Dataset Open Access

Poker Flat Incoherent Scatter Radar (PFISR) Observations of E-region Neutral Winds

Stephen Kaeppler; Weijia Zhan

Updated: 12-15-2021


You are welcome to use the data 'as is', however, please inform me via email if you plan to use the dataset.  There are a number of small issues with the dataset that are best discussed.  We are interested in publications that use the data and derived values that are presented within the dataset.  If you plan to publish these results, please circulate a draft by me (SRK) and we would appreciate an offer of co-authorship or at minimum an acknowledgement.  You should include the NSF funding numbers NSF AGS - 1853408


As a general warning, the data from PFISR are quite noisy and you may need to perform significant averaging to produce usable results.  Again, please contact me and we can discuss this in more detail.

Version v0.6.4.2021.07.12 - This was the final processed version at the time that the final report was submitted to the NSF.


--------------- Previous from before ------------------

This file contains Poker Flat Incoherent Scatter Radar (PFISR) E-region Neutral Winds Data. These data correspond to monthly data files that include the E-region neutral winds and other parameters for the from March 2013-June 2019.

Publications of the Joule Heating Results:


Publication of Neutral Wind Results:

Hopefully we will have something in 2021. 


RAW ISR Data: These data were processed from the following files found in: and Please note that the error on the line of sight velocities may have been overestimated in these data and we scaled them by a eVLOS/sqrt(10).  Interested persons should contact Ashton Reimer or Roger Varney at SRI International for more information about these data, please see

Truthfully, the ISR data should eventually be reprocessed and then the winds algorithm run over it again.  This is a step for future work.


Processing Code is available upon request via email.


File Documentation:


Please see the change log:

Purpose: This is the overarching program and functions which process the
E region neutral winds from the fitted AC and LP data from PFISR.
This is a conversion fo which was originally written by
Nicolls into a more formal python class structure.

2017-10-05 - v0.2

The file has been validated against
using 20161121.001_ac_3min-fitcal.h5, 20170301.013_ac_3min-fitcal.h5, 20170302.001_ac_3min-fitcal.h5.
The program to run these is  At this point these  program match.
I am going to start diverging the code base, first subtly in the Joule Heating
since I found that Mike just looped over Nbeams, which isn't quite right, you need to loop
over the beams that were selected.

Changes from this point forward will produce different results.

2017-10-10 - v0.3.2017.10.10

Version v0.3, I made some IO changes but I may start processing some data with this version.

Version v0.4 - lots of small edits made to the IO and the plotting software.  It all seems to work
I have also included the SNR and Ne into the monthly plots and other information.
Made processing smoother.

03 13 2018 - added solar local time converion

v0.4.1 - 09 08 2018 added some ability to extract out the raw electron and SNR densities for each altitude bin
v0.4.2 - 10 15 2018 added in obtaining the F-region flows - want to check against the electric field.
v0.4.3 - 10 29 2018 added in some more altitude into the Joule Heating so I can make better figures
v0.4.4 - 11 20 2018 made some pretty major changes to IO to include consistent calculation of
Pedersen conductivity from  Made some changes to the Joule heating calculation and checked
formulas.  It is worth checking again.

v0.4.5 - 11 20 2018: added in Hall and Pedersen conductivities from fitted electron density data.
v0.4.6 - 12 03 2018: Tried to fix some of the double counting and time problems in testMakeMonthlyh5

05 22 2019: added some statements to bypass the geophysical parameters.  Also need in config file now.
            Additionally wrote in IOEregionwinds a try except statement

07 29 2019: Running the code for the 06 data reprocessed by Ashton

v0.5.0 - 10-15-2019: put in some filtering on the LOS velocity discharging bad Chi square and bad error codes on the fit.

v0.5.1 - 10-23-2019: changed chi square to 0.01 for lower boundary

v0.5.3 - 12-02-2019: Added in that now passing in the Chi2 and Fitcode filtering by Config file
            Bigger change that I am scaling the AC dVlos by some sort of factor while Ashton figures this out.
            We decided that a conversative scaling would be to reduce the dVLOS by 1/sqrt(10).
            The chi square produced in the data Ashton sent me typically was around 0.01, so the uncertaintiies on the LOS velocities
            may be over estimated.  So we are just changing this as a temporary fix while Ashton fixes the uncertainty estimation.

v0.5.5 02 01 2020 - Added in calculation of Coriolis, Centrifugal, and Lorentz forcing
v0.5.5 02 10 2020 - Added a correction to qvert so that way I can calculate the lorentz term.
                    Found an error where qvert = 0 in the if statement goes to false.

v0.5.5 02 15 2020 - Put in  nuInscaler into the main program, scaling ALL kappas by the scaler number

v0.5.6 02 28 2020 -- Added some more vlos diagostics and the calculation of the scale height. Added Altitude offset

v0.5.6.2020.03.12_nuin_fracoff - testing putting in the Brekke formula for ion neutral collision frequency and took out frac

v0.5.7.2020.04.10 - Put in Ashton's revised ion neutral collision frequency formulas into IO.
                    Also wrote a testscript and at least for the file I used was only different by 2.5%.
                    Revised where the mag data is being pulled from since the URL is deprecated
                    Added in Kappa which is now being interpolate - plan to see where kappa =1 is located for the paper.
                    commented out nuin scaler just so I am not chasing my tail

                    test v0.5.7.2020.04.13_org commented back in original ion neutral collision frequency method
                    possible mistake that not summing up properly.

                    test v0.5.7.2020.04.13_newnuin_orgsum_noTr800 - new formula for nuin except took off Tr>800.
                    I expect this should be almost the same as before since the formulas are basically the same.
                    did the original sum using frac[0] and frac[1] want to see if I am underestimating

                    v0.5.7.2020.04.13_newnuin_orgsum_yesTr800 - same as above except now including Tr>800.

                    'v0.5.7.2020.04.13_newnuin_newsum_noTr800' - using the new sum now and new col freq

                    v0.5.7.2020.04.13_updatedorg - updated original uses original method but including the NO term

v0.6.0.2020.04.15 -- Now think I have the new ion neutral collision frequency working and validated.
                    Found a mistake in how I was calculating the ion neutral collision frequency that
                    the fraction weight I was using only included the O+ and O2+ terms and not NO+
                    Turns out I was basically weighting by about 0.5, so I was effectively reducing the
                    ion neutral collision frequency by about a factor of 0.5 or less...
                    From this point forward need to start using any results from > v0.6
                    This revision has changed previous results signi

v0.6.0.2020.04.21 -- updated to now include the temperature correction for the O2+

v0.6.1.2020.04.23 -- made a number of changes to the geomagnetic files and reprocessed from CDAweb.
                     Wrote new code to be able to process the files from CDAweb in the new format.
                     Also changed the geomagnetic data files

v0.6.1.2020.06.07 -- changed the generation of Monthly files to hopefully be in order now
                    Added in missingIPY files given to me by ashton, maybe improve data covarege
                    Some work going to need to be done to make sure that all of the 10, 15, and 20 minute data are there.

v0.6.1.2020.06.15_Weijia -- Updated the data for Weijia's study in particular since we are missing a lot of IPY data for 02-04 2013 and 2014.

'v0.6.2.2020.07.01' -- Updated the data with new IPY27 mode for 2013 and 2014 Ashton processed.  Also now put in mechanical Joule heating term.
                        Put in the conductance and conductivity now too.

v0.6.2.2020.07.30 -- Made some changes to IO since Weijia noticed the mechanical heating terms were missing from the monthly files.

v0.6.3.2020.10.19 -- Tried to elimated all extra instance of nuinscaler, and also output that variable.  Added in variables
                    To get the Ti, Tn, ion neutral collision frequency along the vertical beam for diagnostic purposes
                    included dVest for F-region plasma drifts for Rafael

v0.6.4.2020.11.20 -- Extracted some more parameters including F107 and the Hall and Pedersen Drags

v0.6.4.2021.07.21 -- Final Run of data for NSF project


Files (16.3 GB)
Name Size
261.6 MB Download
366.1 MB Download
346.7 MB Download
189.6 MB Download
209.4 MB Download
369.6 MB Download
497.7 MB Download
495.5 MB Download
482.6 MB Download
196.7 MB Download
1.8 MB Download
14.2 kB Download
239 Bytes Download
1.3 MB Download
7.9 GB Download
298.5 MB Download
316.3 MB Download
441.3 MB Download
442.9 MB Download
462.2 MB Download
544.2 MB Download
435.2 MB Download
589.4 MB Download
580.0 MB Download
569.2 MB Download
256.5 MB Download
1.1 kB Download
All versions This version
Views 27887
Downloads 17378
Data volume 104.0 GB43.9 GB
Unique views 24179
Unique downloads 7849


Cite as