Quarterly Update: February 2021
DAWG charter
As discussed during the PIs/co-chairs meeting on 11 December, we have added a ‘do no harm’-type statement to the DAWG charter. We phrased this statement in terms of preserving the speed, functionality and usability of the software from release to release.
The updated charter is available here, and the new statement is shown in bold.
FITACF 3.0 White Paper
A white paper on FITACF 3.0 is now available on the DAWG website. This paper provides an overview of the FITACF method in general, and a description of the differences between FITACF 2.5 and FITACF 3.0. A second paper providing comparisons between the two fitting algorithms is in preparation.
Pasha Ponomarenko and Emma Bland are also preparing a manuscript on factors that affect the SuperDARN background noise estimation. Through this work we have been able to understand why the ACF selection (qflg=1
) behaves so differently for the two fitting algorithms for some radars, especially in relation to:
-
The number of “noise” echoes that are isolated in range and time
Multiple causes related to the noise calculation itself, control program design, substitution of the alternate lag zero power, and a data types problem in FITACF 2.5 -
Vertical “streaks” in range-time plots
Probably caused by interference from other HF transmitters, including other SuperDARN radars
Pasha has also written software to filter out these types of echoes from the fitted data. DAWG will begin testing this code soon.
FITACF file format: elevation angle fields for FITACF3.0
The elevation angle fields in FITACF-format files are used slightly differently in FITACF2.5 and FITACF3.0:
FITACF2.5 (documented here)
elv
is the elevation angle calculated from the fitted phaseelv_high
is the upper estimate of the elevation angleelv_low
is the lower estimate of the elevation angle
FITACF3.0 (documented here)
elv
is the elevation angle calculated from the lag zero phase of the ACFelv_high
is the elevation angle calculated from the fitted phase (cf. FITACF2.5elv
)elv_low
is the least square error
FITACF3.0’s use of the elv_high
and elv_low
fields to store other information is misleading. We suggest that two new fields, elv_fitted
and elv_error
, be added to the fitacf file format to capture the different output in FITACF3.0:
Field | FITACF2.5 | FITACF3.0 | Notes |
---|---|---|---|
elv |
Elevation angle calculated from the fitted phase | Elevation angle calculated from the lag zero phase | The elevation angle that is recommended for scientific use |
elv_fitted |
Elevation angle calculated from the fitted phase | The same as elv in FITACF2.5 |
|
elv_low |
Lower estimate of the elevation angle | Often meaningless due to phase flipping. | |
elv_high |
Upper estimate of the elevation angle | (see above) | |
elv_error |
Least square elevation angle error | Replaces elv_low /elv_high |
The above suggestion would not change the FITACF2.5 output. We will only implement this change with PI approval, since it potentially impacts the scientific interpretation of the data. Alternative suggestions are also most welcome. We have not yet investigated the practicality of implementing this change in the RST; we would like PI feedback first.
Removing version numbers from RST directories
We are planning to remove the version numbers from the directory structure in RST. For example, all of these directories would be renamed from [binary name].[version]
to just [binary name]
(e.g. make_fit.1.19
will become make_fit
). The reasons for this change are:
- It is easier to navigate around the RST source code without the version numbers
- Manually maintaining version numbers conflicts with the functionality of our version control software (git)
- Version numbers have not been updated consistently over the years, so they are meaningless
This does not apply to RST libraries where two parallel versions are maintained for scientific reasons, e.g. fitacf2.5
/fitacf3.0
, aacgm
/aacgm_v2
, elevation
/elevation_v2
.
Please contact the DAWG chairs by 31 March 2021 if you have any concerns about this change.
Software releases
RST
RST4.5 was released on 13 January. Please see the Zenodo site for a description of the new release and to download the software.
pyDARN
pyDARN 2.0 is out for testing and planned to be release mid February. The following changes to pyDARN are as follows:
- Removed SuperDARN DMAP and Borealis reading to pyDARNio
- Integration with pyDARNio to keep reading methods available in pyDARN
- Fan plots
- Statistical Power Plots
- Slant range option in plotting
- Utility to get a radar field of view in geomagnetic and geographic coordinates
- Update to obtaining current hardware files from https://github.com/SuperDARN/hdw
- Updated documentation on the following changes in pyDARN
- Documentation on pull requests and how users can test and review the code