GNSS radio occultation ionospheric electron density profiling
As a GNSS signal grazes the ionosphere during a limb occultation, its excess phase delay encodes vertical electron density structure. Abel inversion turns that delay into profiles of the F-layer, sporadic-E patches and total electron content, with honest caveats about where spherical symmetry fails.
Sensors
- COSMIC-2 / FORMOSAT-7: Six satellites in a 24-degree-inclination orbit, each carrying a Tri-GNSS Radio Occultation System (TGRS) receiver. The constellation produces roughly 5,000 ionospheric occultation profiles per day concentrated in the tropics and mid-latitudes. Vertical resolution of electron density profiles is approximately 1–2 km in the F-layer, degrading toward the topside. Data are distributed openly through UCAR COSMIC Data Analysis Center (CDAAC) with typical latency of 1–3 hours for near-real-time products.
- Spire Global LEMUR-2: A commercial constellation of over 100 CubeSats in multiple orbital planes, tracking GPS, GLONASS and Galileo signals. Spire claims global coverage with several thousand occultation soundings per day across all latitudes, including polar regions underserved by COSMIC-2. Vertical resolution is comparable to COSMIC-2 for well-sampled events. Data are licensed commercially and have been used in operational space-weather assimilation trials.
- PlanetiQ GNOMES: A small commercial constellation designed explicitly for high-signal-quality radio occultation, tracking all four GNSS constellations. Multi-frequency reception allows direct first-order ionospheric correction, which also yields differential TEC products. Vertical resolution in the E-layer, where sporadic-E is detected, can approach 0.5–1 km under favourable geometry.
- ESA Swarm: Three satellites in near-polar orbits at 430–530 km altitude. Swarm carries GPS receivers that produce topside ionospheric electron density profiles and in-situ plasma measurements. Particularly valuable for the topside F-layer and plasmasphere, complementing the bottom-side profiles from COSMIC-2. Data are openly distributed through ESA's Swarm data portal.
What the excess phase actually measures
When a GNSS satellite sets behind Earth's limb as seen from a receiver in low Earth orbit, the signal path descends through progressively denser layers of the ionosphere before entering the neutral atmosphere. Free electrons slow the electromagnetic wave's phase velocity relative to its group velocity. The resulting excess phase, measured differentially between two GNSS frequencies (typically L1 at 1575.42 MHz and L2 at 1227.60 MHz), is proportional to the integrated electron density along the ray path. That integral is total electron content (TEC), expressed in TEC units where 1 TECU equals 10¹⁶ electrons per square metre.
The geometry changes continuously as both the GNSS transmitter and the LEO receiver move. Each occultation event sweeps the ray path's tangent point vertically through the ionosphere over roughly 10–20 minutes, sampling a column of electron density from the topside F-layer down through the E-region near 100 km altitude. The resulting profile of excess phase as a function of tangent-point height is the raw observable from which all ionospheric products are derived.
Abel inversion: the assumption that makes profiles possible, and where it fails
Converting a one-dimensional phase profile into a vertical electron density profile requires Abel inversion, a mathematical technique that assumes the ionosphere is spherically symmetric around the tangent point. Under that assumption, the electron density at each height can be retrieved by peeling the profile from the top down, treating each shell as uniform. The method is computationally simple and works well in quiet mid-latitude conditions, where horizontal gradients are modest over the few hundred kilometres that the ray path samples.
The assumption breaks down badly near the equatorial ionisation anomaly (EIA). The EIA is a pair of electron density crests that form roughly 15–20 degrees either side of the magnetic equator, driven by the E×B plasma fountain effect. Horizontal density gradients across the EIA can be steep enough that the Abel-inverted profile produces artefacts, including spurious density peaks or layer height errors of tens of kilometres. COSMIC-2's tropical orbit concentration means it samples the EIA frequently, and users of its products in that latitude band should treat retrieved F-layer peak heights (hmF2) and peak densities (NmF2) with caution, particularly during post-sunset periods when the anomaly intensifies. Tomographic reconstruction using multiple simultaneous occultations can partially correct this, and published methods using ground-based ionosonde assimilation reduce the bias further, but no operational product eliminates it entirely.
Sporadic-E: a thin layer that causes outsized disruption
Sporadic-E (Es) layers are thin, dense patches of metallic ions, typically 1–3 km thick, occurring in the E-region between roughly 90 and 120 km altitude. They appear unpredictably over mid-latitudes in summer, driven by wind shear convergence acting on meteoric metal ions. Their electron densities can reach 10¹² electrons per cubic metre, sufficient to reflect HF signals up to 50 MHz and to cause amplitude scintillation on L-band satellite links.
Radio occultation is one of the few remote-sensing methods that can profile Es layers globally. The vertical resolution of GNSS-RO receivers (approximately 0.5–1 km for the best commercial systems) is just sufficient to resolve the layer's peak, though the retrieved thickness is often an overestimate due to horizontal structure within the layer. COSMIC-2 and Spire data have been used in published climatological studies of Es occurrence frequency, showing a clear summer mid-latitude maximum and a secondary tropical signal. For HF communications planners, a near-real-time Es occurrence map derived from RO data provides situational awareness that ionosonde networks, which are sparse outside Europe and North America, cannot match globally.
Scintillation, the F-layer and what satellite operators actually need to know
Plasma irregularities in the post-sunset equatorial F-layer cause rapid amplitude and phase fluctuations on satellite signals, a phenomenon called ionospheric scintillation. Scintillation degrades GPS positioning accuracy, disrupts satellite communication links and can cause receiver loss-of-lock. The S4 index, the normalised standard deviation of signal amplitude, is the standard metric: values above 0.3 indicate moderate scintillation and above 0.6 indicate strong disruption.
GNSS-RO receivers do not measure scintillation directly in the way that dedicated scintillation monitors do, but the amplitude fluctuations recorded during an occultation event carry information about irregularity structure. The spectral index of amplitude scintillation along the ray path can be estimated from RO amplitude data, and this has been correlated with ground-based S4 measurements in published studies. The practical value is geographic: RO constellations provide global sampling of irregularity activity, including over oceans and regions with no ground infrastructure, giving satellite operators a daily picture of where post-sunset equatorial plasma bubbles are active.
F-layer peak parameters, NmF2 and hmF2, are the most directly useful outputs for satellite operations. NmF2 determines the maximum usable frequency for HF skywave propagation. hmF2 affects the geometry of GPS positioning errors. Both are retrievable from Abel-inverted RO profiles with uncertainties of roughly 10–20% in NmF2 and 10–30 km in hmF2 under quiet conditions, worsening significantly during geomagnetic storms.
Operational products and their honest limits
Near-real-time RO ionospheric products from COSMIC-2 reach CDAAC within one to three hours of the occultation event. Spire's commercial feed can be faster, with latencies under one hour negotiated under data licence. Both are used operationally in global ionospheric map (GIM) assimilation by centres including JPL and the Ionosphere Prediction Service. GIMs produced with RO assimilation show measurably lower TEC errors over data-sparse ocean regions compared to GPS-ground-network-only solutions, though the improvement is modest during severe storms when the ionosphere changes faster than the assimilation cycle.
The key honest limit is horizontal resolution. Each occultation profile represents a horizontal average over a ray path that may extend 200–300 km in the ionosphere. Localised features, including individual plasma bubbles narrower than roughly 100 km, are smoothed or missed entirely. For space weather monitoring at the event scale, RO data is a complement to magnetometer networks, ionosondes and GNSS ground receivers, not a replacement for any of them. Satellize incorporates open RO data streams into space-weather situational awareness products for clients whose satellite operations span equatorial orbits; the Tonga crop programme, separately, has highlighted how ionospheric scintillation during the 2022 Hunga Tonga eruption temporarily degraded GPS-based field boundary data quality, which is an underappreciated operational dependency.
Matching the data to the decision
HF communications planners need NmF2 and hmF2 maps updated at least hourly, with Es occurrence flags at the E-region. Satellite operators need equatorial irregularity activity indices and TEC gradient maps ahead of contact windows. Space weather agencies need assimilation-ready profile feeds in RINEX or netCDF format with full uncertainty metadata.
None of these users needs the raw excess phase, but they all depend on the chain from phase to TEC to Abel inversion being applied correctly and its assumptions being flagged honestly. The most common mistake in procurement is treating a global ionospheric map as a high-resolution product. At 2.5-degree or 5-degree grid spacing, standard GIM products cannot resolve the EIA crest separation of roughly 30–40 degrees in latitude, which means the single most disruptive ionospheric feature for equatorial satellite operations is precisely the one that standard products represent least faithfully. Buyers should ask specifically for EIA-aware retrieval schemes or tomographic products when their operations concentrate between 20 degrees south and 20 degrees north.
Typical figures
| Vertical resolution (F-layer) | 1–2 km (COSMIC-2, Spire); 0.5–1 km (PlanetiQ GNOMES under good geometry) |
| Vertical resolution (E-layer / sporadic-E) | 0.5–1 km for best commercial receivers; retrieved layer thickness typically overestimated |
| Horizontal averaging length | 200–300 km along ray path in the ionosphere; not a point measurement |
| Global profile rate | ~5,000 per day (COSMIC-2, tropical/mid-latitude bias); several thousand additional (Spire, all latitudes) |
| Near-real-time latency | 1–3 hours (COSMIC-2 via CDAAC); under 1 hour (Spire commercial feed, licence-dependent) |
| Primary observing frequencies | L1 (1575.42 MHz) and L2 (1227.60 MHz) GPS; additional GLONASS, Galileo, BeiDou bands on newer receivers |
| NmF2 retrieval uncertainty (quiet conditions) | ~10–20% relative to co-located ionosonde |
| hmF2 retrieval uncertainty (quiet conditions) | ~10–30 km; degrades significantly during geomagnetic storms |
| Archive depth | COSMIC-1 from 2006; COSMIC-2 from 2019; Spire commercial from ~2016 |
| Delivery formats | netCDF (CDAAC standard), RINEX, ASCII ionPrf files; GIM products in IONEX format |
Analytics Satellize can run
| Vertical electron density profiles (NmF2, hmF2) | Abel inversion of differential TEC derived from dual-frequency excess phase; spherical symmetry assumed, EIA-region profiles flagged | Daily netCDF profile archive with per-profile quality flags; hourly updates for near-real-time feed |
| Global ionospheric TEC maps | Assimilation of RO TEC profiles with ground GNSS network data using thin-shell or 3-D model (e.g. IRI-based); IONEX-format output | Gridded IONEX map updated every 15–60 minutes; GIS-compatible raster layer |
| Sporadic-E occurrence and intensity maps | Detection of secondary amplitude and phase perturbations in the E-region during occultation events; thresholded against published Es detection criteria from COSMIC studies | Daily global Es occurrence map with layer intensity estimate; alert feed for HF communications planners |
| Equatorial plasma irregularity activity index | Spectral analysis of amplitude scintillation along occultation ray paths in the equatorial F-region; correlation with published S4 proxy methods | Nightly equatorial irregularity bulletin by longitude sector; satellite contact-window risk rating |
| HF maximum usable frequency (MUF) forecast support | NmF2 profiles ingested into ITU-R P.533 propagation model or IRI; hourly MUF maps for specified circuit paths | Circuit-specific MUF time series report; updated every 1–3 hours |
| Storm-time ionospheric disturbance assessment | Comparison of real-time RO-derived TEC and NmF2 against quiet-day climatology; anomaly detection against geomagnetic index thresholds | Event report with timing, affected latitude bands and estimated positioning error impact; issued within 3 hours of storm onset data availability |
Who does the work
We can get this done for you. Satellize runs its own analyst desk and a strong science team. You do not buy a data feed and work out what it means; our people source the imagery, run the analysis described on this page, and hand you the answer with its confidence limits stated. Discuss this requirement.