Ionospheric Effects on VHF Signals Transmitted from a LEO Satellite

Y. Jade Morton; Harrison Bourne; Steve Taylor,; Chun Yang

doi:10.33012/navi.743

Abstract

We investigated the effects of the ionosphere on a very-high-frequency (VHF) signal transmitted from a polar orbiting National Oceanic and Atmospheric Administration (NOAA) weather satellite, NOAA-19, at an altitude of 861 km. NOAA-19 gathers cloud images and transmits reduced resolution images in the Automatic Picture Transmission format at 137.1 MHz. A software-defined VHF data collection system sampled the signal at Gakona, Alaska, during a high-frequency ionospheric heating experiment. A software algorithm was developed to extract the signal amplitude, frequency, and carrier phase, and the results were used to compute amplitude and phase scintillation indices as the signal traversed the artificially disturbed ionosphere. A strong response was observed in signal amplitude and phase due to the disturbances, in clear contrast to the mild disturbances experienced by global navigation satellite system signals traversing the ionosphere under similar conditions. The results demonstrate the frequency dependence of ionospheric effects on radio waves transmitted from low-Earth-orbit satellites. The study also reveals the deficiency in current ionospheric scintillation modeling in the VHF regime.

Keywords

1 INTRODUCTION

Low-Earth-orbit (LEO) satellite-based navigation has received increasing attention in recent years (Reid et al., 2020; Kassas, 2020). Some applications, such as communication, Internet, and scientific mission applications, rely on signals transmitted from LEO satellites as signals of opportunity for navigation (e.g., Benzerrouk et al., 2019; Kassas, 2020). In other applications, dedicated LEO-based navigation signals have been developed to augment global navigation satellite system (GNSS) services (e.g., Ge et al., 2022; Reid et al., 2020). Regardless of the nature and original purposes of the signals, these LEO-satellite-transmitted signals must traverse the ionosphere to reach users on the surface of the Earth. The ionosphere introduces phase advance and group delay in the signal range measurements, which can be mitigated to varying degrees via dual-frequency measurements, correction services, or modeling, similar to techniques used for GNSSs. The more challenging aspect of the ionospheric propagation effect is due to irregularities in the ionospheric plasma, which cause signal amplitude and carrier phase fluctuations, collectively referred to as ionospheric scintillation (Morton et al., 2020).

Ionospheric scintillation is known to cause disruptions in the operation of GNSSs and communication receivers. Our earlier simulation study showed that for a given ionospheric configuration and radio wave carrier frequency, signals transmitted from a LEO satellite experience more frequent, deeper, and shorter fades and higher signal phase rate changes compared with those from a medium-Earth-orbit (MEO) satellite (Morton et al., 2022). In our recent simulation comparing the ionospheric effect on LEO satellite signals at the ultrahigh-frequency (UHF), very-high-frequency (VHF), L, S, and C bands, carriers at a lower frequency band such as VHF and UHF showed a greater impact than those at a higher frequency band (Sun et al., 2024). Moreover, by assessing the performance of a receiver carrier and code tracking loop on LEO satellite signals traversing ionospheric structures, we found that LEO signals are more prone to carrier cycle slips and loss of lock than the same signals transmitted from a MEO satellite (Xu et al., 2023).

There have been very few reported observations on the impact of ionospheric irregularities on signals transmitted from LEO satellites. While our simulation model has been evaluated using real data at the L band and we have confidence in its performance in carrier frequencies near the L band, some of the assumptions made in this model, such as assumptions regarding small-angle forward propagation and spatial coherency, may not be valid at lower frequencies. Our ongoing model analysis is addressing this issue. The objective of this study is to investigate the effects of ionospheric irregularities in the VHF band using real signals transmitted from LEO satellites.

Naturally occurring ionospheric plasma irregularities most frequently occur at high latitudes and in equatorial areas. We are currently establishing event-driven data collection systems to capture a wide variety of signals transmitted from various LEO constellations. However, a main challenge is to capture real signals traversing the ionosphere with a quantifiable disturbance level and spatial distribution. To address this challenge, we turned to the High-frequency Active Auroral Research Program (HAARP) facility located at Gakona, Alaska (latitude: 62.3911˚N, longitude: 145.1317˚W). HAARP can locally heat specific layers of the ionosphere with 3.6 MW of high-frequency (HF) radiation, digitally controlled by a 180-element phased array antenna. Figure 1 shows an image of the HF antenna array (left), a map of the locations of our GNSS scintillation monitoring receivers and the VHF data collection system (middle), and the radiation power gain pattern of the HF array when the beam is steered toward the local magnetic zenith (202˚ azimuth, 76˚ elevation at HAARP).

FIGURE 1

(Left) HAARP HF antenna array; (middle) locations of GNSS scintillation monitoring receivers and the VHF data collection system; (right) radiation power gain pattern of the HF array when the beam is steered toward the local MZ; Ant: antenna.

During a controlled ionospheric heating experiment, we captured real VHF signals transmitted from NOAA-19, a polar orbiting satellite at an altitude of 861 km. NOAA-19 is a weather satellite that gathers optical and infrared image data and transmits, among other items, reduced-resolution images in the Automatic Picture Transmission (APT) format at a carrier frequency of 137.1 MHz. The signals are right-hand circularly polarized (RHCP) and frequency-modulated (FM) with a bandwidth of 37 kHz. Other NOAA weather satellites, such as NOAA-15 (137.62 MHz) and NOAA-18 (137.9125 MHz), operate in the 1-MHz bandwidth between 137 and 138 MHz. The intensity of an image pixel is quantized into 256 levels (8-bit resolution). Each image line becomes an amplitude-modulated (AM) signal, which is FM-modulated onto the carrier. Figure 2 shows a picture of the satellite (left), an image taken by the satellite (middle), and an example set of orbit tracks (right). Table 1 summarizes the satellite orbit and signal transmission parameters.

FIGURE 2

(Left) The NOAA-19 satellite; (middle) an image taken by NOAA-19; (right) an example set of orbit tracks; acquired from NOAA (n.d.).

View this table:

TABLE 1 NOAA-19 Satellite Orbit and Signal Parameters

We established a software-defined VHF data collection system at HAARP to collect the NOAA-19 signal during a HAARP ionospheric heating campaign in August 2023. A controlled heating experiment was specifically designed to disturb the signals transmitted by NOAA-19 on August 11, 2023. The data were processed to quantify the signal amplitude and phase disturbance level. Intense levels of amplitude scintillation and strong phase scintillation were observed on the VHF signals during the heating experiment. As there were no GNSS satellites in the vicinity of the NOAA-19 satellite, a direct comparison of the ionospheric effects between the L band and VLF band is not possible. However, a GNSS receiver co-located with the VLF receiver was used to gather GNSS signal disturbances during other heating experiments throughout the campaign. The results obtained by processing the VHF and GNSS signals during the heating experiments are described in this paper.

2 NOAA-19 TRANSMISSION SIGNAL STRUCTURE

Let x(t) be the NOAA-APT signal arriving at, say, an RHCP quadrifilar helix antenna; this signal is a constant-modulus sinusoidal wave that is FM-modulated by a voltage message signal s(t) as follows:

$x (t) = A \cos (2 π \int_{0}^{t} (f_{c} + f_{d} (τ) + k_{ν c o} s (τ)) d τ + θ_{0}) + n (t)$ 1a

where A is the signal amplitude and f_c is the carrier frequency set at 137.1000 MHz, 137.6200 MHz, and 137.9152 MHz for NOAA-19, NOAA-15, and NOAA-18, respectively, f_d is the Doppler frequency shift, θ₀ is the initial phase, and n(t) is assumed to be white zero-mean Gaussian noise.

In addition to the Doppler shift f_d, the carrier frequency f_c is varied by the quantity $k_{v c o} s (τ)$ , with $| s (τ) | \leq 1$ being the voltage and $k_{ν c o} = 15 k H z / V$ being the voltage-to-frequency gain of the underlying voltage-controlled oscillator (VCO), which determines the instantaneous frequency deviation. The voltage message signal s(t) is AM-modulated as follows:

$s (t) = a (b + m (t)) \sin (2 π f_{a m} t)$ 1b

where f_am = 2400 Hz is the AM carrier frequency, m(t) = 0 ~ 255 is the modulating message, an 8-bit symbol (an image pixel) with a baud rate of 4160 symbols per second, and a = 1/274 and b = 19 are design parameters such that s(t) has an amplitude of unity and the resulting AM modulation index (MI) is 87%.

Note that the case in which m(t) = 0 or 255 corresponds to the minimum amplitude of 19 or the maximum amplitude of 274, respectively, without considering the normalization factor a. The AM-MI is given by (max – min)/(max + min) = (274 – 19)/(274 + 19) = 87% in this case. However, the signal specification allows the AM-MI to vary by up to 5%.

We note that the baud rate of 4160 symbols per second is close to twice the AM carrier frequency of 2400 Hz. In this case, almost every half cycle of the AM carrier is modulated by a new pixel value.

According to Carson’s rule, the bandwidth of an FM signal is twice the sum of the peak frequency deviation and highest modulating signal frequency, which is 2 × (15 + 2.4) ≈ 34 kHz for NOAA-APT signals.

Aboard each NOAA Polar Operational Environmental Satellite (POES) is an advanced very high-resolution radiometer (AVHRR), which is a scanning sensor with up to six sensing channels in the visible, near-infrared, mid-infrared, and thermal-infrared portions of the electromagnetic spectrum. The AVHRR operates at a rate of 6 scans per second, covering a cross-track swath of 3600 km per scan. Two of the six channels (referred to as Sensor A and Sensor B in Figure 3) are chosen to generate two reduced-resolution image lines in the first two scans in the first half second and two more image lines in the fourth and fifth scans in the second half second, one for Video A and the other for Video B, respectively, as shown at the top of Figure 3.

FIGURE 3

Block diagram of image formation, signal modulation, and transmission of the signal from NOAA POES, as well as ground receiver signal reception and processing; ADC: analog-to-digital converter; AMP: amplifier; BPF: bandpass filter; LNA: low-noise amplifier; LPF: lowpass filter; RFBW: radio frequency bandwidth; SMM: space and minute marker; TLM: telemetry.

The analog signals of the two chosen channels from the AVHRR scanner are sampled into a stream of 909 symbols quantized into 8 bits (resulting in an image pixel with 0 for black and 255 for white), effectively corresponding to a pulse amplitude modulation (PAM). Each pixel has a spatial resolution of approximately 4 km in the along-scan direction and 3.3 km in the along-track direction on the ground. To these 909-pixel sensor data for each sensor, a 39-pixel synchronization sequence and a 47-pixel space and minute marker are pre-pended, and a 45-pixel telemetry wedge is post-pended. This step leads to an image line of 2080 pixels over 0.5 s, where the first half of the image line with 1040 pixels for Video A is placed ahead of the second half of the image line with 1040 pixels for Video B.

This image line constitutes the message m(t) to be transmitted, which is first AM-modulated as in Equation (1b) onto s(t) and then FM-modulated as in Equation (1a) onto the carrier x(t), as illustrated in the middle of Figure 3.

3 DATA COLLECTION SYSTEMS AND EXPERIMENT

a Data Collection Systems

We fabricated a software-defined VHF data collection system to collect signals transmitted by NOAA-19. An RHCP quadrifilar helix antenna is used to receive signals from 135 MHz to 139 MHz. A USRP reconfigurable radio frequency (RF) front-end has its local oscillator tuned to the NOAA-19 transmission signal center frequency. The resulting zero-intermediate-frequency (IF) signal is sampled at a rate of 0.3072 Msps. An algorithm was developed to extract the signal amplitude and carrier phase from the RF front-end output samples, and the results are used to compute the power spectrum density (PSD) and amplitude and phase scintillation indices. The VHF data collection system is set up on a science pad inside the HAARP facility, as shown in the middle panel of Figure 1. Figure 4 presents photographs of the antenna and RF front-end.

FIGURE 4

(Left) RHCP quadrifilar helix antenna with a bandwidth from 135 MHz to 139 MHz; (right) RF front-end with an oscillator tuned to 137.5 MHz and a sampling rate of 0.3072 Msps.

During the experimental campaign, we also utilized three Septentrio PolaRxS ionospheric scintillation monitoring (ISM) receivers that generate 100-Hz carrier phase measurements at all open GNSS bands. The carrier phase measurements are used to compute the total electron content (TEC) and scintillation indices. The middle panel in Figure 1 shows the locations of the three ISM receivers.

b Receiver Signal Processing

The received RF signal is down-converted to the baseband for a given satellite and is then sampled and quantized into baseband samples for further processing, as shown in the bottom of Figure 3. To produce an image, the baseband samples are first FM-demodulated to produce the instantaneous frequency to extract the modulating signal s(t). The envelope of the AM signal is then rectified into a stream of image pixels m(t). The synchronization sequence is used to detect the start of an image line and to obtain a symbol Doppler estimate. An image is formed after the image lines have been segmented out with Doppler compensation to align the pixels for each line. The pixel values of the first eight telemetry wedges of the least noisy frame (128 lines) are compared to the known reference values (a stair with a step size of 1/8 of the full-scale value 256) such that the pixel values are formatted to be between 0 and 255, thus forming a black and white image.

The details of signal processing utilized for this paper are further illustrated in Figure 5. For the initial signal analysis, time-domain waveforms can be displayed for selected time intervals. The PSD of the signal spectrum is calculated every second using Welch’s method. The time-frequency spectrogram can be shown in the waterfall format for the entire data file. These processing steps are shown at the top (green blocks) of Figure 5.

FIGURE 5

Block diagram of signal analysis and amplitude and phase scintillation calculation; Rx: receiver; Tx: transmitter

For amplitude scintillation analysis, we utilize baseband signal samples. The signal power is calculated per sample and then averaged per second. The combined effect of the transmitter and receiver antenna gain patterns is first estimated by fitting a quadratic function to the averaged sample power profile and then removed from the averaged sample power. The remaining power fluctuation is quantified in terms of its mean-normalized variance, which corresponds to the amplitude scintillation index (S4). The processing steps are shown in the middle (blue blocks) of Figure 5.

For phase scintillation analysis, we utilize an aligned and scaled image. The sync-alignment removes the Doppler frequency while telemetry-based scaling allows us to more accurately relate pixel values to the frequency deviation or phase change per pixel duration. As shown in the top right of Figure 3, the telemetry wedges occupy 45 pixels in a staircase of 16 levels (8 image lines per level) that establish a relationship between the received signal amplitude and the nominal amplitude, which is then used to scale the received signal sample values via interpolation. However, any other operations for image quality enhancement such as histogram-based equalization, median filtering, and temperature compensation for a particular sensor are undesirable for our analysis, as such operations may distort the phase information introduced by the ionospheric disturbance.

As shown in the bottom (orange blocks) of Figure 5, lines of pixels are first extracted from the image after the sync, space and minute marker, and telemetry pixels have been removed. The pixel values are converted to phase changes during a pixel duration. Note that the content of an image causes variations in pixel values, which must be removed so that the remaining variations in pixel values are solely produced by noise and ionospheric effects. These variations are removed by finding reference lines to represent the background image content and subtracting these lines from the remaining lines of interest. Assuming that the background content does not change significantly during a short time period, we obtain the reference lines from the image itself. Clearly, this method is only an approximation. As in the case for amplitude scintillation analysis, the antenna gain pattern (main lobe and side lobes) affects the received signal strength, which leads to different phase variations. The effect is first estimated by fitting the average phase profile and then removed. Finally, the phase scintillation index is calculated in terms of the standard deviation of the average detrended phase variations. This process is illustrated with examples in Figures 11 and 12.

FIGURE 11

Frequency deviations along an image line (pixels) for the three segments shown in Figure 6

The three red arrows next to the Segment II image indicate three image lines (283, 285, and 307) that we used to generate the average intensity/frequency plot in Figure 12.

FIGURE 12

(Left) Reference image lines obtained by averaging the three image lines (283, 285, 307) marked in the Segment II image from Figure 11; (right) frequency deviation between Segment II and the reference image lines

c Artificial Heating Experiment

To generate scintillation signals, the outputs of the 180 HF dipole antennas (in a 15 × 12 rectangular array) are combined to form an intense beam that can create sufficient energy to penetrate the F-region ionosphere at desired directions under appropriate atmospheric and ionospheric conditions. We predicted that the NOAA-19 satellite traveled close to the MZ over HAARP on August 11, 2023, at approximately 05:19:30 Coordinated Universal Time (UTC; 21:19:30 local time). This was a geomagnetically quiet day. The local MZ is the point in the sky where the local magnetic field lines point directly downward (in the northern hemisphere) or upward (in the southern hemisphere), perpendicular to the Earth’s surface. The MZ is known to be the most efficient direction for generating disturbances in the ionosphere (Pelgrum et al., 2011). Figure 6 illustrates the experimental setting. In the sky plot centered at the receiver antenna (shown in the left of Figure 6), the passage is time-tagged with the time point in UTC (red text) and the relative time from the beginning of the data recording (blue text). As an advantage, the recording time can be related to an image line for a precise analysis of events. Table 2 lists the UTC time, local time (Alaska Daylight Time [AKDT]), the corresponding NOAA-19 satellite azimuth and elevation, and the second count, with 0 corresponding to 05:12:00 UTC, which is the beginning of the VHF receiver data recording. The main beam footprint of the array (gain of -6 dB) and local MZ are also marked in the sky plot.

FIGURE 6

Illustration of the experimental setting and field data collection scheme

View this table:

TABLE 2 UTC, AKDT, and Relative Time Since Data Recording and NOAA-19 Azimuth and Elevation at Several Critical Time Points

At the test site, the geomagnetic field has a dip angle of 76˚. The time of the closest approach to the MZ occurs at approximately 450 s after the start of recording. For maximum effect, the heating was started 1.5 min prior to the MZ passage, at 360 s, and ended half a minute after the MZ passage, at 480 s. Note that the heating start time at UTC 5:18:00 corresponds to the local time of 21:18:00 in Alaska (8 h behind UTC), and soon after, at 400 s, a day/night switch occurred from a visible light to an infrared sensor, which is reflected in the left image (Video A), whose background intensity is shown in Figure 7.

FIGURE 7

Image transmitted by the NOAA-19 satellite during heating

The satellite-transmitted image was captured by two sensors: picture A is from Sensor 1, and picture B is from Sensor 2. Sensor 1 went through an operation mode transition at approximately 400 s. Three image segments were extracted for comparison, as shown on the right: before (I), during (II), and after (III) the signal entered the main beam. Speckles are visible in Segment II. Some residual effects can be seen in Segment III after the signal exited the main beam. Segment I is free of disturbances.

A local riometer array measurement showed that the atmospheric absorption was relatively low at the time of the satellite passing, indicating that the HF heating could penetrate the ionosphere. We tuned the HF array transmission to the fifth electron gyrofrequency at 7.7 MHz, based on an analysis of local ionosphere plasma frequency distributions retrieved from co-located ionosonde measurements. The HF beam was transmitted in the ordinary mode (O-mode) with a 100% duty cycle toward the MZ for 2 min, starting at 05:18:00 and ending at 5:20:00 UTC.

The ISM GNSS receivers collected data continuously throughout the campaign. There were no GNSS satellites near the MZ at the time of the HF heating. However, there were GNSS satellites near the MZ at other heating times when NOAA-19 was not in the area. GNSS data were collected during those times.

4 SIGNAL PROCESSING RESULTS AND ANALYSIS

a Transmitted Image Corruption

Corruption is clearly visible in the transmitted image as the signal traversed the heating beam main lobe. Two images were obtained from NOAA-19: image 1 switches between near-visible (0.86 μm) and mid-wave infrared (3.75 μm) depending on whether or not the ground is illuminated by sunlight; image 2 is long-wave infrared (10.8 μm) (Sigidwiki.com (2025)). Figure 7 presents the images captured by the two sensors side-by-side. Owing to the local evening time of the experiment, Sensor 1 changed its mode of operation at ~400 s, as indicated by the background color change from yellow to blue.

b Amplitude Disturbance

Signal amplitude disturbances can be observed in the PSD, where the signal power is redistributed to different spectral components. In other words, large amplitude changes are reflected in the PSD plot as a power surge, enhancing our interpretation of the plot. Figure 8 presents the PSD time series, where a PSD profile is generated once every second via Welch’s method with a 4096-point fast Fourier transform, showing the characteristic pattern of FM with 17 sidebands or tones in this case. More specifically, the complex signal samples over 1 s (307,200 samples) are divided into 100 segments with 3072 samples per segment. Overlapping of 50% between segments is used together with the Hamming window. The display frequency spacing is 75 Hz. The PSD disturbance appears to start shortly after 360 s, corresponding to the time when the heater was turned on (05:18:00), 6 min into the data collection. At 400 s, there is a dramatic PSD disturbance, which is due to the Sensor 1 day/night mode change. At 450 s, the signal is propagating nearly parallel along the magnetic field line at the MZ and is in the center of the heating beam main lobe; at this time point, some prominent structures can be seen in the PSD. The disturbance tapered off at approximately 480 s, when the heating stopped. Notably, there is a null at zero frequency across all times. This result is due to the receiver front-end having a zero-IF design with a direct-current (DC) notch. We will redesign the front-end in the future to avoid this issue.

FIGURE 8

PSD of the received NOAA-19 signal during the heating experiment

Two PSD cuts at 2.775 kHz and 6.525 kHz are shown in the top left, showing the detailed structure of the PSD response to heating.

The amplitude disturbance can also be observed in the in-phase/quadrature (I/Q) channel magnitude. Figure 9 displays time series of the I/Q channel magnitude and signal amplitude for several time points of interest. Prior to 440 s, the signals appear to have a constant modulus, as expected for an FM signal. From 445 s to 500 s, especially near 450 s, large amplitude variations can be observed, whose appearance is consistent with amplitude scintillation. Some residual effects can also be observed in the I/Q magnitude disturbance, which lasted until approximately 500 s, as shown in Figure 9.

FIGURE 9

Time series of I/Q channel magnitude and signal amplitude at several time points of interest

We extracted the signal amplitude from complex samples of the received FM signal, which ideally has a constant modulus, and computed the average and standard deviations of the signal power per second, as shown in the top left panel of Figure 10. Roughly, the average received signal power as a function of time (blue curve in the top left of Figure 9) traces out the antenna gain pattern of the satellite transmitter as it sweeps by the receiver with an omnidirectional hemispheric antenna pattern facing toward the sky. Consistent with the images in Figure 7 and the spectrogram in Figure 8, the signal exhibits significant power between 350 s and 550 s (approximately 3.3 min), which is shown above the dashed line in the top right panel of Figure 10. The amplitude scintillation index S4 was computed based on the signal amplitude using the technique described by Niu et al. (2012), as shown in the bottom right panel of Figure 10. Namely, a quadratic form is fit to the main lobe response, as shown by a red line in the top right panel of Figure 10. This fitted curve is then removed to generate the “detrended” power. The detrended power is then normalized by the mean power level, which is represented by the trend. The normalized signal power in dB is plotted in the bottom left panel of Figure 10. Both the signal power average and standard deviation values are then calculated over a sliding window of 10 s; their ratio corresponds to the S4 index, shown in the bottom right panel of Figure 10. The S4 index peaks inside the heating beam main lobe around the time when the signal traverses the main beam near the MZ, indicating that the time and location of the maximum disturbance effect are associated with the heating. At the peak disturbance, the S4 index value is approximately 0.8.

FIGURE 10

(Top left) Average and standard deviation (std) of the signal power, computed every second; (top right) curve-fitting to the antenna main lobe; (bottom left) detrended and normalized power over the main beam in dB; (bottom right) amplitude scintillation index S4 during the heating experiment.

The peak S4 index occurred when the signal propagated nearly parallel to the magnetic field line at the MZ at the center of the main beam.

c Phase Disturbance

To obtain phase measurements from the NOAA-19 signal, we applied an APT modulation scheme, as described in detail here. In the scheme, each image has two pictures (A and B), side by side, of the same scene captured by Sensors 1 and 2, respectively. Each picture line represents a scanned swathe of the Earth’s surface. The analog image signal is 8-bit quantized, with 0 corresponding to black and 255 corresponding to white, and discretized into 1040 pixels over 0.25 s per picture line. Of the 1040 pixels, 909 pixels are used for the image content and 171 pixels are used for line synchronization and telemetry, respectively. In total, there are 2080 pixels per image line over 0.5 s. The resulting signal is first modulated as a 2.4-kHz AM signal and then modulated again as an FM signal with a 34-kHz bandwidth on the carrier at 137.1 MHz.

For demodulation at reception, the instantaneous frequency deviation from the carrier frequency during a pixel time interval of 0.24 ms (0.5 s/2080 = 0.24 ms/ pixel) represents the intensity of the image pixel. That is, a frequency deviation in the range of [0, 34 kHz] is mapped to a pixel value in the range of [0, 255]. Given a pixel value, the corresponding phase change in degrees over the pixel time interval (i.e., the instantaneous frequency over the pixel) can be obtained with the following conversion scaling factor:

$34 kHz / 256 * 0.24 ms * 180 ° / π = 1.83 ° / unit pixel value$ 2

The signal’s instantaneous frequency reflects the combined effects of image content intensity, noise, Doppler frequency, and phase changes due to propagation through plasma structures created by heating. Most of the effects have relatively smooth frequency deviations, except for the ionospheric propagation effects if an area is filled with turbulent structures. We computed the signal frequency deviations from the carrier by first mixing the received RF signal with the nominal carrier frequency and then calculating the instantaneous frequency of the lowpass-filtered complex signal samples with either a numerical differentiation or a phase-locked loop method. Figure 11 shows the frequency deviation results for the three segments defined in Figure 7. The levels of frequency deviations for these three segments are clearly different. Segment I is taken from a time period when the signal is outside the main heating beam and its frequency deviations are relatively clean. Segment II corresponds to a time period when the signal traverses the main lobe of the heating beam and its frequency shows wild swings. Segment III shows mild disturbances but there appears to be a frequency bias associated with the Sensor 1 measurements. This bias could be due to the gray stripe of an unknown origin observed in Figure 7.

To isolate the ionospheric effects caused by the heating experiment, we assumed that the image content does not change substantially from one segment to another, which is evident in this particular area as the major vertical features run across the image lines and we can therefore subtract the clean image lines from the disturbed segment lines. The retained portion of the continuously transmitted image with better quality (less noisy without spikes) has 400 lines (0.5 s per line) corresponding to the recording time from 350 s to 550 s. We selected lines 283, 295, and 307, which are free of large spikes, as a reference, as these lines exhibit the average trend of the three segments in Figure 11. These three lines are marked on the Segment II image. The average intensity/frequency is used as a reference, shown in the left plot of Figure 12, where the sync, minute marker, and telemetry pixels are omitted. The right panel in Figure 12 shows the frequency deviations of the disturbed image segment after the reference image lines shown in the left plot have been subtracted. Notice that the Sensor 1 outputs (left portion of the plot) contain larger positive deviations, whereas the Sensor 2 outputs (right portion of the plot) contain larger negative deviations. The reason for this asymmetrical response is unknown.

The carrier phase is directly measured from complex samples of the baseband signal, and changes in carrier phase during a pixel time interval provide an estimate of the instantaneous frequency or the frequency deviation from the carrier center frequency (the FM part). The envelope of the frequency deviations (the AM part) is taken as the image pixel content (intensity). Inversely, we computed the carrier phase change per pixel by subtracting the reference image content from the measured frequency deviations along an image line (akin to detrending used by Niu et al. (2012)) and the standard deviation of such pixel phase changes per line. The results are plotted in the left panel of Figure 13. The phase scintillation index was computed as the standard deviation of the carrier phase computed over 2080 pixels per line (0.5 s), as shown in the right panel of Figure 13. As shown, the phase scintillation index indicated high levels of disturbance between 440 and 460 s, when the signal propagated through the main lobe of the heating beam. This result is in agreement with the amplitude scintillation results.

FIGURE 13

(Left) Carrier phase variation associated with each pixel in the NOAA-19 image during the heating experiment arranged in lines (0.5 s per line in the x-axis) and pixels (2080 pixels per line in the y-axis); (right) carrier phase scintillation index (carrier phase standard deviation calculated over pixels per line).

The spikes in the left plot and the large phase scintillation index values in the right plot are primarily associated with times when the signal was propagating through the main lobe of the heating beam.

The observed frequency deviation, with a maximum range of 34 kHz, was scaled to a pixel value, with the full scale being 0–256. The pixel value was then scaled to the phase change in degrees over a pixel, which lasts 0.24 ms. The results are plotted in the right panel of Figure 13. In this study, we assumed that the same scaling factor (34 kHz to 256 over 0.24 ms) was used for both transmission and reception.

Doppler frequency shifts may not have been perfectly removed, which may explain why some larger values are observed toward the end while smaller values are observed in the middle, corresponding to a minimal Doppler shift as the satellite flies overhead. Imperfect detrending with reference image lines may be another source of error, which we will work toward reducing in future work.

d GNSS Signal Response

We have conducted numerous experiments in which a heating beam is directed toward a GNSS satellite to observe the TEC and amplitude and phase scintillation responses. Pelgrum et al. (2011) described some of these previous experimental results. The magnitude of TEC disturbance varies, depending on the background ionosphere and the position of the GNSS satellite. The strongest disturbances occur when a GNSS satellite is over the MZ. The largest observed TEC disturbance is approximately 0.5 TEC units, and the largest S4 index is 0.15.

The current work qualitatively validates an ionospheric scintillation simulation study (Sun et al., 2024). In that study, simulated GNSS signals transmitted from realistic GNSS orbits and GNSS-like signals transmitted from various LEOs over a wide range of bands (VHF, UHF, L, S) traversed the same ionosphere plasma structures. The results showed that for the same ionosphere plasma structure, signals transmitted from LEO satellites have more frequent and deeper fades compared with those from MEO satellites, and lower-frequency signals transmitted from the same orbits experience stronger scintillation. The simulator was based on earlier studies that defined ionospheric structures as a phase screen characterized by a space-to-time scale factor and turbulence strength (Rino, 1979; Rino et al., 2018; Xu et al., 2020). The space-to-time scale factor was determined by the structure spatial scale and the effective signal scan velocity across the structure. A joint analysis of signals transmitted from the same satellite at different frequencies traversing the same ionospheric structure could lead to a better understanding of the scale sizes of these structures; such efforts are ongoing.

5 CONCLUSIONS AND FUTURE WORK

This paper has presented the results of an experiment to artificially disturb the ionosphere in a controlled manner to enable observation and quantification of VHF radio wave scintillation for signals transmitted from NOAA-19, a LEO satellite orbiting in the upper region of the ionosphere. The satellite transmitted FM-modulated images at a carrier frequency of 137.1 MHz within a bandwidth of 34 kHz. We observed image corruptions and intense scintillation, with the amplitude scintillation index reaching 0.8 and phase scintillation standard deviations reaching near 45˚ when the signal propagated through the main lobe of the heating beam. Similar levels of heating power directed towards GNSS satellites in the same areas of the local sky only yielded a fraction of TEC changes and did not produce noticeable scintillation effects.

Future heating experiments are already being planned to further quantify the effects of the ionosphere on signals transmitted from LEO satellites. In future experiments, we will modify the front-end to avoid the impact of the zero-IF configuration on PSD observations. We will also conduct heating experiments under different background atmospheric and ionospheric conditions, such as conditions exhibiting various background TEC and TEC gradient distributions, solar and geomagnetic activities, D region absorption, etc., to better quantify their effects on the signals. We also plan to investigate LEO satellite signals at different frequencies and with different modulations. Finally, we will compare experimental observations with our updated model simulation results to improve our understanding of these effects.

HOW TO CITE THIS ARTICLE:

Morton, Y.J., Bourne, H., Taylor S., & Yang, C. (2026). Ionospheric effects on VHF signals transmitted from a LEO satellite. NAVIGATION, 73. https://doi.org/10.33012/navi.743

ACKNOWLEDGMENTS

This research was funded by Air Force Research Laboratory Space Vehicles Directorate contract # 282109-874X. HAARP is operated by the University of Alaska Fairbanks, Geophysics Institute. The HF heating experiment was conducted as part of the Polar Aeronomy and Radio Science Summer School (PARS) campaign sponsored by the National Science Foundation. We thank Dr. Madeleine Naudeau from the Air Force Research Laboratory for her technical guidance and support and the HAARP staff for making the experiment possible. Travel support for this experiment was funded by Air Force Office of Scientific Research grant #AWD-004534-G2.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

REFERENCES

↵
Benzerrouk, H., Nguyen, Q., Xiaoxing, F., Amrhar, A., Nebylov, A. V., & Landry, R. (2019, May). Alternative PNT based on iridium next LEO satellites Doppler/INS integrated navigation system. In 2019 26th Saint Petersburg International Conference on Integrated Navigation Systems (ICINS), 1–10). IEEE. https://doi.org/10.23919/ICINS.2019.8769440
Google Scholar
↵
Ge, H., Li, B., Jia, S., Nie, L., Wu, T., Yang, Z., Shang, J., Zheng, Y., & Ge, M. (2022). LEO enhanced global navigation satellite system (LeGNSS): Progress, opportunities, and challenges. Geo-spatial Information Science, 25(1), 1–13. https://www.tandfonline.com/doi/full/10.1080/10095020.2021.1978277#abstract
Google Scholar
↵
Kassas, Z. (2020). Navigation from low earth orbit—part 2: Models, implementation, and performance. In Y. J. Morton, F. van Diggelen, J. J. Spilker, & B. Parkinson (Eds.), Position, navigation, and timing technologies in the 21st century, Vol. 2. Wiley-IEEE Press. https://onlinelibrary.wiley.com/doi/abs/10.1002/9781119458555.ch43b
Google Scholar
↵
NOAA. (n.d.). History of NOAA Satellites. NOAA, National Environmental Satellite Data, and Information Service. https://www.nesdis.noaa.gov/our-satellites/related-information/historyof-noaa-satellites
Google Scholar
↵
Morton, Y., Xu, D., & Jiao, Y. (2022). Ionospheric scintillation effects on signals transmitted from LEO satellites. Proc. of the 35th International Technical Meeting of the Satellite Division of the Institute of Navigation (ION GNSS+ 2022), Denver, CO, 2980–2988. https://doi.org/10.33012/2022.18341
Google Scholar
↵
Morton, Y., Yang, Z., Breitsch, B., Borne, H., & Rino, C. (2020). Ionospheric effects, monitoring, and mitigation techniques. In Y. J. Morton, F. van Diggelen, J. J. Spilker, & B. Parkinson (Eds.), Position, navigation, and timing technologies in the 21st century (879–938). Wiley-IEEE Press. http://doi.org/10.1002/9781119458449
Google Scholar
↵
Niu, F., Morton, Y., Taylor, S., Pelgrum, W., & Van Dierendonck, A. J. (2012). Performances of GPS signal observables detrending methods for ionosphere scintillation studies. Proc. of the 25th International Technical Meeting of the Satellite Division of the Institute of Navigation (ION GNSS 2012), Nashville, TN, 3435–3443. https://www.ion.org/publications/abstract.cfm?articleID=10515
Google Scholar
↵
Pelgrum, W., Morton, J., van Graas, F., Gunawardena, S., Bakich, M., Chemey, D., Peng, S., Triplett, J., Vermuru, A., & Vikram, P. (2011). Measurement and analysis of artificially-generated and natural ionosphere scintillations effects on GNSS signals. Proc. of the 2011 International Technical Meeting of the Institute of Navigation, San Diego, CA, 950–958. https://www.ion.org/publications/abstract.cfm?articleID=9541
Google Scholar
↵
Reid, T. G., Walter, T., Enge, P. K., Lawrence, D., Cobb, H. S., Gutt, G., O’Connor, M., & Whelan, D. (2020). Navigation from low Earth orbit: Part 1: Concept, current capability, and future promise. In Y. J. Morton, F. van Diggelen, J. J. Spilker, & B. Parkinson (Eds.), Position, navigation, and timing technologies in the 21st century (1359–1379). Wiley-IEEE Press. http://doi.org/10.1002/9781119458449
Google Scholar
↵
Rino, C., Breitsch, B., Morton, Y., Jiao, Y., Xu, D., & Carrano, C. (2018). A compact multi-frequency GNSS scintillation model. NAVIGATION, 65(4), 563–569. https://doi.org/10.1002/navi.263
Google Scholar
↵
Rino, C. L. (1979). A power law phase screen model for ionospheric scintillation: 2. Strong scatter. Radio Science, 14(6), 1147–1155. https://doi.org/10.1029/RS014i006p01147
Google Scholar
↵
Sigidwiki.com (2025). Automatic Picture Transmission (APT). Sigidwiki.com Signal identification Guide. https://www.sigidwiki.com/wiki/Automatic_Picture_Transmission_(APT)
Google Scholar
↵
Sun, A. K., Morton, Y., & Lee, J. (2024). Ionospheric scintillation effects across multiple carrier frequency bands transmitted from LEO satellites. Proc. of the 2024 International Technical Meeting of the Institute of Navigation, Long Beach, CA, 109–125. https://doi.org/10.33012/2024.19525
Google Scholar
↵
Xu, D., Morton, Y. T. J., Rino, C. L., Carrano, C. S., & Jiao, Y. (2020). A two-parameter multifrequency GPS signal simulator for strong equatorial ionospheric scintillation: Modeling and parameter characterization. NAVIGATION, 67(1), 181–195. https://doi.org/10.1002/navi.350
Google Scholar
↵
Xu, J., Morton, Y., Xu, D., Jiao, Y., & Hinks, J. (2023). Tracking GNSS-like signals transmitted from LEO satellites and propagated through ionospheric plasma irregularities. Proc. of the 36th International Technical Meeting of the Satellite Division of the Institute of Navigation (ION GNSS+ 2023), Denver, CO, 2813–2822. https://doi.org/10.33012/2023.19327
Google Scholar

PDF

Previous Next

Ionospheric Effects on VHF Signals Transmitted from a LEO Satellite

Abstract

1 INTRODUCTION

2 NOAA-19 TRANSMISSION SIGNAL STRUCTURE