Modern HF Signal Detection and Direction Finding

1

Introduction

1.1 Overview

As most readers know, the electromagnetic spectrum is divided into segments based on the radio frequency (RF) of the relevant electromagnetic wave. The high-frequency (HF) segment of this spectrum is specified to be 3 to 30 MHz. Waves at this frequency are refracted as they pass through the Earth’s atmospheric layer known as the ionosphere;¹ as a result, a wave transmitted skyward by a ground-based antenna may not continue in a straight line but may be bent back toward the Earth, where it can be received by an appropriately tuned radio receiver. The details of this phenomenon depend on the specific radio frequency, on the launch angle of the wave, and of course on the ionization conditions. Favorable propagation conditions are common enough to have spawned a wide variety of HF communications-signaling applications, including transcontinental radio operations, commonly known as shortwave (SW) radio, and over-the-horizon radar (OTHR)

But the often-favorable propagation conditions that induced these applications, and their associated long communication ranges, have led to a locally crowded spectrum. The spectrogram of figure 1.1, often called a waterfall plot, illustrates this crowding. It shows the power present in each narrowband frequency cell in a 2 MHz frequency span, during a 10 second interval, as a function of the frequency specified by the ordinate. Many signals continue unabated over this interval; others last only a short time. Noise bursts abound.

Figure 1.1 Spectrogram (waterfall plot) example

Such spectral crowding is common in the HF band. After all, a transmitted HF signal that can reach a distant intended receiver will also illuminate a large area where many unintended receivers are located; these receivers may be communicating with other transmitters, but the undisciplined nature of most HF signaling protocols leaves many possibilities for multiple users to employ a common frequency. Such cochannel interference is therefore a very common occurrence at HF, and as a result, the need for interference mitigation approaches is clear. Adaptive antenna technology, a powerful interference-suppression technology based on multielement antenna arrays at the receiver to exploit spatial differences among cochannel signals, has recently become affordable for such an application. This technology is the primary focus of this work.

Adaptive antenna arrays have been an academic study topic for decades. Elements of such arrays are equipped with hardware that can modify the amplitudes and phases of each element signal, each represented mathematically as a complex signal, by a complex weight prior to their summation in a beamformer. These weights are computed adaptively to instill the beamformer with the desired characteristics. Mathematical analysis and computer simulations have shown such systems to have great power in suppressing interference and in mitigating multipath effects. Yet until recently, implementation costs have denied the benefits of this technology for many applications. Today, modern electronics technology makes such adaptive antenna systems practicable and affordable. As a result, these systems are now appearing in an increasing number of military and commercial settings.

Initially, these applications focused on environments where line-of-sight propagation phenomenology prevailed. Early examples included applications in radar and communications. The successes in these situations invited the exploration of more complex scenarios, especially in the communications field. In particular, HF communications in which the long-range propagation effects and limited spectrum guarantee cochannel interference in a multipath-rich environment became a target for the emerging adaptive antenna technology. As any user of the HF band will attest, this domain cries out for help.

This book addresses the theory and application of modern adaptive antenna arrays as applied to HF signal sensing and reception. Earlier books [1, 2, 3] have covered classical adaptive array theory in some detail and have provided the foundation for many arrays operating today. But a number of recent theoretical developments especially applicable to the multipath-rich HF sensing/reception problem have not been treated in a comprehensive, self-consistent way. Furthermore, although most of these developments have been published in a variety of journals, notational approaches vary widely. Consequently, reconciliation of the recent results and comparisons of them with classical theory becomes tedious at best. This shortfall is the central theme here. Thus, the goals of this book are to:

Briefly describe the HF signal environment to provide background for readers unfamiliar with the challenges facing an HF receiver system.

Define an appropriate mathematical framework for describing a wide range of theoretical adaptive antenna array models.

Present the theoretical foundations for those array-processing algorithms applicable to the HF communications regime.

Provide example algorithm results based on simulations of appropriate HF signal scenarios.

Thus, the audience for this book includes members of both the civil and military communities who are involved with processing HF signals, especially those received in an interference-rich environment. Although the antenna array-based techniques considered here are likely to impose additional costs for the antenna and the multichannel processing, often the interference environment is such that these approaches are the only options with potential for success. One hope is that this work will stimulate the use of advanced adaptive antenna array algorithms in various HF communication applications and will inspire the development of even more sophisticated algorithms for improved performance in this complex, unpredictable, and fascinating propagation environment.

1.2 High-Frequency Propagation

Almost since the beginning of radio, the term HF communications has referred to communications in the frequency band from 2 to 30 MHz. At these frequencies, radio waves transmitted by ground-based antennas can be refracted by various layers of the ionosphere, returning sufficient energy to Earth for reliable ground-to-ground, over-the-horizon signal reception. But ionospheric layers are the result of solar heating, and as a result, the spatial, spectral, and temporal characteristics of the signals reaching a receiving antenna array become intimately tied to the diurnal physics of the ionosphere along the signal path. As a result, any treatise on HF signal processing would be incomplete without at least superficial discussion of ionospheric phenomenology as it effects HF radio signal propagation. Such an introduction is included in chapter 2.

The atmosphere above the Earth consists of a gaseous mixture with an approximately exponential decreasing density, characterized by a decay constant of about 7 km. As the sun illuminates this atmosphere, solar ultraviolet rays ionize air gas molecules in an absorption process. But as the sun’s rays penetrate deeper into the atmosphere, their energy decreases rapidly due to the increase in the ambient air density. Overall, the concentration of free electrons resulting from this ionization is low near the Earth surface, increases with altitude, becomes greatest at about 400 km altitude, and then declines as the atmosphere fades away to the nothingness of free space. This overall picture is illustrated in figure 1.2.

Figure 1.2 Atmospheric ion density vs. altitude

Also, as seen in figure 1.2, this superficially smooth model is actually modulated by irregularities that are often described as layers. A layer exists where the free electron density gradient is sufficient to refract a radio signal back toward the Earth’s surface. Although this refraction is gradual and causes a signal to traverse a curved path, the phenomenon can be viewed simply as a reflection. In fact, the layers are only partially reflective with much of the signal energy continuing through the layer.

Quantitatively, the influence of the ionized medium on upwardly propagating radio signals depends on free electron density, ambient magnetic field, signal frequency, signal polarization, and transmitter-receiver-layer geometry, but as just discussed, the first of these is driven by ultraviolet radiation from the sun and thus depends on time of day, season, altitude, latitude and sunspot number. In addition to reflections, signal energy absorption can occur when the free electron density is high. These various effects make propagation phenomenology quite variable and difficult to describe concisely.

From the standpoint of this book, only a few basic ideas are important:

The ionospheric reflections allow signals to be received well beyond the optical horizon, even at transcontinental distances.

Signal strength at the receive station varies a great deal with time of day, season, and frequency.

A transmitted signal can reach a specific receive station by several different length paths, resulting in severe fading as the received signal components combine with different phases at the receive system antenna.

Such long-distance propagation is responsible for significant levels of cochannel interference that must somehow be mitigated if the intended communication link is to operate successfully.

These factors influence the strategies for reliable communications in the HF band and the signal-processing algorithms that can be used to improve them.

1.3 HF Band Utilization

Because of its potential for long-range communication at relatively low cost, use of the HF band has flourished despite its complex propagation phenomenology. Shortwave broadcast, emergency messaging, ship-to-shore telegraphy, and amateur radio transmissions are some of the most well-known applications. But a relatively unstructured frequency-management approach, arising from the uncertain nature of the propagation medium, the uncertain maximum communication range, and the international regulation environment associated with the long-range nature of HF communications, has led to great variability in frequency occupancy across the band. There are no well-defined frequency slots or standardized signal bandwidths. As a result, transmitting users can never be sure of having a clear broadcast channel, and receivers can never be sure that only one signal will be present at a specific frequency.

Originally, most HF radio traffic could be identified as voice or Morse code (colloquially, CW, or continuous wave). But today, many HF users are transmitting data at a variety of rates, with a variety of signal modulations. A very recent protocol, automatic link establishment (ALE), has developed a significant following. This protocol utilizes a 2.4 kHz frequency-shift keyed (FSK) signal format with an adaptive frequency selection protocol, and has been used as a foundation for a low-cost computer-to-computer networking system. Chapter 3 covers a number of common HF signal formats indicating their instantaneous frequency occupancy.

A set of spectrograms covering the entire HF band, from 2 to 30 MHz, as a function of time of day would show the great variability in the signal environment experienced at HF. The dynamic range is very wide, and weak signals are often overwhelmed by much stronger cochannel signals; such characteristics significantly limit the ability of a single-channel receiver system to receive a weak signal. Multichannel receivers, with their outputs processed by adaptive antenna array algorithms, have the potential for much-improved prosecution of such weak signals. This potential is explored herein.

1.4 Motivations for Adaptive Antenna Array Application at HF

Propagation phenomenology and spectral occupancy associated with HF communications have been outlined in the previous sections. Although the description was superficial, the reader should come away with this message: the HF band is characterized by extreme variability and crowding. If one tunes a narrowband (2.4 kHz bandwidth) receiver to any specific frequency, one is likely to find several signals within its bandwidth. In addition, at least some of the signals in this bandwidth will have originated from the same transmitter but will have been refracted by different ionospheric layers and thus will exhibit different Doppler offsets and time delays. The strength of the various signals will also manifest a large dynamic range, often exceeding 60 dB.

Thus, an HF receiver is faced with a challenging signal environment. Co-channel interference and substantial levels of multipath abound, compromising quality reception. These challenges are the very ones that adaptive antenna array technology has been developed to address. Its success in suppressing interference in the line-of-sight radio communications environment for frequencies from VHF (30 to 300 MHz) through L-band ( ≈1500 MHz) [1, 2, 3], in radar interference suppression at frequencies from UHF (300 to 600 MHz) through C-band (5000 MHz) [4], and in passive sonar where there is little control over the form of the signals to be detected and processed has been documented [5, 6]. It is unnecessary to summarize these successes here. Needless to say, the evidence suggests that adaptive antennas could well play a role in the quality reception of HF communication signals.

However, there is one important issue that is much more prominent in the HF band: signal models used in adaptive antenna array-processing technology always assume plane wave propagation, yet HF skywave signals have been refracted by the ionosphere and reach the receiver system by much more complex propagation phenomenology than in line-of-sight communications. At the very least, this aspect of HF communications signaling injects an iota of uncertainty about the applicability of conventional adaptive antenna array technology in the HF environment. A first step necessarily involves assessment of the line-of-sight, plane wave signal model to HF. Once this basic question has been resolved and the potential for performance enhancement is established, it will be appropriate to quantitatively assess the extent of adaptive antenna performance benefits.

1.5 Overall Plan of the Book

With the stage set for presenting adaptive antenna theory and its application to HF communications in detail, a brief outline of the succeeding chapters will be described. First, a more detailed description of the HF propagation environment will be presented in chapter 2. This chapter will motivate the multipath models used in later discussions of adaptive array algorithms and will provide an overall characterization of the temporal, spectral, and spatial effects of the ionospheric refraction on the transmitted signals as they propagate from their sources to a receiving system equipped with an array antenna.

Next, chapter 3 will provide a description of the more important modulation approaches and signaling protocols used at HF. It will include a discussion of how HF propagation affects a particular modulation scheme and how these effects influence the choice of signal protocol parameters.

After these descriptions of the propagation environment and its component signals, an overview of the important receiver system architectural features needed in an HF receiver system is presented in chapter 4. This chapter will cover coherence and dynamic range requirements for a suitable receiver module to be included in an HF receiving system. Also, digitization of the receiver outputs will be discussed, leading to a specification for an appropriate analog-to-digital converter (ADC). A brief comparison of two commercially-available ADCs will be given to illustrate the evaluation process. No attempt will be made to suggest particular ADCs; such an endeavor will be outdated before this book reaches the reader’s hands.

Chapter 5 will address two important topics, the signal processor architecture and its implementation. Throughout this book, the focus will primarily be on narrowband signals. In actual system practice, a wideband system may be built and operated, but the architecture includes an initial stage of signal processing using digital methods to break the signal seen at each array element into a dense collection of narrowband frequency subchannel signals. Channelization processing is necessary to accomplish this. Then, the narrowband algorithms covered in this text can be applied to individual subchannels separately. Methods to accomplish the creation of multiple adjacent subchannels are covered here, addressing issues such as subchannel leakage and signal fidelity.

Chapter 6 will provide the technical foundation for this book with a description of the vector model for a signal received at an adaptive array antenna. It will start with a summary of the necessary mathematical background, not only to fill gaps in the reader’s knowledge but also to present the notation to be used when specific algorithms are described later. The basics of wave propagation will be reviewed and presented in the coordinate system that will be used throughout the book. A simple array example is included to explain the importance of aperture in resolving individual signals in a multisignal environment. The notation needed to address antenna arrays that manifest polarization diversity will be described. Finally, this chapter will describe the sum and difference processing associated with conventional monopulse antenna systems.

Next, attention shifts to actual signal-processing techniques starting with the detection topic. Chapter 7 addresses the important question: Is there a signal present? At HF this question has a critical extension due to the crowded nature of the HF band: "Is there a new signal present? And to the further, not unrelated question: Has a previously detected signal turned off? And to another very important question: Does a new energy alarm correspond to a signal of interest?" Here again, examples of these algorithms applied to HF channel data are given.

Next, chapters 8 and 9 will consider the direction-finding goal. Starting with the computation of relevant performance bounds in chapter 8, the topic will continue with a myriad of specific direction-finding algorithms in chapter 9. This chapter will include a discussion of array calibration and how it can contribute to achieving direction-finding results that approach the bounds described in chapter 8. Example results, using simulated HF channel data, are included.

Chapter 10 addresses an important topic for consumers of the direction-finding results, geolocation. This term refers to methods for estimating the location of the source from which a signal has been detected and for which a direction has been estimated. Methods include those that utilize lines of bearing from a single sensing system, as well as those that combine such estimates from multiple sensors.

Finally, chapters 11 and 12 will focus on signal estimation, often referred to as copy. This term arises from the common communications operator lingo Do you copy me? In effect, this refers to determining a set of complex antenna element weights that can be applied to the signals observed at the various array elements, thereby producing a single receiver system output that is an estimate of the transmitted signal. Chapter 11 focuses on methods that estimate a signal steering vector and use this estimate to compute copy weights. In contrast, chapter 12 exploits a priori knowledge of a specific signal type to compute appropriate antenna element weights. Least mean square methods, the earliest approach used for signal copy, is one example of this class, but constant amplitude signal copy is a more important algorithm for HF signals where phase-shift keyed signals are frequently transmitted. Example results of these algorithms applied to HF channel data are given with each algorithm; also included is a discussion of weight update approaches that are necessary if the dynamic nature of HF signal propagation is to be addressed effectively.

References

1. R. A. Monzingo and T. W. Miller, Introduction to Adaptive Arrays, New York: John Wiley and Sons, 1980.

2. P. J. D. Gething, Radio Direction Finding and Superresolution, 2nd ed., London: Peter Peregrinus, 1991.

3. J. Li and P. Stoica, Robust Adaptive Beamforming, Hoboken, NJ: John Wiley and Sons, 2006.

4. G. A. Fabrizio, High Frequency Over-the-Horizon Radar, New York: McGraw Hill, 2015.

5. L. J. Ziomek, Fundamentals of Acoustic Field Theory and Space-Time Signal Processing, Boca Raton FL: CRC Press, 1995.

6. X. Lurton, An Introduction to Underwater Acoustics, 2nd ed., Berlin: Springer-Verlag, 2010.

Note

1. From 80 to ∼400 km in altitude.

2

HF Propagation Physics and Its Effect on Signals

2.1 Ionospheric Medium

This section describes the ionospheric medium and the natural influences that affect its characteristics.

2.1.1 Overview

The roughly spherical Earth is surrounded by an atmosphere composed of nitrogen, oxygen, carbon dioxide, and other gases; this atmosphere is densest near the Earth’s surface, but as indicated in figure 2.1, its density decreases approximately exponentially. The lapse constant is about −7.027 km, which is the altitude change resulting in a decrease in density by a factor 1/e. Thus, at an altitude of 50 km, the atmosphere is e(−50/7.027) = 0.0008125 times the density at the surface.

Figure 2.1 Standard atmospheric density vs. altitude compared to exponential fit

This upper atmosphere is bombarded by the sun’s radiation, ionizing the air and generating free electrons and ionized molecules of both positive and negative persuasion. Energy at short ultraviolet (UV) wavelengths is especially effective in its interactions with the atmosphere. Free electrons are released from neutral atoms, creating positive ions, but these electrons can in turn bombard neutral atoms and be captured to form negative ions. As a result, this upper atmosphere is an ionized medium, or plasma, that is known as the ionosphere. An impinging electromagnetic wave may propagate through this plasma, or not, depending on various factors detailed later in this chapter. This section concentrates on ionospheric characteristics; the next section outlines the basic interaction between such a propagating wave and the ionosphere.

Of course, at any specific location in the ionosphere, solar flux is not constant. Even if the sun’s radiation did not vary, which it does, Earth rotation imposes a diurnal cycle, and the Earth’s annual orbit around the sun, combined with the tilt of the Earth axis with respect to the ecliptic plane, superimposes a second cyclic component on the impinging radiation. But there are also variations in the solar emission level due to thermal activity on the sun, including solar flares and coronal mass ejections (CMEs). These solar events also have cyclic features. Solar flares manifest a period synchronous with the eleven-year sunspot cycle; the effects of CMEs show variations that follow the twenty-eight-day rotation period of the solar sphere. In addition to these multiple cyclic variations in solar flux at a specific location, there are variations induced by magnetic storms that have unstructured temporal variations.

But in addition to these interactions of the sun with the Earth’s atmosphere, there is another source of ionization, that produced by meteorites captured by Earth gravity. These penetrate the atmosphere, give up their energy, and ionize gas molecules encountered along their path. This contributor to the ionospheric characteristics is especially important at higher latitudes in the winter when solar insolation is at its ebb.

Thus, ionospheric physics is fairly complex, resulting in a complicated ionospheric geography. The composite effects of the various physical processes produce an ionosphere that varies with altitude, latitude, sunspot number, time-of-day, and season. Fortunately, scientists have observed ionospheric conditions for more than a century and have been able to develop models for those features important to HF communications. These models have been incorporated into readily available and effective computer codes. The most frequently used of these models will be described briefly later in this chapter.

2.1.2 Solar Wind

Radiation and particle flow from the sun make up the so-called solar wind, which interacts with the Earth. Solar wind is a term given to the ionized gas, or plasma, that is continually ejected from the sun as a result of its inherent thermonuclear activity. The gross structure of the solar wind phenomenon, as it reaches Earth, is illustrated in figure 2.2. Because the plasma is fully ionized, it can be regarded as a conducting medium that, as it flows outward from the sun, encounters the Earth’s magnetic field. The plasma cannot easily penetrate the Earth’s magnetic field, which represents an obstruction to the flow. Much as a submerged boulder is an obstruction to water flow in a mountain stream, the Earth’s magnetic field interacts with the plasma, distorting the magnetic field from the simple magnetic dipole structure that would exist if there were no solar wind. The result is a compressed magnetic field extending approximately ten Earth radii on the sunward side and an elongated tail extending approximately hundred Earth radii in the Earth’s shadow. The envelope of the resultant distorted magnetic field is known as the magnetopause. As the solar wind penetrates to lower altitudes inside the magnetopause, these interactions change due to the integrated effects on the radiation spectrum and the atmospheric species mix that remains after interactions at higher altitudes. Further, geometry plays a role because the radiation flux is reduced as the zenith angle increases, as shown in figure 2.3. This latter effect is the same one producing more sunburn in summer when the sun is directly overhead. The consequence is the fairly complex magnetosphere, the asymmetrical region surrounding the Earth extending from about 100 to several thousand kilometers altitude, as figure 2.2 shows.

Figure 2.2 Interaction of the solar wind with Earth and its atmosphere [1]

Figure 2.3 Variation of radiation flux with zenith angle

The level of ionization must exceed ∼ 10⁸ ions/m³ for it to have an appreciable effect on a propagating electromagnetic wave. During the day, short UV radiation is strong enough to generate such ionization levels down to 65 km altitude. But at night, when solar radiation is not providing energy to form new ions, recombination processes dominate, and at the lower altitudes, the atmosphere returns to its neutral state. At higher altitudes, however, some solar UV-induced ions remain uncombined, and, together with the meteorite-induced ions provide sufficient ion densities to leave a sensible ionosphere above 100 km altitude.

2.1.3 Altitude Variations

The detailed interaction of the solar radiation with the atmosphere is complex and beyond the scope of this work. Nevertheless, it can be summarized by noting that the absorption cross section of the atmospheric atoms depends on the species and on radiation wavelength. Thus, the radiative energy spectrum evolves as it penetrates deeper into the atmosphere. As a result, the ion and free-electron densities become strong functions of altitude. Figure 2.4 provides four example electron-density profiles corresponding to noon and midnight for both a low and a high sunspot number. Generally, these profiles manifest a low electron density near the Earth’s surface, and a gradual increase with altitude. Peak density occurs at ∼ 300+ km altitude, at summertime noon when the sunspot number is high.

Figure 2.4 Election density vs. altitude: various times of day and sunspot number [1]

In addition to neutral winds, there are other weather-like dynamic processes driving ionospheric dynamics well into the upper atmosphere, for example, plasma drifts driven by electric fields, which in turn are generated by a variety of neutral wind, magnetic field, and plasma interactions. As altitude increases, the radiative-chemistry and plasma-neutral wind effects dominate over the neutral atmospheric processes of the troposphere and mesosphere (e.g., turbulent mixing). These interactions produce an altitude-dependent, irregular electron-density profile that is often characterized using the term layers, even though the altitude variation is more subtle than that term implies.

Although boundaries between these layers are not distinct, there are five terms used to describe them. The lowest of these, the D layer, extends from 65 to 90 km and primarily induces signal attenuation during the daytime hours. Above it is the E layer, from 90 to 140 km; higher yet are two layers, F1 and F2, with the first representing a daytime layer around 170 km and the second a nighttime layer above 250 km extending to as high as 500 km. As the diurnal cycle transitions from night to day in the morning and then back again in the evening, both F1 and F2 layers can be present. Finally, there is sporadic E, an erratic phenomenon with a fairly narrow altitude extent at about 105 km.

Solar flares are the result of plasma instabilities in the sun’s outer layers (convection zone, photosphere, chromosphere, and corona) driven by a magneto-dynamic process. If a solar flare occurs within view of the Earth, a few minutes later brief bursts of energetic protons reach the Earth. If the flare produces energetic particles (a solar particle event [SPE]) and field lines emanating from the flare connect to Earth, then these particles arrive at Earth on the order of tens of minutes later. Furthermore, if lower-energy coronal mass is ejected during the event (CME), it will reach Earth in a period of one to three days. Energetic particles increase ionization at low altitudes, in particular in the D layer. CMEs reaching the Earth’s magnetosphere drive currents in the ionosphere that cause variations in electron density, predominantly at higher altitudes.

Both SPEs and CMEs contribute to the ionizing processes already occurring in the atmosphere as a result of the more normal short-UV solar radiation. Although these solar flares are short lived, they precipitate a dramatic increase in ionospheric plasma density, especially at D layer altitudes. Effects on E and F layers are less significant because at their altitudes the ion density is low enough to make collisions rare; as a result, there is insignificant nondeviative absorption, the process by which a propagating electromagnetic wave is attenuated without deviation from a straight-line path. More discussion of this phenomenon is covered in the next section. Since solar flares are more common when the sunspot number is high, these sudden increases in ionization, though random, follow the eleven-year sunspot cycle in their overall frequency of occurrence.

2.1.4 Geographic Variation

Another unpredictable ionospheric feature is identified by the term traveling ionospheric disturbances, or TIDs. TIDs are a common ionospheric feature characterized by wavelike disturbances in the electron-density profile. They are driven by both gravity waves propagating up from the neutral atmosphere and electrodynamic effects resulting from geomagnetic disturbances.

These general characteristics of the ionosphere are modified in ways that depend on geographic latitude. Four specific latitude bands manifest identifiable characteristics: equatorial, midlatitude, auroral, and polar. A brief summary of the principal differences among them follows.

Equatorial ionosphere. As will become apparent when the auroral and polar ionospheres are discussed, the equatorial ionosphere manifests greater plasma density irregularities than the midlatitude ionosphere. The underlying causes for these features lie with the orientation of the Earth’s magnetic field, often modeled as a magnetic dipole tilted approximately 15° to the Earth’s rotation axis. At equatorial latitudes, the magnetic field lines are almost parallel to the Earth’s surface. So-called equatorial spread F (ESF), also known as equatorial clutter, involves low-electron-density globules that rise from the bottom to the top of the F layer shortly after sunset when the solar flux abruptly ends. These bubbles can be as small as a few centimeters in diameter or as large as 1000 km across. Because the loss of sunlight is their root cause, these bubbles move away from the sun, that is, in an easterly direction.

Midlatitude ionosphere. Many factors already mentioned are important contributors to those ionospheric features affecting HF signal propagation: the day/night terminator, spread F, solar flares, magnetic storms, TIDs, and sporadic E. At mid-latitudes, sporadic E can persist for many hours. Ionic motion induced by neutral air molecule wind shears in the lower atmosphere, interacting with the Earth’s magnetic field, causes the ions to form thin layers, already identified as sporadic E, at 105 km altitude. The density of such layers can exceed that of the F layer above.

Magnetic storms