Technical Details


In his new position at the Air Force Research Lab, Todd Parris with other researchers T.R. Pedersen and J.M. Holmes requested special time on the northern hemisphere SuperDARN radars to support a multi-instrument campaign including ionosondes, incoherent scatter radars, RF beacon receivers, GPS receivers, and optical instruments. The campaign’s objectives were to determine the evolution of polar cap patches and irregularity structures within them as they convect across the polar cap along a chain of heavily instrumented sites.

Todd writes in his request: SuperDARN measurements, especially if collected at high resolution and re-analyzed with the benefit of raytracing through data-based ionospheric reconstructions, will provide a critical measurement of irregularity locations for correlation with the larger scale structures as a function of patch age and location as determined by optical and other observations along the chain of ground stations. Todd outlined SuperDARN’s role in the bigger experiment as being able to observe and locate field-align irregularities, create 1 minute estimates of the global convection pattern, and provide information about ionospheric densities through ground backscatter.

After some discussion, it was agreed that Todd would write the control program necessary for the special time request but the individual operators could modify the control program to their needs. The special time request was mixed in with discretionary time to fulfill a request from Keisuke Hosokawa involving Pykkvibaer, Stokkseyri, Syowa South, and Syowa East. However, as it was noted to the schedule working group (SWG), other northern hemisphere radars should considering running per Todd’s special time request between 1800 utc and 2400 utc during the discretionary time. To make things simpler with radar modes, Kevin determined that the VT radars would run the Todd’s mode throughout the discretionary times.

The PCP Control Programs

The new control program, as outline in Todd’s request, would have a scanning scheme much like a regular 1 minute scan, sounding on a single frequency band. However, the time per sounding on each beam would be set such that the scan through the radar’s field of view would finish with about 10 to 12 seconds before the next minute boundary occurred. During the extra time, the radar would be set to a single beam and then 8, 1 second soundings would occur. Each of these 8 short soundings would use a different frequency band and perform a frequency sweep of sorts. The beam to be used during this frequency sweep mode would be determined per radar and would point towards other instruments being used in the campaign like Qaanaaq, Greenland or Resolute Bay, Nunavut. In order to observe the polar cap patches, Todd requested that the radars run with a smaller range gate separation, 15 km for the older radars and 6 km for the new MSI radars.

For the QNX4 style radars, the new control program was written using normalscan as a basis. Extra parts of code were added by Todd in order to perform the single beam frequency sweep soundings at the end of the field of view scan. Todd sent this code to Kevin in Dec. 2011 for testing and it was found that the normalscan code was several versions out of date. Kevin took a newer version of the normalscan code and was able to insert the single beam frequency sweep parts of the code Todd had written. This code then compiled and ran without any major problems at the Blackstone, Kapuskasing and Goose Bay radars. The new control program was called pcpscan and used a cpid of 9211 as given by Bill Bristow.

For the QNX6 style radars, the new control program was written using pcodescan as a basis. As with the QNX4 version, this code would have extra parts inserted in by Todd in order to perform the single beam frequency sweep soundings. Todd sent this code to Kevin in Dec. 2011 for testing on the Fort Hays radars and it was found that the mode compiled and ran on the radar without much of a problem. However at the time, with a 6 km range separation and 565 ranges, the data could not be plotted or visualized by the VT lab. As the size of the files was non-zero and not very small, it was believed that the control program was writing good files. The new control program was called pcppcodescan and used a cpid of 9213 as given by Bill Bristow.

In the weeks leading up to the special time, it was realized that Simon Shepherd, P.I. for the Christmas Valley radars, had been left out of the discussion on radars that would be operating with the new control program. Upon receiving the control program code on Jan. 5, Simon returned the code with several questions on Jan. 9. Simon believe that the code he had tweaked would run on the MSI radars and so his version of the code was compiled at these radars. However, without any discretionary time left before the special time, the new code was left untested. One key change made to the control program in the discussion leading up to the special time was to extend the range of the radar out to the usual 4500 km. This meant that the number of ranges would need to be increased to 750.

Special Time Operations and the Aftermath

The first requested special time period occurred on Jan. 15 between 0600 and 1200 utc. As this time period was in the early morning hours of a Sunday on the east coast, the radars were not observed in real-time. Of the VT radars, the older QNX 4 radars all ran without much problem and the data downloads showed that data collected during this time had some ionospheric returns in the northern radars. The MSI radars on the other hand did not run at all and no data was collected during this time. It was believed at the time that a small bug in Simon’s code could have caused the radars to stop operating. In order to try to fix the problem, on the morning (EST) of Jan. 15, Kevin took a few additions from Simon’s code and inserted them into code that had previously run on the Fort Hays radar. The semi-reverted code was compiled and the Fort Hays radars ran during the next special time period beginning Jan. 15 at 1800 utc. Simon abandoned the special time mode for Christmas Valley until the pcppcodescan control program could be further debugged.

During the later Jan. 15 time period and following days of special time, the data from the Fort Hays radars could not be plotted or visualized very well. The VT lab software appeared to break or crash upon trying to process and plot the data from the PCP periods. The only indication that the Fort Hays radars were collecting data appeared on the real-time java display. However, the data that was plotted here was not very good and could not adequately justify that the radar was collecting good data.

Feedback from Todd on Jan. 17 indicated that the real time display served up by VT was being observed for radar operations and data. However, the main use of SuperDARN during this campaign was to observe the real-time convection mapping as a way to predict where PCPs being observed would go and estimate the time it will take for a patch to travel from one ISR to another. This real-time convection mapping was served up by JHU/APL and Todd expressed disappointment that few radars were feeding data into this data product. At this time, Todd also noted that mid-latitude radars were included in the experiment because other colleagues had noticed the field-of-view boundaries extending into the polar cap and thus these radars should be included.

With a growing frustration on not being able to plot or visualize the data coming from the Fort Hays radars combined with increased solar activity, the special mode was abandoned around 1900 utc on Jan 23. Some debate was put forth to why we were running this mode and if we had lost data during rare geomagnetic storm periods. The mode was also abandoned at Blackstone since it was believe it was too far south to be good for polar cap patch detection. Upon the news of a massive CME arrival on around 0900 EST on Jan. 24, Ray requested that the special mode campaign also be abandoned for the high latitude VT radars, Kapuskasing and Goose Bay. The special mode was abandoned after the scheduled 1200 utc change over to a 1-minute common time.

In the weeks since the special mode campaign, some discussions have arisen on how new control programs and special modes from outside scientific communities are to be developed and tested. It was the feeling that during this campaign several steps of displaying the scientific merits of this mode were overlooked.

Also, in the time since the special mode campaign, VT has corrected some limiting factors on its processing and visualization software that originally limited its ability to monitor the data in semi-real-time. First looks at this data have showed that the Fort Hays radars were collecting data very well during the special time and have already showed some promise in contributing to areas other than detecting polar cap patches. The ground backscatter during this time shows the expected 6 dB degradation in power over a normalscan mode which was expected by Todd.

Lastly, several coding issues have been analyzed and corrected that caused the Fort Hays and Christmas Valley radars to not operate during the first special time period. Jef Spaleta of UAF, quickly spotted the factor that caused the MSI radars to not operate during the first special time period. This factor was a limiting variable definition that would not allow the radars to operate with more than 700 range gates. Jef has also found that putting command line arguments out of order in the schedule files could cause some of the arguments to be misinterpreted or not read altogether. Simon took a look at some of the code in the control programs and seems to have the issues with the out of order arguments sorted out. Some of these issues could have been come up during other operating times. Jef has been coordinating several tests in order to track down any other issues with the control programs or limiting variables.

Ongoing Discussions

- Procedures/methods for outsiders of SuperDARN to make use of time of SuperDARN radars.

- Data analysis from the Fort Hays radars as it relatively unseen.

- Code debugging, corrections, and small program logic changes.

- Who is allowed to analyze the data for scientific purposes?