Methodology:

Our data set was derived from a reanalysis of SuperDARN plasma velocity data: Shore, R., Freeman, M., Chisham, G., Lam, M. M., & Breen, P. (2022). Dominant spatial and temporal patterns of horizontal ionospheric plasma velocity variation covering the northern polar region, from 1997.0 to 2009.0 - VERSION 2.0 (Version 2.0) [Data set]. NERC EDS UK Polar Data Centre. https://doi.org/10.5285/2b9f0e9f-34ec-4467-9e02-abc771070cd9. The Shore et al. (2022) reanalysis data covers the period 1997.0 to 2009.0 at 5-min resolution. The scientific motivation and details of the methodology of its production are described in Shore et al., 2021, https://doi.org/10.1029/2021JA029272.

We performed a regression analysis on the Shore et al. (2022) reanalysis data, which resulted in the forecast model regression coefficients published here. The regression analysis occurred via two steps described in Section 3 of Lam et al. (see References below), submitted to the journal Space Weather in January 2023:

Step A: see Section 3.2 of Lam et al. (2023). Regression of each component of the SuperDARN plasma velocity with respect to 5-min averages of the epsilon solar wind coupling function (Koskinen and Tanskanen, 2012 https://doi.org/10.1029/2002JA009283 ), IMF By, and a constant. This produced 3 regression coefficients (the slope in epsilon, the slope in By and the constant) for every location and month (3 step A coefficients x 559 locations x 144 months).

Step B: see Section 3.3 of Lam et al. (2023). Regression of each set of regression coefficients (each of size 559 x 144) produced in Step A with respect to 4 variables: sin x, cos x, the monthly mean value of f10.7, and a constant, where x = 2pi (tj - 79)/365.25 and tj is the day-of-year in the middle of the 12 months of the year (j = 1 to 12). This produced 12 regression coefficients for each velocity component, so 24 in total (3 step A coefficients x 4 step B coefficients x 2 velocity components).

Use of regression coefficients. The plasma velocity can be forecast (or hindcast) by reading in these 24 regression coefficients from two ACSII files and using them in Equations 6, and subsequently Equations 5 of Lam et al. (2023). An example of how this is done is given by the IDL programme 'read_BAS_convection.pro'. Please see Data structure and data format below for details. 

Data collection:

The authors gratefully acknowledge the use of SuperDARN data. The original plasma velocity observations were gathered using the northern hemisphere radars of the SuperDARN global array. SuperDARN is a collection of radars funded by the national scientific funding agencies of Australia, Canada, China, France, Italy, Japan, Norway, South Africa, United Kingdom, and the United States. The fitted Doppler velocities were processed from the original autocorrelation functions using version 4.5 of the radar software toolkit (RSTv4.5) and within that toolkit, fitting routine 'FitACF v2.5'. The fitted Doppler velocities were analysed using multiple data-interpolating empirical orthogonal function eigenanalyses for the period 1997.0 to 2009.0 at 5-min resolution, extending between the north magnetic pole and 30 degrees colatitude, as described in Shore et al., 2021 DOI: 10.1029/2021JA029272.

Data quality:

The SuperDARN radar observations were processed to remove ground scatter, and to eliminate measurements with too low power (lower than 3dB), or which had a poor-quality flag (identified in RSTv4.0). When binning the data, range gates below 11 and above 150 (where those values correspond to a multiple of the 45 km range distance from the radar array location) were not used, since these gave inaccurate locational estimates.

The Shore et al. (2022) eigenanalyses of the radar data were performed on a spatial grid with 559 locations, so we keep the same grid for consistency. We recombined all 10 spatial and temporal modes for the 12-year interval. The resulting velocities have NaN (not a number) values for co-latitudes greater than 30 degrees, and for the single point exactly at the pole. For this reason, our forecast model regression coefficients also have NaN values for co-latitudes greater than 30 degrees and the pole.