed bins: RhoHV shows a wide variability: it's worthing to add some comments trying to explain possible causes.
  Response: First of all, RhoHV is the cross correlation coefficient from time series where no Doppler-Clutter filter has been applied. There is no further processing or quality control (e.g. thresholding; this information will be in the revised manuscript) applied. The large variability in RhoHV reflects the typical observed value range if there is no precipitation, and where the backscattered signal is caused by targets (e.g. insects, moving ground clutter) that are often without correlated backscatter in the two polarization planes.
  
  % The uncorrected RhoHV also has a strong dependence on SNR: if SNR is below 0, RhoHV never reaches values above 0.6 even in pure rain. at SNR ~10 dB, RhoHV can take almost all values between 0 and 1 and only above that RhoHV takes on stable values. RhoHV in figure 6 therefore has a small variability, since most SNR values lie stably above 10 dB. In figure 7, variability in RhoHV is higher, because more bins have a SNR clearly below 10 dB which results in a lower RhoHV. We will add a few new sentences at the end of section 3.2 that explain this connection.
  
  Also please note that with your suggestion of having matching axes in both figures it becomes very clear that variability of RhoHV is only high in figure 7. The boxes in Figure 6 for RhoHV become very small when showing the whole range from 0 to 1.
  Reviewer: The German radar network shows large overlapping areas where 2D or even 3D winds can be retrieved. Do the authors believe that the "Radar Tango" scan strategy affects or limits such retrieval? The reviewer suggests a discussion on the possible drawbacks of "Radar Tango".
  Response: We do not expect a disadvantage. This is supported by the evaluation of the dual-Doppler retrieval of the 3-D wind vector from volume data with multiple sweeps we just did for another study. Even without the Radar-Tango, there always will be time differences in sampling a specific volume from 2 or more radars. So a stationarity assumption always has to be made, which is, depending on the meteorological phenomena, more or less an approximation, especially if data from multiple sweeps are used.
  Minor remarks of the reviewer:
  Figure 1 - Add scale bar and North Arrow, please.
  - We will add a scale bar to the figure. A North Arrow would not be useful in our opinion. The projection of the map is such that the longitudes converge towards north. If we put the arrow in the lower left corner, for example, it would only be valid for this exact longitude. The north direction in the center or on the right hand side of the map is a different one. To clarify the position and direction of the map we will add the following sentence to the description: "The grid is in latitudes north and longitudes east."
  Figure 2 - Add scale bar and North Arrow, please. Adding distance in km at each arrow could improve clearness.
  -We will add all requested items to the map in the revised version.
  Figure 6 - change "Boxplot of the number of disturbed pixels" to "Boxplot of the number of disturbed bins", please
  -We will change the annotations from "pixels" to "bins" in Figures 6 and 7.
  Figure 6 and Figure 7 should report the same range on both axes.
  -We agree, the difference is more clearly seen with matching ranges. We will adopt your suggestion in the revised version of the manuscript.
  
  Citation: https://doi.org/10.5194/egusphere-2022-532-AC2

Interactive discussion

Status: closed

RC1:
'Comment on egusphere-2022-532', Anonymous Referee #1, 25 Oct 2022

This manuscript describes how DWD tries to minimize the effects of radar - radar interferences in the German weather radar network. In an unique experiment using five operational and one research radar the effects were assessed and evaluated. The manuscript is very comprehensive and certainly deserves publication in AMT after minor changes.

There are two points which could be elaborated in more detail:

- the definition of STD is not precise and seems to be a special product of the GAMIC signal processor. Also, the Schaper et al. 2022 preprint does not give sufficient details. In line 56 STD is defined as the standard deviation of the power using I&Q data. In Schaper et al. 2022 it is defined as a normalized standard deviation of received power. Obviously, the expected range of 0 .. 1 implies normalization. But to which value? What is “mean” STD (line 62) in this context? The authors should give more details for users not familiar with the GAMIC signal processor. I also can not follow the statement that STD app. 0.5 indicated all pulses have the same power (line 156). If in a timeseries all data have the same value, standard deviation is 0! SQI has a more common definition, it is provided by a number of signal processors, and is generally defined as the magnitude of autocorrelation lag 1 divided by autocorrelation lag 0; or simply 1 by (spectral width normalized with Nyquist interval). Assuming that Tx phase is removed from measured phase angle.

- the authors should give some more explanation why the interferences are reduced in the case where all radars operate with the same PRF (line 187).

Minor remarks:

Line 30: is 2 km in Fig.1

Line 41: explain PRF

Line 110: receive end: MHP? Transmit end: other radars?

Line 125: H_b above radar?

Line 154+155: better: The received power is independent from range.

Line 229: refer to Table A2 for the elevations used by PCP at MHP

Figure 10: to what does 1.4 MHz refer? Red numbers at lower left side

Line 283: in Table A3 sweep 5 is at 0.5° elevation

Citation: https://doi.org/10.5194/egusphere-2022-532-RC1
- AC1: 'Reply on RC1', Michael Frech, 01 Dec 2022
  
  We thank the reviewer for his helpful and constructive comments which will help to improve the paper. We respond to the issues raised in point by point reply and incorporate this response in the revised manuscript.
  Reviewer comment: "the definition of STD is not precise and seems to be a special product of the GAMIC signal processor. Also, the Schaper et al. 2022 preprint does not give sufficient details. In line 56 STD is defined as the standard deviation of the power using I&Q data. In Schaper et al. 2022 it is defined as a normalized standard deviation of received power. Obviously, the expected range of 0 .. 1 implies normalization. But to which value? What is “mean” STD (line 62) in this context? The authors should give more details for users not familiar with the GAMIC signal processor. I also can not follow the statement that STD app. 0.5 indicated all pulses have the same power (line 156). If in a timeseries all data have the same value, standard deviation is 0! SQI has a more common definition, it is provided by a number of signal processors, and is generally defined as the magnitude of autocorrelation lag 1 divided by autocorrelation lag 0; or simply 1 by (spectral width normalized with Nyquist interval). Assuming that Tx phase is removed from measured phase angle."
  Response: STD: We operate the signal processor Enigma in a DAS ("dynamic angle syncing" mode). Radar moments are computed from the batch of received pulses that are collected for the predefined angle width (1°). The number of processed pulses depends on the antenna speed and the PRF. For a batch of pules, STD is the normalized standard deviation of the I/Q magnitude in a given range bin. The standard deviation of the I/Q power in the batch is normalized with the mean I/Q power. Pulsed interferences usually are not correlated with the radar PRF, and the standard deviation of the I/Q power magnitude is increased because the interference is expected to be not present in every single pulse. The noisiness then increases, and the SQI decreases. This is the principle idea behind the usage of STD and SQI for the interference detection.
  To clarify the description, we will change "standard deviation of the power" in line 56 to "normalized standard deviation", remove the word "mean" in line 62 and add some more in depth explanation of STD in the revised version. Line 156 will be rewritten to: "In most of the scanned area, STD is at ~0.5, indicating a typical value for recorded noise. A stable weather signal usually is at STD < 0.2.".
  Reviewer comment: the authors should give some more explanation why the interferences are reduced in the case where all radars operate with the same PRF (line 187).
  Response: We have not found a clear reason for this feature. We assume that the identical PRF settings might somehow minimize the amount of time spans that the receiving MHP radar can record the interference. Possible influencing factors are the PRFs, the maximum range set by them, and the fact that our operational scans only record the first 150 km of range. We will add a sentence in the revised manuscript stating that reason for this yet unknown."
  Response to the minor comments:
  Line 30: is 2 km in Fig.1
  -Thank you. We will change the number from 3 to 2 in the revised manuscript
  Line 41: explain PRF
  -PRF (pulse repetition frequency) is already explained in line 21/22.
  Line 110: receive end: MHP? Transmit end: other radars?
  -Correct. We will add the short names of the respective transmitting and receiving radars in brackets.
  Line 125: H_b above radar?
  -H_b in this case describes the beam height over an imaginary "earth" with a starting radius of r_e (which depends on latitude, elevation and finally the radars height over ground) and the 4/3 factor. So we think "above ground" is more fitting than "above radar".
  Line 154+155: better: The received power is independent from range.
  -Agreed. We will change the text to your suggestion.
  Line 229: refer to Table A2 for the elevations used by PCP at MHP
  -Good suggestion! We will add a cross reference to the table in the revised manuscript.
  Figure 10: to what does 1.4 MHz refer? Red numbers at lower left side
  -The number of 0.4 µs was wrong in the figure and will be replaced by 0.8 µs in the revised version. The 1.4 MHz refers to the 3dB bandwidth of the transmitted pulse with a pulse duration of 0.8 µs.
  Line 283: in Table A3 sweep 5 is at 0.5° elevation
  -We will change the number in the sentence to 0.5°. For some additional info on the chosen elevations:
  
  Sweep number 5 is indeed at 0.5° elevation. The elevation steps of 1° are chosen according to the antenna beam width (roughly 1°) at low elevation (up to 5.5°) in order to obtain a good volume sampling close to the surface. Compared to the precipitation scan, a 0.5° elevation scan may contain more clutter, but at the same time also contains valuable radar information close to the surface. The 0.8° elevation of the precipitation scan has been chosen empirically in order to provide good quantitative precipitation estimates close to the surface and to avoid ground clutter.
  
  Citation: https://doi.org/10.5194/egusphere-2022-532-AC1
RC2:
'Comment on egusphere-2022-532', Anonymous Referee #2, 01 Nov 2022

The paper investigates radar-to-radar interferences on the national german weather radar network operated by DWD. Experimenting with radar located in Southern Germany by setting up dedicated scans, the authors demonstrate that interferences are strongest when close radars operate within 15 MHz. The authors also showed that the "Radar Tango" scan strategy can reduce interferences.

The work is of interest for operational applications in densely populated weather radar areas operating at the same band. The study is scientifically based and the research is quite clearly exposed. Hereafter, some remarks that can be improved clearness.

Major remarks

RFIs at C-band are nowadays widespread in Europe. The study focuses on RFIs caused by other weather radars. In the data analysis, it is not clear how, in the data analysis, wifi RFIs are distinguished from weather radar ones. The amount of wifi RFIs can be partially inferred by Figure 6, where about 200 bins are classified as disturbed.

A short description of current meteorological conditions, focusing on atmospheric refractivity, could help to better interpret the analysis. Moreover, it is unclear which PW is set up in the experimental scans.

Figure 6 and Figure 7 summarize the number and properties of disturbed bins: RhoHV shows a wide variability: it's worthing to add some comments trying to explain possible causes.

The German radar network shows large overlapping areas where 2D or even 3D winds can be retrieved. Do the authors believe that the "Radar Tango" scan strategy affects or limits such retrieval? The reviewer suggests a discussion on the possible drawbacks of "Radar Tango".

Minor remarks

Figure 1 - Add scale bar and North Arrow, please.

Figure 2 - Add scale bar and North Arrow, please. Adding distance in km at each arrow could improve clearness.

Figure 6 - change "Boxplot of the number of disturbed pixels" to "Boxplot of the number of disturbed bins", please

Figure 6 and Figure 7 should report the same range on both axes.

Citation: https://doi.org/10.5194/egusphere-2022-532-RC2
- AC2: 'Reply on RC2', Michael Frech, 01 Dec 2022
  
  We thank the reviewer for his helpful and constructive comments which will help to improve the paper. We respond to the issues raised in point by point reply and incorporate this response in the revised manuscript.
  
  Reviewer comment: RFIs at C-band are nowadays widespread in Europe. The study focuses on RFIs caused by other weather radars. In the data analysis, it is not clear how, in the data analysis, wifi RFIs are distinguished from weather radar ones. The amount of wifi RFIs can be partially inferred by Figure 6, where about 200 bins are classified as disturbed.
  Response: We did not try to eliminate the WIFI disturbances. We analyze the relative difference in the numbers of the identified classified rangebins. Any differences are then attributed to the respective radar configuration of the transmitting radar. This makes the assumption, that the WIFI disturbance is constant throughout the considered time period. For the MHP radar this is a valid assumption based on the continuous monitoring for WIFI interferences as described in Schaper at al. (2022).
  Reviewer comment: A short description of current meteorological conditions, focusing on atmospheric refractivity, could help to better interpret the analysis. Moreover, it is unclear which PW is set up in the experimental scans.
  Response First, all transmitting radars were using a PW of 0.8 µs during the experiment. We will add this information to table 3. The PWs of the receiving radar MHP is described in the appendix, tables A.1 and A.3. All evaluated sweeps use a PW of 0.8 µs as well.
  
  Second, we agree that showing the connection between disturbed bins and refractivity gradients would be helpful. We will add a new paragraph at the end of section 2 where we calculated refractivity gradients from radiosonde data and compare the frequency of occurrence of strongly negative values with the amount of recorded interferences. This will be supported by an additional figure. It will show that we have more disturbed bins when the radiosondes show lower values for the refractivity gradient.
  Reviewer: Figure 6 and Figure 7 summarize the number and properties of disturbed bins: RhoHV shows a wide variability: it's worthing to add some comments trying to explain possible causes.
  Response: First of all, RhoHV is the cross correlation coefficient from time series where no Doppler-Clutter filter has been applied. There is no further processing or quality control (e.g. thresholding; this information will be in the revised manuscript) applied. The large variability in RhoHV reflects the typical observed value range if there is no precipitation, and where the backscattered signal is caused by targets (e.g. insects, moving ground clutter) that are often without correlated backscatter in the two polarization planes.
  
  % The uncorrected RhoHV also has a strong dependence on SNR: if SNR is below 0, RhoHV never reaches values above 0.6 even in pure rain. at SNR ~10 dB, RhoHV can take almost all values between 0 and 1 and only above that RhoHV takes on stable values. RhoHV in figure 6 therefore has a small variability, since most SNR values lie stably above 10 dB. In figure 7, variability in RhoHV is higher, because more bins have a SNR clearly below 10 dB which results in a lower RhoHV. We will add a few new sentences at the end of section 3.2 that explain this connection.
  
  Also please note that with your suggestion of having matching axes in both figures it becomes very clear that variability of RhoHV is only high in figure 7. The boxes in Figure 6 for RhoHV become very small when showing the whole range from 0 to 1.
  Reviewer: The German radar network shows large overlapping areas where 2D or even 3D winds can be retrieved. Do the authors believe that the "Radar Tango" scan strategy affects or limits such retrieval? The reviewer suggests a discussion on the possible drawbacks of "Radar Tango".
  Response: We do not expect a disadvantage. This is supported by the evaluation of the dual-Doppler retrieval of the 3-D wind vector from volume data with multiple sweeps we just did for another study. Even without the Radar-Tango, there always will be time differences in sampling a specific volume from 2 or more radars. So a stationarity assumption always has to be made, which is, depending on the meteorological phenomena, more or less an approximation, especially if data from multiple sweeps are used.
  Minor remarks of the reviewer:
  Figure 1 - Add scale bar and North Arrow, please.
  - We will add a scale bar to the figure. A North Arrow would not be useful in our opinion. The projection of the map is such that the longitudes converge towards north. If we put the arrow in the lower left corner, for example, it would only be valid for this exact longitude. The north direction in the center or on the right hand side of the map is a different one. To clarify the position and direction of the map we will add the following sentence to the description: "The grid is in latitudes north and longitudes east."
  Figure 2 - Add scale bar and North Arrow, please. Adding distance in km at each arrow could improve clearness.
  -We will add all requested items to the map in the revised version.
  Figure 6 - change "Boxplot of the number of disturbed pixels" to "Boxplot of the number of disturbed bins", please
  -We will change the annotations from "pixels" to "bins" in Figures 6 and 7.
  Figure 6 and Figure 7 should report the same range on both axes.
  -We agree, the difference is more clearly seen with matching ranges. We will adopt your suggestion in the revised version of the manuscript.
  
  Citation: https://doi.org/10.5194/egusphere-2022-532-AC2

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision | EF: Editorial file upload

AR by Michael Frech on behalf of the Authors (16 Dec 2022) Author's response Author's tracked changes Manuscript

ED: Publish as is (19 Dec 2022) by Gianfranco Vulpiani

AR by Michael Frech on behalf of the Authors (19 Dec 2022)

Journal article(s) based on this preprint

20 Jan 2023

Assessing and mitigating the radar–radar interference in the German C-band weather radar network

Michael Frech, Cornelius Hald, Maximilian Schaper, Bertram Lange, and Benjamin Rohrdantz

Atmos. Meas. Tech., 16, 295–309, https://doi.org/10.5194/amt-16-295-2023,https://doi.org/10.5194/amt-16-295-2023, 2023

Short summary

Michael Frech, Cornelius Hald, Maximilian Schaper, Bertram Lange, and Benjamin Rohrdantz

Viewed

Total article views: 287 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
170	108	9	287	4	3

HTML: 170
PDF: 108
XML: 9
Total: 287
BibTeX: 4
EndNote: 3

Views and downloads (calculated since 29 Aug 2022)

Month	HTML	PDF	XML	Total
Aug 2022	42	16	3	61
Sep 2022	30	21	0	51
Oct 2022	19	16	2	37
Nov 2022	37	18	0	55
Dec 2022	34	29	4	67
Jan 2023	8	8	0	16
Feb 2023	0
Mar 2023	0
Apr 2023	0
May 2023	0
Jun 2023	0
Jul 2023	0
Aug 2023	0
Sep 2023	0
Oct 2023	0
Nov 2023	0
Dec 2023	0
Jan 2024	0
Feb 2024	0
Mar 2024	0
Apr 2024	0
May 2024	0
Jun 2024	0
Jul 2024	0
Aug 2024	0

Cumulative views and downloads (calculated since 29 Aug 2022)

Month	HTML	PDF	XML	Total
Aug 2022	42	16	3	61
Sep 2022	30	21	0	51
Oct 2022	19	16	2	37
Nov 2022	37	18	0	55
Dec 2022	34	29	4	67
Jan 2023	8	8	0	16
Feb 2023	0
Mar 2023	0
Apr 2023	0
May 2023	0
Jun 2023	0
Jul 2023	0
Aug 2023	0
Sep 2023	0
Oct 2023	0
Nov 2023	0
Dec 2023	0
Jan 2024	0
Feb 2024	0
Mar 2024	0
Apr 2024	0
May 2024	0
Jun 2024	0
Jul 2024	0
Aug 2024	0

Viewed (geographical distribution)

Total article views: 292 (including HTML, PDF, and XML) Thereof 292 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 31 Aug 2024

Download

The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.

Preprint (2694 KB)
Metadata XML

Short summary

Weather radar data are the backbone of a lot of meteorological products. In order to obtain a better low level coverage with radar data, additional systems have to be included. The frequency range in which radars are allowed to operate is limited. A potential radar to radar interference has to be avoided. The paper derives guidelines how additional radars can be included into a weather radar network and how interferences can be avoided.


Total:	0
HTML:	0
PDF:	0
XML:	0