**1. Introduction**

The ionosphere, situated between 60 and 2000 km above the Earth's surface, plays a vital role in electromagnetic wave propagation, influenced by solar-radiation-induced ionization [1]. The speed at which the transmitted electromagnetic signals from the GNSS (global navigation satellite system) satellites propagate through the ionosphere depends on the electron density along the line of sight between the satellite and the receiver. Upon traversing the ionosphere, GNSS signals may encounter two distinct forms of perturbations: Firstly, the introduction of an error in the estimated range due to the signal's delay that is proportional to the integrated electron density (slant total electron content—sTEC), and secondly, the occurrence of signal characteristic fluctuations resulting from irregularities in the ionosphere's electron density distribution [2].

The use of GNSS signals, renowned for their global availability and signal propagation characteristics, has been widely investigated and exploited as a powerful tool for

**Citation:** Moreno, M.; Semmling, M.; Stienne, G.; Hoque, M.; Wickert, J. Characterizing Ionospheric Effects on GNSS Reflectometry at Grazing Angles from Space. *Remote Sens.* **2023**, *15*, 5049. https://doi.org/10.3390/ rs15205049

Academic Editor: Mehrez Zribi

Received: 12 September 2023 Revised: 15 October 2023 Accepted: 17 October 2023 Published: 20 October 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

ionospheric studies across diverse spatial and temporal scales. Ground-based atmosphericsounding techniques employing continuously operating reference station (CORS) networks and GNSS receivers, which operate on low Earth orbit (LEO) satellites for the analysis of refracted radio signals via GNSS radio occultation (GNSS-RO), provide key observations for improving global weather forecasts [3]. To further broaden the observations, GNSS reflectometry (GNSS-R) has emerged as a complementary technique that leverages signals reflected off the Earth's surface. This approach not only facilitates the retrieval of reflecting surface properties but also serves as an atmospheric-sounding tool.

In order to understand ionospheric ranging delays within space-borne GNSS-R, simulations are conducted as detailed in [4]. The simulation is based on the Cyclone GNSS (CYGNSS) [5] mission and encompasses different elevation angles, latitudes, and solar activities. The results reveal an inverse relationship between the satellite elevation angle and ionospheric delay, with a larger ionospheric influence at low latitudes. In [6], the impact of scintillation effects on reflectometry has been explored using data from UK TechDemoSat-1 [7]. These effects lead to a degradation of the signal-to-noise ratio that can be utilized for altimetry and scatterometry performance assessments. More recently, studies have been carried out to retrieve the total electron content (TEC) from coherent reflectometry observations. In the work presented in [8], a methodology was introduced for sTEC estimation along the paths of incident and reflected signal rays. This estimation is based on coherent dual-frequency GNSS-R measurements obtained from Spire Global low Earth orbit (LEO) CubeSats. The outcomes have demonstrated a favorable alignment between reflectometry sTEC estimations and the global ionospheric TEC maps (GIM). Furthermore, an algorithm outlined in [9] combines sTEC observation from space-borne reflectometry using CubeSats and data collected from ground-based GNSS stations to generate vertical TEC (vTEC) maps in the Arctic region. Simulations conducted within this study under diverse conditions, involving variations in temporal resolution, solar activity levels, and the number of reflection events, have demonstrated enhanced accuracy in vTEC estimations when coherent GNSS-R observations are incorporated.

In the domain of GNSS-R, it has been empirically established that coherent observations are more frequently observed in the presence of smooth reflecting surfaces, such as sea ice, regions with low sea states, or inland waters, and at low grazing angles [10–12]. Nonetheless, within this range of elevation angles, it is important to note that the trajectories of the LEO GNSS-R rays entail a longer path through the ionosphere. This extended path results in a more pronounced ionospheric impact on the signals themselves. The representation (not to scale) of the LEO GNSS-R configuration along the grazing angle rays' paths and its interaction with the ionosphere are illustrated in Figure 1.

**Figure 1.** (**a**) LEO GNSS-R representation at 30◦ elevation angle at specular point. (**b**) LEO GNSS-R representation at 5◦ elevation angle. *sTECx* denotes the slant total electron content. Subscripts dr, in, and re correspond to the direct ray (transmitter *Tx* to receiver *Rx*), incident ray (transmitter to specular point *SP*), and reflected ray (specular point to receiver), respectively. *Hmx* represents the peak electron density height for the incident and reflected ray paths.

As described in [13], dual-frequency receivers possess the capability to mitigate these first-order ionospheric effects through the utilization of a linear combination (ionospherefree) of either code or carrier measurements. Conversely, single-frequency receivers must rely on applying a model to correct for ionospheric refraction, which can introduce delays of several tens of meters. For the Galileo GNSS constellation, the European GNSS Open Service has adopted the Neustrelitz Total Electron Content Model NTCM [14] (NTCM-G) or NeQuick 2 [15] (NeQuick-G) models to provide real-time ionospheric corrections for single-frequency receivers [16].

This study is in preparation for the European Space Agency's GNSS-R CubeSat mission "PRETTY" (passive reflectometry and dosimetry) [17]. The mission's primary goal is to retrieve sea surface height using grazing angle observations. Since PRETTY operates at a single frequency (L5), it requires model-based ionospheric corrections. This study provides a comprehensive characterization of ionospheric effects, at the grazing angle range (5◦–30◦), considering satellite geometry, latitude-dependent regions, temporal variations, and solar activity. It analyzes variability in the ionospheric group delay, Doppler shift, and peak electron density height. Additionally, the uncertainty in model-based ionospheric corrections for GNSS-R group delay altimetry is assessed.

The analysis is based on utilizing the sTEC obtained from three-dimensional, timedependent models. To assess model uncertainty, the sTEC values computed using the Neustrelitz Electron Density Model (NEDM2020) [18] are used as a reference and compared with the sTEC retrievals from NeQuick 2. Simulations are conducted to replicate conditions similar to those of the PRETTY mission, utilizing orbit data from the GNSS-R Spire Global Lemur-2 constellation. To provide a comprehensive analysis, the results are categorized into three elevation angle ranges: very-low (5◦–10◦), low (10◦–20◦), and mid-low (20◦–30◦). These categories are further grouped by latitude into three distinct regions: north, tropics, and south. Additionally, this study considers variations in local time and solar activity. Low solar activity (LSA) is represented by F10.7 = 75 in March 2021 and high solar activity (HSA) by F10.7 = 180 in March 2023.

The structure of this paper is outlined as follows: Section 2 presents the GNSS-R data descriptions, reflection events, and ray point settings for the simulations. Section 3 illustrates the methodologies utilized for the determination of parameters such as sTEC, relative ionospheric delay, Doppler shift, and ionospheric piercing points. Subsequently, Section 4 presents the results and analysis of the parameters explained in Section 3. Finally, in Section 5, a discussion of the findings is presented along with the conclusions in Section 6.
