Calibration and Validation of Swarm Plasma Densities and Electron Temperatures Using Ground‐Based Radars and Satellite Radio Occultation Measurements

Abstract

In this study we calibrate and validate in situ ionospheric electron density (Ne) and temperature (Te) measured with Langmuir probes (LPs) on the three Swarm satellites orbiting the Earth in circular, nearly polar orbits at ~500 km altitude. We assess the accuracy and reliability of the LP data (December 2013 to June 2016) by using nearly coincident measurements from low‐ and middle‐latitude incoherent scatter radars (ISRs), low‐latitude ionosondes, and Constellation Observing System for Meteorology, Ionosphere, and Climate (COSMIC) satellites, covering all latitudes. The comparison results for plasma frequency ( f∝Ne) for each Swarm satellite are consistent across these three, principally different measurement techniques. It shows that the Swarm LPs systematically underestimate plasma frequency by about 10% (0.5–0.6 MHz). The correlation coefficients are high (≥0.97), indicating accurate relative variation in the Swarm LP densities. The comparison of Te from high‐gain LPs and those from ISRs reveals that all three satellites overestimate it by 300–400 K but exhibit high correlations (0.92–0.97) against the validation data. The low‐gain LP Te data show larger overestimation (~700 K) and lower correlation (0.86–0.90). The adjustment of the Swarm LP data based on Swarm‐ISR comparison results removes the systematic biases in the Swarm data and gives plasma frequencies and high‐ and low‐gain electron temperatures that are precise within about 0.4 MHz (8%), 150–230 K, and 260–360 K, respectively. We demonstrate that the applied correction significantly improves the agreement between (1) the plasma densities from Swarm, and from ionosondes and COSMIC, and (2) the Te from Swarm LPs and International Reference Ionosphere 2016.

Steve: The Optical Signature of Intense Subauroral Ion Drifts

“…The measurements from the satellites are well suited for the study of ionospheric subauroral phenomena. While a direct intersatellite comparison has not been done for the Swarm satellites, scientific publications using multiple Swarm satellites (Archer et al, 2015;Goodwin et al, 2015;Spicher et al, 2015) as well as validation studies (Lomidze et al, 2017(Lomidze et al, , 2019 all show the Swarm satellites to behave similarly to one another. The Swarm satellites measure electric and magnetic fields, as well as electron density and temperature; see Jørgensen et al (2008) and Knudsen et al (2017) for more information on the Swarm instrument package.…”

Geomagnetically Conjugate Observations of Equatorial Plasma Irregularities From Swarm Constellation and Ground‐Based GPS Stations

“…Two Langmuir probes onboard of each Swarm satellite provide the electron density and temperature along the satellite track. The detailed data resolution, calibration, and validation of Swarm plasma density are summarized in Knudsen et al () and Lomidze et al (). Dual‐frequency GPS receivers are equipped on onboard the Swarm satellites for providing the GPS data, including the pseduorange, carrier‐phase, and C / N 0 measurements (van den Ijssel et al, ).…”

An Empirical Model of Vertical Plasma Drift Over the African Sector

“…The MAD is calculated as follows: where M i and M j are the median values for the i th absolute deviation and j th observations, X represents the observations and b is the numerical value (1.4826) describing the distribution of data assumed to be normally distributed (Huber & Ronchetti, ). The scaled MAD in Lomidze et al () is adopted, and therefore, for each satellite pass within the defined grid, the C/NOFS vertical E × B drift that was considered fell within the range of median±3 MAD. The spatial grid used throughout the analysis around the magnetometer station AAE (9.0°N, 38.8°E, geographic) is 0.18°N ± 4° geomagnetic latitude, 38.8°E ± 11.2° geographic longitude within the altitude range of 400–450 km.…”

“…The median and scaled MAD filtering technique removed 3% (as outliers) from the entire data set (2008–2014). The median filtering technique is the preferred option in removing outliers because it has a breakpoint of 50% (it starts failing only when observations consist of outliers reaching about 50%) and has recently been used in related analyses (e.g., Lomidze et al, ). In Figures b and d, the final average C/NOFS vertical E × B values are plotted as red dots.…”