1. Introduction

This chapter covers studies of radar capabilities and limitations as they may be related to the apparent manifestation of unidentified flying objects. The studies were carried out by the Stanford Research Institute pursuant to a contract with University of Colorado (Order No. 73403) dated 23 June 1967, under sub-contract to the U.S. Air Force.

The preceding chapter of this report, entitled "Optical Mirage - A Survey of the Literature," by William Viezee, covers optical phenomena due to atmospheric light refraction.

As they became available other information and interim results of these studies were informally communicated to the University of Colorado study project in accordance with the referenced contract.

The purpose of this chapter is to provide a basic understanding of radar, the types of targets it can detect under various conditions, and a basis upon which specific radar reports may be studied. Studies of specific UFO incidents were performed by the Colorado project (see Section III, Chapter 5).

At first consideration, radar might appear to offer a positive, non-subjective method of observing UFOs. Radar seems to reduce data to ranges, altitudes, velocities, and such characteristics as radar reflectivity. On closer examination however, the radar method of looking at an object, although mechanically and electronically precise, is in many aspects substantially less comprehensive than the visual approach. In addition, the very techniques that provide the objective measurements are themselves susceptible to errors and anomalies that can be very misleading.

In this chapter we will consider how the radar principle applies to detection of targets that may be or appear to be UFOs, and attempt to establish the criteria by which such apparent manifestations must be judged in order to identify them. Since we make no assumptions regarding the nature of UFOs we limit ourselves to describing the principles by which radars detect targets and the ways in which targets appear when detected. In a word, we can only specify the nature of radar detection of targets in terms of physical principles, both in regard to real and actual targets and in regard to mechanisms which give rise to the apparent manifestation of targets. It is hoped that these specifications will assist in the review of specific instances as they arise. Even in cases where radar may identify target properties that cannot be explained within the accepted frame of understanding of our physical world, the authentic observation of a target having such properties will shed little or no light on its nature beyond the characteristics observed, and it will therefore remain unidentified.

2. Radar Systems

RADAR is an acronym for RAdio Detection And Ranging. It is a device for detecting certain types of targets and determining the range to the target. The majority of radars are also capable of

measuring the azimuth and elevation angles of targets.

Radars operate on three fundamental principles:

1. that radio energy is propagated at uniform and known velocity;
2. that radio energy is normally propagated in nearly straight lines, the direction of which can be controlled or recognized; and
3. that radio energy may be reradiated or "reflected" by matter intercepting the transmitted energy.

Basically radar consists of a transmitter that radiates pulses of electromagnetic energy through a steerable antenna, a receiver that detects and amplifies returned signals, and some type of display that presents information on received signals.

Radar systems can be separated into three general categories:

1. operational systems,
2. special usage systems and
3. experimental and research systems. These include fixed and portable ground-mounted systems, airborne, and shipborne systems.

Many types of radars are specifically designed to perform specialized functions. In general, radars provide either a tracking or a surveillance function. The surveillance radar may scan a limited sector or 360° and display the range and azimuth of all targets on a PPI (plan position indicator). Tracking radar locks onto the target of interest and continually tracks it, providing target coordinates including range, velocity, altitude, and other data. The data are usually in the form of punched or magnetic tape with digital display readout. Air traffic control, ship navigation, and weather radars fall into the surveillance category; whereas instrumentation, aircraft automatic landing, missile guidance, and fire control radars are usually tracking radars. Some of the newer generation of radar systems can provide both functions, but at this time these are very specialized systems of limited number and will not be discussed further.

In addition to the above general applications, each of the radar systems have special selective functions for various purposes. For example, some radar systems are designed so that they can track moving targets. Signals from stationary targets such as the ground, buildings, or even slow-moving objects are excluded from the display. This simplifies the display and makes it possible to track aircraft even though they are moving through an area from which strong ground clutter signals would otherwise mask the echo from the aircraft.

In addition to the many radar types, the radar operator has at his disposal many control functions enabling system parameters to be changed in order to improve the radar performance for increasing the detectability of particular types of targets, thereby minimizing interference, weather, and/or clutter effects. These radar system controls can modify any one or any combination of the following characteristics:

• Transmitter output power
• Pulse repetition rate
• Sensitivity time control
• Transmitted pulse width
• AGC response time
• IF receiver bandwidth
• Transmitter operating frequency
• Antenna scan rate
• Polarization control of radiated and received energy
• Skin or transponder beacon tracking
• Receiver RF and IF gain
• Display control functions
• Numerous signal processing techniques for clutter suppression, weather effects, moving target indication, false alarm rate, and threshold controls.

The radar operator himself is an important part of radar systems. He must be well trained and familiar with all of the interacting factors affecting the operation and performance of his equipment. When an experienced operator is moved to a new location, an

important part of his retraining is learning pertinent factors related to expected anomalies due to local geographical and meteorological factors.

Two other groups of persons also affect the performance of the radar system. They are the radar design engineer and the radar maintenance personnel. The designer seeks to engineer a radar which achieves the performance desired, in addition to being a system which is both reliable and maintainable. Highly trained maintenance technicians routinely monitor the system insuring that it is functioning properly and is not being degraded by component system failures or being affected by other electronic systems that could cause electrical interference or system failure.

During the past 30 years, radar systems design has considerably improved. Radars manufactured today are more complex, versatile, sensitive, accurate, more powerful, and provide more data-processing aids to the operator at the display console. They are also more reliable and easier to maintain. In the process, they have become more sensitive to clutter, interference, propagation anomalies, and require better trained operating and maintenance personnel. Furthermore, with the increased data-processing aids to the operator, the more difficult becomes his target interpretation problem when the radar systems components begin gradually to degrade or when the propagation environment varies far from average conditions. The more sophisticated radar systems become, the more sensitive the system is to human, component, and environmental degradations.

3. Radar Fundamentals

Radar detection of targets is based on the fact that radio energy is reflected or reradiated back to the radar by various mechanisms. By transmitting pulses of energy and then 'listening' for a reflected return signal, the target is located. The period of time the radar

is transmitting one pulse is called the pulse length and is generally measured in microseconds (millionths of a second) or expressed in terms of the length from the front to the back edge of the pulse. (A one microsecond pulse is 984 ft. long, since radio waves, like light travel 186,000 statute mps.) The rate at which pulses are transmitted is called the pulse repetition rate. When pulses are transmitted at a high rate, the receiver listening time between pulses for return echoes is reduced as well as the corresponding distance to which the energy can travel and return. This means that the maximum unambiguous range is decreased with increasing pulse repetition rate. More distant targets may still return an echo to the radar after the next pulse has been transmitted but they are displayed by the radar as being from the most recent pulse. These so-called multiple trip echoes may be misleading, since they are displayed at much shorter ranges than their actual position.

Other important operating characteristics of a radar are its transmitted power and wavelength (or frequency). The strength of an echo from a target varies directly with the transmitted power. The wavelength is important in the detection of certain types of targets such as those composed of many small particles. When the particles are small relative to the wavelength, their detectability is greatly reduced. Thus drizzle is detectable by short wavelength (0.86 cm.) radars but is not generally detectable by longer (23 cm.) wavelength radars.

The outgoing radar energy is concentrated into a beam by the antenna. This radiation of the signal in a specific direction makes it possible to determine the coordinates of the target from knowledge of the azimuth and elevation angle of the antenna. The desired antenna pattern varies with the specific purpose for which the radar was designed. Search radars may have broad vertical beams and narrow horizontal beams so that the azimuth of targets can be accurately determined. Height finders on the other hand have broad horizontal beams so that the height of targets can be accurately determined. In either case the radiating and receiving surface of the antenna is usually a section of a paraboloid.

A circular beam may be described as a cone with maximum radiation along its axis and tapering off with angular distance from the center. The beam is described by the angle between the half power points (the angular distance at which the radiated power is half that along the axis of the beam). In the case of non-circular beams two angles are used, one to describe the horizontal beamwidth, a second to describe the vertical beamwidth. Later in this report the detection of targets by stray energy outside the main beam will be discussed.

The size of the beam for a given wavelength depends on the size of the parabola. For a given size parabola the longer the wavelength, the broader the beam.

When the radiated energy illuminates an object, the energy (except for a small amount that is absorbed as heat) is reradiated in all directions. The amount that is radiated directly back to the radar depends on the radar cross-section of the target. Differences between geometrical cross-section and radar cross-section are related to the material of which the object is composed, its shape, and also to the wavelength of the incident radiation. The radar cross-section of a target is customarily defined as the cross-sectional area of a perfectly conducting sphere that would return the same amount of energy to the radar as that returned by the actual target. The radar cross-section of complicated targets such as aircraft depends on the object's orientation with respect to the radar. A jet aircraft has a much smaller radar (and geometric) cross-section when viewed from the nose or the tail than when viewed broadside.

Equations relating the various parameters are given, in varying degrees of complexity, in textbooks on radar. In their simplest form the equations for average received power are:

For point targets (birds, insects, aircraft, balloons, etc.):

For plane targets (earth's surface at small depression angles):

     (2)

For volume targets (precipitation):

     (3)

Where:

• P_r = average received power
• P_t = transmitted power
• G = Antenna gain
• lambda = wavelength
• sigma = radar cross-section
• R = range of target
• theta = horizontal beamwidth of antenna
• c = velocity of radio waves
• tau = length of transmitted pulse
• psi = vertical beamwidth of antenna
• eta = reflectivity per unit volume

These equations show that the intensity of echo signal varies according to whether the target is a point, a relatively small area, or a very large volume such as an extensive region of precipitation. The echo signal intensity of point targets varies inversely with the fourth power of the distance from the radar to the targets. The, intensity of area targets varies with the cube of the distance, and that of large volume targets, with the square of the distance.

Figure 1 illustrates how the radar beamwidth and the cross section area or volume of the target interact to give these different

variations with range of the returned signal. In Fig Aa, the point target has a radar cross-section [sigma]. In Fig. Ab there may be a number of targets with radar cross-section sigma over an area with dimensions of half the pulse length and the beam width at range R. Replacing sigma in equation (1) with this new expression for radar cross-section cancels one R in the denominator giving the R³ relationship. When the target is many sigma's spread over a volume with dimensions determined by range, horizontal and vertical beamwidth, and half the pulse length (Fig. Ac) R appears in the numerator twice, thus cancelling an R² in the denominator of equation (1).

Because of differences in variation with distance of the return signal from various types of targets it is apparent that with combinations of targets the point targets might not be detectable. For example, an aircraft cannot be detected when it is flying through precipitation or in an area of ground targets unless special techniques are used to reduce the echo from precipitation or ground clutter.

Information on signals returned to the radar by a target may be presented to an operator in a number of ways; by lights or sounds that indicate there is a target at a selected location; by numbers that give the azimuth, elevation angle, and range of a selected target; or in 'picture' form showing all targets within range that are detected as the antenna rotates. The latter form of presentation is called a Plan Position Indicator (PPI). Plate 65 shows a photograph of a PPI. This photograph is a time exposure equal to the time for one antenna revolution. The center of the photograph is the location of the radar station. Concentric circles around the center indicate distance from the station. In this case the range circles are at 10 mi. intervals, so the total displayed range is 150 mi. North is at the top of the photograph and lines radiating from the center are at 10° intervals.

Plate 65: PPI (Radar Screen)

Click on Thumbnail to see Full-size image.

A PPI display such as this corresponds very closely to a map. Often overlays with locations of cities, state boundaries, or other pertinent coordinates are superimposed over the PPI to aid

in locating echoes. The plate shows a number of white dots or areas at various locations. These may be echoes from a variety of different targets, or they may be the result of interference or system malfunction.

The radar operator must keep watch of this entire area (70,650 sq. mi. in this example) and try to determine the nature of the targets. If he is a meteorologist he watches for and tracks weather phenomena and ignores echoes which are obviously not weather-related. If he is an air traffic controller he concentrates on those echoes that are from aircraft for which he is responsible. Many unexplained radar echoes are not studied or reported for several reasons. One of the reasons might be that the operators in general only track targets that they can positively identify and control. Since a radar operator can only handle a limited number (6 to 8) of targets simultaneously, he might not take serious note of any strange targets unless they appear to interfere with the normal traffic he is vectoring. Even when the unexplained extraordinary targets are displayed, he has little time available to track and analyze these targets. His time is fully occupied observing the known targets for which he is responsible. In addition, the operator is familiar with locally recurring strange phenomena due to propagation conditions and suspects the meteorological environment as being the cause. In general, the operator seldom has a way in which to record the displayed data for later study and analysis by specialists.

In addition to the tracking of various targets he must also be aware of the possibility of malfunction of the radar.

4. System Reliability

Two types of failures occur in a radar system: those that are catastrophic and those that cause a gradual degradation. In spite of good maintenance procedures, there will be system component failures that occur due to external events such as ice or wind loading, rain

on the cabling and connectors, bugs and birds in the feed structure. The operator is not always immediately aware of such failures. He is usually located in a soundproofed and windowless room remote from the transmitter, antenna, and receiving hardware. The operator has available to him only the console display and readout equipment. Catastrophic systems failure is usually self-evident to the operator. When the transmitter power tube fails, or the antenna drive unit fails, the operator is aware of this immediately on his PPI display. But when the gain in a receiving tube decreases, or the system noise slowly increases due to a component degradation, or the AFC in the transmitter section begins to go out of tolerance over a period of days causing increased frequency modulation or "pulse jitter" in the transmitted pulse, time may elapse before the operator becomes aware of the slowly deteriorating performance. Reduced sensitivity or the increased reception of extraneous targets from ground clutter or nearby reflecting structure is often evidence that the radar system is deteriorating.

It can be considered that a major system component of a typical radar might be subject to catastropic failure every 250 to 2,000 hours of operation (5 to 36 average failure-free days) and that graceful degradations of components occur continually. Possible failure thus becomes one of the first causes to be considered in analyzing unusual radar sightings. The next factor will be possible unusual propagation effects to which the radar is subject. Analysis of extraordinary sightings is further handicapped by the fact that the displayed data of the sighting usually are not recorded and that any explanations must frequently be based upon interpretations by the operators present at the time of the sighting. The point is that the operator, the radar, and the propagation medium are all fallible parts of the system.

5. Relationships between Echoes and Targets

There are five possible relationships between radar echoes and targets. These are:

1. no echo - no target;
2. no echo - when a visual object appears to be in a position to be detected;
3. echo - unrelated to a target;
4. echo - from a target in a position other than that indicated;
5. echo - from a target at the indicated location.

The first and last possibilities are indicative of normal function. Possibility (b.) becomes of importance where there is an object that is seen visually. Then, from knowledge of the types of targets that are detectable by the radar, some knowledge of the characteristics of the visual object could be obtained.

The situations (c.) where there is an apparent echo but no target are those when the manifestation on the PPI is due to a signal that is not a reradiated portion of the transmitted pulse but is due to another source. These are discussed in a subsequent section of this chapter.

Situations where the echo is from a target not at the indicated location (d.) may arise due to one or a combination of the following reasons. First, abnormal bending of the radar beam may take place due to atmospheric conditions. Second, a detectable target may be present beyond the designed range of the radar and be presented on the display as if it were within the designed range, for example, multiple-trip echos from artificial satellites with large radar cross-sections. Third, stray energy from the antenna may be reflected from an obstacle to a target in a direction quite different from that in which the antenna is pointed. Since the echo is presented on the display along the azimuth toward which the antenna is pointed the displayed position will be incorrect. Finally, targets could be detected

by radiation in side lobes and would be presented on the display as if they were detected by the main beam.

Possibility (e) listed above encompasses the broad range of situations where there is a target at the location indicated on the display system. Of primary concern in this case is the identification of the target.

The possible relationships listed above show that radarscope interpretation is not simple. To attempt to identify targets, the operator must know the characteristics of his radar; whether it is operating properly; and the type of targets it is capable of detecting. He must be very aware of the conditions or events by which echoes will be presented on the radar in a position that is different from the true target location (or in the case of interference by no target). Finally, the operator must acquire collateral information (weather data, transponder, voice communication, visual observations or handover information from another radar before he can be absolutely sure he has identified an unusual echo.

6. Signal Sources

Sources of electromagnetic radiation that may cause real or apparent echoes on the radar display include both radiators and reradiators. Some sources, such as ionospheric electron backscatter, the sun, and the planets, are not considered, since they can be detected only by the most sensitive of research radars. As a radiator the sun does emit enough energy at microwave wavelengths to produce a noise signal. This signal has been used for research purposes (Walker 1962) to check the alignment of the radar antenna. Radio sextants have been built which track the sun at cm. wavelengths by Collins Radio Co. Since this signal is quite weak it is unlikely it would be noticed during routine operation of a search radar.

Reradiators include objects or atmospheric conditions that intercept and reradiate energy transmitted by the radar. Objects range in size from the side of a mountain to insects. Atmospheric conditions include ionized regions such as those caused by lightning discharges

and inhomogeneities in refractive index caused by sharp discontinuities in temperature and moisture.

Table 1 lists some radiators and reradiators. This list is incomplete since continuing development of new types of radars or improvements due to evolutionary growth of existing radars results in new types of targets becoming detectable.

The signal sources listed have relatively unique sets of characteristics although in many cases there is some overlap. For example, a fast flying bird with a tailwind could have ground speeds comparable to a light aircraft with a headwind. At comparable range, however, the signal intensity would be quite different unless the bird were in the main beam and the aircraft in a side lobe. This section will discuss the typical characteristics and behavior of the return signals and the auxiliary information needed to confirm or reject them as the sources of a given echo will be mentioned. For example, as mentioned above, knowledge of wind speed is necessary to determine the air speed of a target.

In the discussion of detectability of the various signal sources some specific frequency bands may be mentioned. Figure 2 illustrates the relationships between wavelengths and frequency in the various bands and shows specific radar bands within the frequency and wavelength spectrum.

Precipitation

In the 1940's when radar technology advanced to the point where wavelengths less than half a meter began to be feasible, precipitation became a radar-detectable target. Ligda (1961) states that the first radar storm observation was made on 20 February 1941 in England with a 10 cm. (S band) wavelength radar. Since that time, radar has been widely used for meteorological purposes and special meteorological radars have been designed and constructed specifically for precipitation studies (Williams, 1952; Rockney, 1958). Many radars designed for purposes other than weather detection were found to be very adequate as precipitation detectors. Ligda (1957) studied the distribution of precipitation over large areas of the United States using PPI photographs from Air Defense Command (ADC) Radars during the period 1954 to 1958 and during 1959 studied the distribution of maritime precipitation shown by PPI photographs from radars aboard ships of Radar Picket Squadron I stationed off the west coast of the United States. Later programs concurrent with several of the meteorological satellites (Nagle, 1963; Blackmer 1968) have also utilized data from ADC and Navy radars. Thus radars designed for other specific missions are often capable of detecting precipitation and an understanding of the characteristic behavior and appearance of precipitation is essential if the radar operator is to interpret properly the targets his radar detects.

Detailed studies have been made of characteristics of radar returns from precipitation. In a review of the microwave properties of precipitation particles Gunn and East (1954) discuss variations in return signal with wavelength and differences between the return signal from liquid and frozen water particles. Precipitation consists of a large volume of particles that generally fill the beam at moderate ranges. The

received power at any instant is the resultant of the signals from the large number of individual particles. The particles are constantly changing position relative to each other (and to the radar site). As a result the signals from the individual particles sometimes add to give a strong return, sometimes subtract to give a weaker signal. This fluctuation in echo from precipitation is readily apparent on scopes that permit examination of the return from individual transmitted pulses. The fluctuation of the return signal is not, however, apparent to a radar operator monitoring the PPI of a search radar. This is because the persistence of the cathode ray tube used for PPI displays averages or integrates a number of pulses. Of importance to a radar operator concerned with interpreting the PPI is the variation of signal intensity with wavelength, with pulse length and with precipitation type. Particles that are large compared to the wavelength are more readily detectable than those that are small compared to the wavelength. Light drizzle may be barely detectable at short ranges while severe thunderstorms with large raindrops are detectable at ranges of 300 - 400 mi. When there is large hail falling from a severe thunderstorm the return signal may be quite strong.

Radar-detected precipitation may be in a variety of forms from very widespread continuous areas of stratiform precipitation of sufficient vertical extent to nearly cover the PPI of a long-range (150 n.mi.) search radar to only one or two isolated small sharp edged convective showers. The former is likely to persist for many hours, the latter for only a fraction of an hour. Between these two extremes there are many complex mixtures of convective and stratiform precipitation areas of various sizes. One of the distinguishing features of precipitation echoes is their vertical extent and maximum altitude. Usually precipitation echoes extend from the surface to altitudes up to 60,000 ft., although a more common altitude of tops is 20,000 - 40,000 ft. Further, isolated small volumes of precipitation seldom remain suspended in the atmosphere. The initial echoes from showers and thunderstorms may appear as small targets at moderate altitudes but subsequently grow

rapidly. For example, Hilst and MacDowell (1950) examined the initial echoes from a thunderstorm. Horizontal measurements were made with a 10 cm. radar and the vertical measurements were made with a 3 cm. radar. Their first measurement showed a small horizontal area and a vertical extent from 11,000 - 18,000 ft. Presumably measurements a short time earlier would have shown smaller dimensions. Subsequently there was rapid growth to an area of 200 sq. mi. and a vertical extent from the surface to about 30,000 ft. The importance of this large vertical extent is that such an echo on the PPI of a search radar with a narrow beam can be present at a variety of ranges; that is, the beam will not be below the target at short ranges or above it at long ranges as would be the case with targets of limited vertical extent.

Since precipitation is less detectable at longer wavelengths and showers may have a quite short lifetime, it is possible that on rare occasions precipitation targets could confuse the radar operator. Consider for example a search radar operating at wavelengths of greater than 20 cm. in an environment where short-lived showers were occurring. A study by Blackmer (1955) using photographs from a 10 cm. radar showed a peak in echo lifetimes of 25 - 30 min. while the mean lifetime was 42 min. Also using data from an S band radar, Battan (1953) found a mean echo duration of 23 min. with the greatest number having lifetimes of 20.0 - 24.9 min. At longer wavelengths with short lifetimes, it is not impossible that an intense shower would be detectable only in the brief period during which it was producing hail, because a long wavelength radar might not detect small precipitation particles but could detect hail. Water-coated hail acts as a large water sphere and thus gives very strong return signals even at long wavelengths. Geotis (1963) found that hail echoes are very intense subcells on the order of 100 M. in size. When a number of short-lived showers or long-lived showers that were detectable only when hail is falling, are within range of a long wavelength radar, the PPI display could show over a period of time, a brief echo at one location, then an echo at a new location for a

short period, etc. This might be interpreted as a single echo that was nearly stationary for a short period then moving abruptly to a new position.

One of the characteristics of precipitation echoes is that their motion is very close to that of the wind direction and speed. This wind velocity may not be the same as that observed at the radar site if the distance to the precipitation is great. Occasions have also been noted when precipitation echoes within a relatively small area have shown differences in motion due to being moved by different wind directions at various levels.

In general, however, precipitation is a relatively well behaved radar target and except for rare instances its extensiveness and orderly movement readily identifies it to the radar operator monitoring a PPI display.

Aircraft

The term aircraft includes a wide variety of vehicles from unpowered sailplanes to the most advanced military jets with speeds several times that of sound. A target such as an aircraft has a very complex shape that is many times the wavelength of the incident radar energy. As the energy scattered from different parts of the aircraft adds or subtracts from other parts, the signal returned to the radar fluctuates. Fluctuations in the echo can also result from changes in the angle at which the aircraft is viewed. That is, when an aircraft is viewed broadside, its radar (and visual) cross-section is much larger than when viewed from the nose or tail. Skolnik (1962) reports a 15 dB change in echo intensity with an aspect change of only 1/3 of a degree. High frequency fluctuations due to jet turbines (Edrington, 1965) and propellors (Skolnik, 1962) have also been reported. These fluctuations are on the order of 1000 cycles per second and would not be apparent on a PPI.

Although aircraft echoes fluctuate due to aspect and propulsion modulations, there is a general correlation between size of aircraft and the amount of signal returned to the radar. An indication of the

relative detectability of several aircraft as given by the Air Force (1954) is F-86 = 0.46, B-45 = 0.75, B-17 = 1.0, B-29 = 1.2. The numbers mean that, if on a given radar a B-17 was just detectable at 100 mi., an F-86 would be just detectable at 46 mi.

The radar cross-sections of components of a large jet aircraft was measured with a 71 cm. radar (Skolnik 1962) and maximum values in excess of 100 m² were found. The fuselage of the large jet when viewed from the front or rear had a cross-section of about one-half square meter. Smaller aircraft would have much smaller radar cross-section of about one-half square meter. Smaller aircraft would have much smaller radar cross-sections and light aircraft or sailplanes of fiberglass or wooden construction could have extremely small radar cross-sections.

Another type of fluctuation in echo signal from aircraft and similar point targets is due to the nature of radio wave propagation. When a radar wave is propagated over a plane reflecting surface there will be reflections from that surface to a target in addition to the direct path from the radar to the target. Figure 3 illustrates the geometry of beam distortion due to such a plane reflecting surface. In Fig.3a an idealized beam pattern in free space is shown. When a reflecting surface such as the ground or sea surface is introduced a portion of the beam will be reflected from the surface as in Fig.3b. A target will thus be illuminated both by a direct wave and a reflected wave. The echo signal from the target back to the radar travels over the two paths so that the echo is composed of two components. The resulting echo intensity will depend on the extent to which the two components are in phase. Areas along which the two components are in phase resulting in a stronger signal lie along lines of angular elevation of

The two components are out of phase and nearly cancel each other between the maxima. The resulting beam pattern thus consists of a series of lobes as presented schematically in Fig. 3c. As an aircraft flies along it will progress through the regions of maxima and minima, and the signal will fluctuate from near zero in the minima to a value near twice the free-space intensity in the maxima.

The foregoing assumes a plane, perfectly reflecting surface. Since the surface in the vicinity of a radar station is generally not a plane and its reflecting qualities vary the situation is much more complex than the idealized case.

The effect of these fade areas is to cause aircraft targets to sometimes disappear and then (if the target has not reached a range such that the return signal is no longer detectable) to reappear. With a number of aircraft flying about it is not inconceivable that the fadings and reappearances of the several aircraft would be difficult to keep track of and could be misinterpreted as a smaller number of targets that were moving quite erratically.

Considering the whole spectrum of vehicles that travel in the atmosphere, there may be speeds as low as zero (hovering helicopter) or speeds exceeding Mach 3.0. Correspondingly, altitudes vary from the surface to 50,000 - 60,000 ft. (in some cases above 100,000 ft.) Different types of aircraft, however, are limited in their range of speeds and altitudes. A hovering helicopter cannot suddenly accelerate to three times the speed of sound. Neither can a supersonic jet hover at 60,000 ft. A characteristic of an aircraft echo on a PPI is therefore its relative uniformity of movement. To monitor this movement allowance must be made for fades. The direction of movement also will be quite independent of wind direction at flight level.

Birds and insects

Possibly the earliest observation of a radar echo from a bird was made by R. M. Page (1939) of the Naval Research Laboratory in February, 1939. It was made with an experimental 200 MHz. radar (the XAF) on the U.S.S. New York near Puerto Rico. Bird echoes, as reported by Lack and Varley (1945), were observed on a 10 cm. coast-watching radar set near Dover during 1941. Visual checks confirmed both of these early detections by radar as being returns of individual birds. Numerous bird observations by radar have been made since,

especially of bird migrations as is evidenced in a bibliography compiled by Myres (1964) listing 89 papers, and a text written by Eastwood (1967). Radar cross-sections (sigma) have been measured of birds in a fixed position suspended in a non-reflecting sling and of birds in flight. The values obtained, shown in Tables 2 and 3, vary with species, aspect, and radar wavelength.

Because of the inverse-fourth-power variation with range, a bird at short range in the main beam can give a radar echo comparable in intensity to that from an aircraft in the main beam at a long range. For example, if a pigeon with a broadside radar cross-section of 100 cm² were flying within the radar main beam at a range of 10 mi., it would produce as strong a signal to the radar as a jet aircraft with a a value of 10⁶ cm² (100 m²) flying within the radar main beam at a range of 100 mi. However, if the aircraft were flying in a side-lobe 40 dB less powerful than the main beam in which the bird is flying both would produce equal intensity signals at the same range. If the side lobe were 30 dB down, a bird in the main beam at 10 mi. would look like an aircraft at 17.8 mi., and if the side lobe were 20 dB down, the bird at 10 mi. would look like an aircraft at 31.6 mi.

Theoretically the maximum detectable range as dictated by the amount of radar signal returned from birds can be calculated. However, verification is not easy due to the difficulty of spotting a bird and establishing that it belongs to a particular blip on a radar scope. This is particularly difficult in the presence of sea clutter as experienced during an experiment conducted by Allen and Ligda (1966) at Stanford Research Institute. During an experiment conducted by Konrad (1968), individual birds were released from an aircraft flying over water at 5,500 - 6,000 ft. from 8 - 10 n.mi. from the radars. After separation of the aircraft from the bird in the radar scope, each individual bird was automatically tracked for periods up to five minutes, so that the target observed was positively identified as a bird. Flocks of birds have been detected to ranges of at least 51 n.mi. as reported by Eastwood and Rider (1965).

*VV, Transmit vertical polarization and receive vertical polarization.

†VH, Transmit vertical polarization and receive cross-polarized or horizontal component.

*For the cross-section measurements of the starling, pigeon, sparrow, and rook, the birds were suspended from a tower with their wings folded; the radar elevation angle was 18°. Measurements of the turkey buzzard were made when the bird was in flight; measurements of the duck and chicken were made when the birds were standing or squatting.

†400 megacycles.

Very few birds fly over 13,000 ft.; most fly below 5,000 ft. In a survey conducted by Farrari (1966) of USAF reports of bird-aircraft collisions during 1965, 27% of all collisions were under 100 ft. 28% between 100 - 2,000 ft., 21% between 2,000 - 3,000 ft. and the 24% above 3,000 ft. If it can be assumed that the probability of a bird-aircraft collision is equally likely at all altitudes (which may not be fully valid due to climb and descent) this should be somewhat of a representative figure of the height of flight for birds. There was one reported bird-aircraft strike at 17,000 ft. and a few sightings above 20,000 ft., however the number of birds flying at these altitudes appears to be extremely small.

Eastwood and Rider (1965) reported a rather complete analysis of the height of flight of various birds observed by radar at the Marconi Research Laboratory in England. Their findings agreed very closely with the above; about 90% of all birds were below 5,000 ft. Birds fly higher at night and during the spring and fall migration periods. A plot of the average altitude distribution over the year is shown in Fig. 4. All of these figures are probably applicable as height above the general terrain; i.e., at 5,000 ft. above mean sea level, 90% of the birds would fly at altitudes below 10,000 ft.m.s.l. The amount of cloud cover also affects the height at which birds fly. Diagrams included by Eastwood and Rider (1965) clearly indicate a marked tendency for higher mean altitudes to be flown in the presence of complete cloud cover.

Target airspeed is another means for identifying a bird. It can be obtained vectorially from a knowledge of the wind velocity and the radar-measured target velocity. Houghton (1964) determined the airspeed of a limited sampling of the birds by visually identifying each through a telescope aimed by tracking radar Fig. 5. In all cases the wind speeds were less than 5 knots. Target air speed cannot invariably distinguish between a helicopter, a slow moving aircraft and a bird flying in a high wind without precise knowledge of the wind at the bird altitude.

Flocks of birds sometimes produce rings on a radar scope which expand from a number of fixed points. These have been called "ring angels" and were first attributed to birds by Ligda (1958). Visual confirming observations were lacking at that time. Later, Eastwood, Isted and Rider (1962) verified that radar ring angels were definitely caused by the dispersal of starlings (Sturnus vulgaris) from their roosts at sunrise. After several radar scope observations were studied, it became possible to pinpoint the centers of the rings and the approximate locations of the roosts. A number of observers equipped with radio telephones were stationed at each location and signaled the precise moment of emergence of the successive flocks of starlings from the roost under observations. These data were correlated with the radar scope presentations to confirm definitely the generation of ring angels by birds. The mean air speed of starlings leaving the roost was measured as 37 knots.

Under some conditions, slow-moving ring echoes may be produced by the rise of a temperature inversion layer in the early morning hours after sunrise. Sea-breeze fronts have occasionally been seen on radar as a line, and at other times as a boundary between scattered and concentrated signal returns as shown by Eastwood (1967). How much of the line produced is due to the meteorological effects and how much by birds and insects is still a matter for speculation. However, Eastwood (1967) cites reports by glider pilots sharing upcurrents with birds taking advantages of the lift provided. This and some limited study of the characteristics of the radar scope signals, produce some indication as to the validity of the bird theory.

Some studies have been made on target signal fluctuation and other signature analysis techniques in connection with birds (Eastwood, 1967) and even with insects (Glover, 1966). Some of the signal characteristics have been attributed to aspect of the target and others to wing motion. There is ample evidence that insects are to be found in the atmosphere well above the surface. Apart from flying insects, creatures such as spiders can become airborne on strands of gossamer and be borne aloft in convective air currents. Glick (1939) reports in considerable detail the results of collecting insects from aircraft over the southern U.S. and Mexico. He found concentrations of insects of the order 1 per

2 cubic kilometers in the layer between 1000 ft. and 4000 ft. above the ground, with more widely spaced encounters up to four or five times the latter height. Although more recent data do not appear to have been collected, it is common for sailplane pilots to experience many types of insects impinging on the canopy or the leading edges of the wings at altitudes exceeding 10,000 ft. above terrain. Less commonly, birds feeding on insects carried aloft by thermals are observed at similar altitudes.

The radar cross-sections (sigma) of the various insects listed in Table 4 (measured at wavelenths of 3.2 cm.) range from 0.01 cm² to 1.22 cm² for all but the locust which has a maximum sigma value of 9.6 cm². The ability of any given radar system to detect radar cross sections of these low values is a function of its design, its current performance, and the ability of the operator. Ultra-sensitive radar systems such as the MIT Lincoln Laboratory radars at Wallops Island, Va. have reported minimum detectable cross-sections at 10 km. of 6x10^-4 cm² for the X-band, 2.5x10^-5 cm² for the S-band, and 3.4x10^-5 cm² for the UHF radars (Hardy, 1966). The X-band radar is two orders more sensitive than required to detect the listed insects at a range of 10 km. and probably is functioning close to the limit of detectability. The majority of other radar systems in general use today are less sensitive. Some are not able to detect insects in the lower range of a values. Tabulation of a large number of radar system characteristics has been published in classified documents by RAND. Major radar parameters for some airborne sets are listed in an article by Senn and Hiser (1963)

Insects are commonly found at surprisingly high altitudes. Swarms of butterflies and other insects are found in summer on 14,000-ft. mountain peaks in the Rockies. A few insects have been reported at over 25,000-ft. altitudes in the Himalayas.

Verification of insects as causing a particular blip on a radar scope is even more difficult than birds. Flowever, this was accomplished as reported by Glover, et al (1966). Single insects were released from an aircraft and tracked by radar at altitudes from 1.6 to 3.0 km. and at ranges up to 18 km. Experiments of this sort and other studies involving clear atmosphere probing with high-power radars (Atlas, 1966; Hardy, 1966 and 1968) have led to valid conclusions that most of the dot echoes are caused by insects or birds.

Attention has been given by Browning (1966) to the determination of

the velocity characteristics of some clear-air dot angels. A 5.42 cm. pulse Doppler radar with a 1° beam elevated at 30° and rotating at 4 rpm was used in the study. A series of radar soundings spaced about half to one hour apart were obtained at 500 ft. altitude intervals up to 3000 ft. using range-gating techniques. Temperature, humidity and wind data were collected simultaneously with the radar soundings.

Three kinds of angel population were distinguished according to their mean deviation from the swarm velocity, their average vertical motion, their maximum relative velocities and their sigma values. Atmospheric inhomogeneities or the presence of plant seeds appeared to be ruled out because of the small back-scattering cross-sections of individual angels (less than approximately 0.1 cm²), their discreteness in space and velocity, their often quite large mean deviations (up to 4 m sec^-1) from a uniform velocity, and the fact that the only major upward velocities occurred after sunset, at a time when the lapse rate was becoming increasingly stable. The same data suggest insects as the likeliest cause.

Satellites and space debris

Some of the larger man-made objects in space (such as the Echo I and Echo II metallized balloons, Pegasus, and large boosters) have large radar cross-sections and can be detected by search radars^*. For example, Peterson, (1960) found that occasionally the radar cross-section of Sputnik II approached 1000 m². Such space objects at altitudes of around 120 mi. and with speeds of around 18,000 mph could appear as multiple trip echoes if they were detected on a search radar.

Fig. 6 illustrates the possible appearance of the track of a satellite on the PPI of a search radar. The figure assumes a satellite at 120 n. mi. altitude moving radially at a distance of 500 n. mi. from a radar with an unambiguous range of 200 mi. (The elevation angle of the satellite would be about 8° which is within the vertical coverage of many search radars.) When the satellite is at point A the echo is displayed on the PPI at point A', 400 mi. less than the actual range. As the satellite moves to point B its range closes to less than 450 mi. so the echo moves to within 50 mi. on the PPI. From B to C the range of the satellite opens to 500 mi. so the echo moves

---

*However search radars are beamed to low altitudes above the horizon and so would miss satellites when at high altitudes.

NCAS EDITORS' NOTE: This footnote was listed on the errata sheet, but not the text to which it applies; we have placed the asterisk on the page at a spot that seems reasonable from the context.

out to 100 mi. again. An interesting feature of this example is that while the actual path length from A to C is 500 mi. the length of the echo track is only 140 mi. Thus, if the satellite was moving at 18,000 mph the echo would move only 140/500x18,000 or 5,040 mph. At the speed of 18,000 mph the satellite would move 5 mi/sec and take 100 sec. to move from A to C. It is obvious that the rotation rate of the antenna would have to be high to map the entire track of the satellite as it moved from A to C. An antenna rotating at 6 rpm would detect the satellite every 10 sec. and thus get an echo 10 times as the satellite moved from A to C. At slower rotation rates fewer points along the track would be displayed.

Detection of satellites by search radars would therefore result in high-speed echoes on the PPI. If the satellite were moving toward the radar the echo would move at the satellite velocity but would probably be detected for a shorter period since as it approached the radar it would rise above the vertical coverage of the radar beam.

Ionization phenomena

In 1906 J.J. Thomson showed that ionized particles are capable of scattering electromagnetic waves. Sources of ionized particles include lightning strokes, meteors, reentry vehicles, corona discharges from high voltage lines, and static discharges from high- speed aircraft. Ionospheric 'layers and the aurora are also ionization phenomena. These ionization phenomena or plasmas may under certain conditions produce radar echoes on the PPI of a typical search radar.

Plasmas resulting from lightning discharges return echoes which may be seen on the PPI if the operator is looking at the right spot at the right time. A number of investigators (Ligda, 1956; Atlas 1958a) have discussed the appearance of lightning echoes on the PPI. The echoes typically vary from a point to irregular elongated shapes up to 100 mi. or more in length.

A salient feature of lightning echoes is the short duration of the echo from a given lightning discharge. Since the echo lasts about 0.5 sec., it will be evident only on one scan.

The radar cross-section of the ionized column of plasma produced by lightning has been estimated by Ligda (1956) to be 60 m² depending on ion density within the plasma and on the wavelength of the radar illuminating the plasma. Electron densities of 10¹¹/cc are required for critical (100%) reflection of 3 cm. radar energy; only 10⁹ electrons / cc are required with a 30 cm. radar. Thus, longer wavelength radars are more apt to detect lightning than the shorter wavelength radars. There is another factor which aids lightning detection at longer wavelengths. The longer wavelength radars detect less precipitation than the shorter wavelength radars. Therefore, a lightning discharge inside an area of light precipitation might be hidden within the precipitation echo on the PPI of a 3 cm. radar, while a 23 cm. radar might detect the lightning-produced plasmas but not the precipitation.

Confirmation that short-lived (one scan) echoes were caused by lightning was based on the fact that there were visual lightning discharges in the area from which the radar received the echoes. Atlas (1958a), however, estimated (from echo intensities and dimensions) that discharges may occur that are radar detectable, but are not visible to the eye. Whether or not there is visible lightning in the area of these short echoes, there will undoubtedly be precipitation areas in the vicinity. The exact distance from precipitation that lightning may occur has not been adequately studied. It is known that the probability of radar detection of lightning is greatest when the radar beam intercepts the upper levels (ice crystal regions) of thunderstorms. In a mature thunderstorm the ice crystal blowoff or anvil may extend many tens of miles downwind of the precipitation area. Atlas (1958a) illustrates a lightning echo some 10 to 20 mi. ahead of the precipitation echo but within the anvil cloud extending downwind from the storm.

In addition to short duration lightning strokes there is the longer-lived "ball lightning." Ritchie (1961) mentions the controversy surrounding ball lightning and also some of its alleged characteristics such as sliding along telephone wires, fences, or other metallic objects. Radar detection of ball lightning under these conditions is difficult since echoes of the metallic objects and the ground would tend to mask ball lightning near the surface.

Since search radars can detect echoes of very short duration returned by plasmas created by lightning flashes, there is no reason to assume that other plasmas could not be detected by search radars if the plasmas were sufficiently separated from other targets. The radar echoes would probably appear as point targets and if the duration were sufficient to compute a speed, it would correspond to that of the plasma. The possible range of speeds of plasma blobs cannot be given since so little is known about the phenomenon.

In addition to reflections of the radar pulse there is another source of signals from the lightning discharge, those that are radiated by the lightning discharge itself. These signals, called sferics, appear on the PPI as radial rows of dots, as one or more short radial lines, or as a combination of dots and lines (Ligda,1956). Atlas (1958b) states that 10 cm. and 23 cm. radars are good sferics detectors while radars such as the 3 cm. CPS-9 have moderately low range capabilities in detecting sferics.

As with the lightning echo, the sferic duration is very short Atlas (1958b) found an average 480 microseconds for 489 sferics measured during a severe squall line on 19 June 1957. As a result such sferic signals from a given lightning discharge would only be displayed on one scan of the PPIL.

The aurora is a complex phenomenon caused by ionization of the upper atmospheric gases by high-speed charged particles emitted by the sun. Upon entering the earth's upper atmosphere, these charged particles are guided by the earth's magnetic field and give rise to

a luminous display visible only at night. The aurora occurs most often in the vicinity of 67° geomagnetic latitude. In the zone of maximum auroral activity, visual displays can be seen almost every clear night.

Increased auroral activity is found to follow solar magnetic storms. A direct correlation exists between sunspot activity and the intensity and extent of aurora. The increased auroral activity follows a solar disturbance by about one or two days, the time required for the charged particles to travel from the sun to the earth. During these times, auroras may be seen at latitudes far removed from the normal auroral zones.

Auroral displays occur in the ionosphere at altitudes ranging from 54-67 mi. The ionization which is seen as a visual auroral display is formed into long slender columns which are aligned with the earth's magnetic field. This formation results in strong aspect sensitivity which means that radar reflections occur only when the radar beam is approximately at right angles to the earth's magnetic field. Echo strength is proportional to the radar wavelength raised to the third or fifth power; consequently, most radar observations occur at VHF or lower UHF.

As a result only lower frequency UHF search radars within 1000 mi. of the Arctic or Antarctic Circles would be capable of detecting auroral echoes. The echoes would generally appear at true ranges of 60 - 180 mi. for a few minutes to several hours. The echoes would be mainly stationary and could be either distributed or point targets usually in the magnetic north azimuths in the northern hemisphere or magnetic south azimuths in the southern hemisphere.

Meteors are small solid particles that, when they enter the earth's atmosphere, leave an ionized trail from which radar echoes are returned. The majority are completely ablated at altitudes ranging from 50 - 75 mi. Visible meteors vary in size from about 1 gm. to about 1 microgram. The ionized trail produced by a 0.1 gm. meteor is miles long and only a few feet in diameter.

The meteor particle itself is far too small to be detected. Meteors are observed both visually and by radar by the trail of ionization they produce. Because of the distance and the small cross-section of the trail, meteor ionization can be detected by radar only when the trail is orientated at right angles to the radar beam.

Although most meteor echoes last no more than a fraction of a second when observed with VHF radar, a few echoes persist for many seconds. The duration of the meteor echo is theoretically proportional to the square of radar wavelength, and the power returned is proportional to the wavelength cubed. For these reasons, meteor echoes are seldom detected at frequencies above VHF.

Meteor echoes on a low frequency UHF radar usually appear as point targets with a duration of a few seconds or less. Ranges center around 120 mi.

Very, very infrequently meteors occur that are large enough to survive atmospheric entry. They usually produce a spectacular visual display, referred to as fireballs. Such meteorites are detectable by sensitive search radars operating at any frequency and at any angle to its path. Echoes appear as point targets with a duration of a few seconds. The true range would be less than 120 mi. and the range rate generally would be less than 20,000 mph.

Balloons

Balloons and instrument packages or reflectors carried by balloons can be detected by search radars. More than 100 balloons are released over the United States at least twice a day from Weather Bureau, Navy, and Air Force Stations for the measurement of upper atmospheric conditions. A number of these balloons carry radar reflectors as well as an instrument package, and some are lighted for theodolite (visual) tracking. Echoes from these point targets move at the speed of the wind at the altitude of the balloon. Balloon altitudes vary widely and may reach 100,000 ft. so that ground speeds vary from near zero to well over 100 knots. When a balloon bursts and the instrument package abruptly starts a descent which is normally slowed by

parachute, there could be an abrupt change in the behavior of the echo on the PPI. A balloon that had been rising in a direction away from the station would show the range gradually increasing. Then if it descended rapidly the range could appear to decrease which could be interpreted as a reversal of course.

"Chaff," "Window," and "Rope"

When radar was developed as a means for aiming searchlights and antiaircraft guns during World War II, countermeasures were promptly devised. What was needed was something inexpensive and expendable that would give a radar return comparable with the echo from the aircraft. Small metallic foil strips which act as dipole reflectors were employed. The strips are released from an aircraft, and they are wind-scattered which results in a cloud with a radar cross-section comparable to a large aircraft.

The terms "chaff," "window," and "rope" are used to designate particular types of materials. Chaff consists of various lengths of material. Chaff having the same length is called window. Rope is a long roll of metallic foil or wire designed for broad, low frequency response.

Metallized nylon monofilaments have replaced metal foil in the construction of chaff and window. The nylon type is lighter, hence has a slower rate of descent, and is more compact. A typical package of X-band chaff is a cylinder 1 in. in diameter and 1.5 cm. (one half the 3 cm. wavelength) long. The cylinder contains approximately 150,000 filaments and weighs 6.5 gm. and forms a cloud with a radar cross-section of about 25 m². The filaments descend at about 2 ft/sec in still air at lower altitudes, so that if dispensed at 40,000 ft. they take about four hours to reach the ground. Turbulence causes the chaff cloud to grow and disperse, so that generally the signal becomes so much weaker that sometimes the chaff cloud cannot be tracked all the way to the ground.

Since chaff contains a large number of elements the radar signal is similar to that from precipitation. Also it moves with the wind at its altitude. Therefore, it is difficult to distinguish between precipitation and a cloud of chaff by briefly examining the PPI display. When chaff is distributed along a relatively extended path as opposed to only a point distribution, the echo is elongated and does appear to be dissimilar to precipitation.

Rope is a 60 - 80 ft. piece of narrow metallized material such as mylar. It is weighted at one end and has a drag mechanism at the other. When deployed it has a rate of descent about twice as fast as chaff so it would take about two hours to fall from 40,000 ft. to the surface. Usually a number of rope elements are deployed together so there will be some increase in the size of the cloud as it descends.

Smoke

Hiser (1955) reports detecting smoke from fires at a city disposal dump about 15 mi. from the site of a 10 cm. search radar. The radar echo from the smoke plume was evident on the PPI extending in a northeasterly direction to a range of 50 mi. Goldstein (1951) mentions a case where an airplane was directed to an echo observed by a 10 cm. radar. Only several columns of smoke from brush fires were found. Smoke particle size and concentrations are so small that one would be highly skeptical about echoes from the smoke itself. The returns may arise from refractive index discontinuities at the boundaries of the smoke plume. Plank (1956) suggests that echoes from the vicinity of fires may be from either particles (neutral or ionized) carried aloft by convective currents or from atmospheric inhomogeneities created by the fire.

Distant Ground Return and "Angels"

Local terrain features and, at sea, the ocean surface are detected by radar. The range to which such clutter is detected is a function of antenna height, elevation angle and beamwidth, and the distribution of temperature and humidity along the propagation path.

Since normal ground clutter is present day after day, radar operators become familiar with it arid may even use some prominent points to check the azimuthal accuracy of the radar. There are circumstances in which distant, rarely detected terrain features or surface objects return echoes to a radar. The phenomenon referred to as "angels" is also included in this section since at least some of the angels appear to be distant ground return that is detected by reflection or forward scatter of the radar beam by atmospheric inhomogeneities.

To investigate the phenomena of distant ground return it is first necessary to review some of the fundamentals of the propagation of electromagnetic radiation through the atmosphere. The interested reader can find a comprehensive treatment of tropospheric radar propagation in a book on radio meteorology by Bean (1966) which covers in detail the topics in the following brief review.

In a vacuum, electromagnetic energy is propagated in straight lines at the velocity of light, 3x10⁸ m/sec. This constant is usually designated by the symbol "c." In a homogeneous medium, the direction of propagation remains constant, but velocity (V) is reduced and

where mu is the magnetic permeability of the medium and kappa is its dielectric constant and

where n is the index of refraction.

When electromagnetic wave energy encounters a surface of discontinuity in refractive index in a medium, the wave is partly reflected and partly refracted*.

The angle of the incident ray (theta) is related to the angle of the refracted ray (theta') by the equation:

where theta and theta' are the angles of incidence and refraction respectively in the first and second medium, and n and n' are the values of the refractive index for the first and second medium respectively.

The ray is always refracted towards the medium of higher refractive index. A portion of the energy will also be reflected in the same plane and at an angle equal to the angle of incidence if the energy encounters a sudden change in the refractive index; this is a partial reflection. Total reflection occurs when the angle of incidence exceeds a critical value given by (with n₁ < n):

NCAS Editors' Note: this formula appears to be incorrect in the original manuscript, and should probably have been:

In the atmosphere, discontinuities in refractive index sharp enough to cause reflection of the incident wave back to the radar are believed to exist on occasion. Because of the difficulty in making suitable measurements of the physical factors involved, some uncertainty attends the understanding of this mechanism under practical conditions. Detailed discussion of this aspect of propagation is deferred until later where radar 'angels' are described. In the present context, discussion of the effects of refractive index inhomogeneities will be confined to refraction.

---

*For a more complete discussion of atmospheric refraction of electromagnetic rays, see Section III, Chapter 5 and Section VI, Chapter 4. Note, however, the difference in the factors contributing to the refractive index at radar and at optical frequencies.

Where the refractive index gradient is changing continuously as is normally the case in the natural atmosphere as the height above the earth's surface increases, a ray of electromagnetic energy will follow a curved path. The change of direction that this produces may be evaluated by reference to Snell's law by the expression :

n_h(a + h) cos beta = n_s a cos beta₀     (4)

where n_h is the refractive index at height h, n_s is the refractive index at the surface, a is the radius of the spherical earth, beta is the ray elevation angle at height h and beta₀ is the ray elevation angle at the earth's surface (See Fig. 7).

A most important consequence of this is that the effects of a vertical gradient of refractive index are most apparent at low (10° or less) angles of elevation.

Where the refractive index gradient is constant (n_h = n_s) or varies regularly, the curvature of the path of rays of radar energy may be readily determined by reference to the foregoing expressions. In more complicated conditions more sophisticated techniques are available for tracing the path of such rays.

In terms of the real atmosphere, at radar frequencies the refractive index varies as a function of pressure, temperature, and water vapor content. An equation relating the various parameters as given by Smith (1953) is:

where:

• P = total pressure (millibars)
• T = absolute temperature (degrees Kelvin)
• e = partial pressure of water vapor (millibars)

When the available data are given in terms of relative humidity, e may be replaced by e_s R.H., where e_s is saturation vapor pressure at the pressure and temperature of interest and R.H. is relative humidity expressed as a decimal.

For convenience, the left-hand side of the equation is commonly designated N (refractivity) and is expressed in equation is commonly N-units, i.e., N = (n - 1) 10⁶.

Values of N are conveniently derived from meteorological parameters by the use of tables or nomograms, such as those given by the U.S. Navy (1960).

At sea level, a typical value of n is 1.00035, i.e., the refractivity is 350 N units. But depending upon pressure, temperature and humidity the sea level refractivity may range from 250 to 450 N units.

Since pressure, temperature, and water vapor normally decrease with height the refractivity normally decreases with altitude. In a 'standard' atmosphere, typical of temperate latitudes (with a thermal lapse of 2°C/1000 ft. and uniform R.H. of 63%, the gradient (lapse rate) of refractivity is 12 N-units/l000 ft. 39 N. km^-1 in the lower levels. For a constant gradient of this magnitude, a ray will have a curvature of about 1/4^th that of the earth's surface (the radar horizon in this case is about 15% further than the geometrical horizon). For short distances the geometry is equivalent to straight-line propagation over an effective earth with a radius 4/3 as large as the true earth.

A device frequently used to facilitate the consideration of propagation geometry and radar coverage takes advantage of this fact. If a fictitious earth radius is adopted that is 4/3 the earth's true radius, radar rays in the standard atmosphere may be drawn as straight lines, which will preserve the same relationship to the redrawn earth's surface as is the case in reality.

In atmospheres having different constant gradients of refractivity appropriate factors may be applied to the earth's true radius to accomplish a similar result. Typical values ire given in Table 5.

*For an average temperate zone climate; northern climates (e.g. England) tend to be "standard," tropical climates tend to be near-superrefractive (e.g. -80 km^-1).

It is important to recognize the limitations of this device, for even in standard atmospheres initially horizontal rays rapidly reach higher atmospheric levels, at which the refractivity gradient can no longer be represented by the same constant. Again, as will be discussed below, atmospheric conditions frequently depart from the "standard" conditions. The effect of variation in the refractivity gradient on the curvature of radar rays is shown in Fig. 8. Apart from showing the range of curvatures in atmospheres having constant refractivity gradients, this figure indicates the way in which rays can be deflected in passing through atmospheric layers. More specifically, the deflection of a ray in milliradians (Delta tau) in passing through a layer with constant N-gradient is given by:

where the subscripts B and T refer to the bottom and top of the layer respectively. The values of beta are determined at each level in terms of beta₀, N_s (surface refractivity), N_h (refractivity at height h) and h, using Snell's Law (equation 4).

Procedures based on these relationships may be used to trace the path of rays to determine the detailed effect of refraction on radar propagation under any given condition of atmospheric stratification.

The broad pattern of refractive effects, however, is as follows:

• Where the general refractivity gradient lies between 0 N-units/l,000 ft. and 24 N-units/l,000 ft. (100 km^-1) propagation is described as normal.
• Refractivity gradients less than 0 N-units/1,000 ft. are subrefractive and cause upward bending of radar waves with a reduction of distance to the radar horizon. Such conditions may occur where the temperature lapse rate is well above average, or where the atmosphere is drier at lower levels than aloft.
• Where the refractivity gradient exceeds 24-N units/1,000 ft. conditions are said to be superrefractive and radar waves curve down

more strongly. Such conditions result from thermal inversions, i.e., where temperature increases with height, or where the decrease of water vapor content with height is excessive.

For refractivity gradients greater than 48-N/l,000 ft. (157 kin1), the ray curvature will be greater than that of the earth's surface and trapping is said to occur.

This condition gives rise to marked anomalies in propagation and, provided the layer through which such a gradient occurs is deep enough, the radar energy will be guided within a duct bounded by the earth's surface and the upper level of the layer. In such cases, exceptionally long detection ranges are achieved, well beyond the normal radar horizon (See Fig. 8). Where a marked negative refractive gradient occurs in a layer adjacent to the ground, a surface duct is formed (Fig. 9a). An elevated layer of strong negative gradient can also produce ducting (Fig. 9b).

Surface ducts are commonly caused by radiative cooling of the earth's surface at night, leading to a thermal inversion in the air near the surface. In this case, the extreme refractivity gradient is mainly due to temperature effects and such ducts can occur in quite dry air. Where humidity at the surface is higher than usual and falls off rapidly with height, a strong negative refractivity gradient is also established. Evaporation from water surfaces or wet soil can produce these conditions and a particularly common example occurs in warm dry air from the land when it is advected over the sea. This type of duct is commonly found in tropical areas, where temperature and humidity both decrease with height; the inversion type of duct is more common in temperate and artic areas (Bean, 1966).

Elevated layers of extreme refractivity gradient are caused by similar meteorological mechanisms but often occur on a somewhat broader scale. Certain areas of the world are particularly prone to such layers; the California coastal area is a good example. Plate 66 (Blackmer, 1960) shows an example of the PPI during a trapping situation off the California coast.

Plate 66: PPI During Elevated Duct

Click on Thumbnail to see Full-size image.

In this case echoes were presented on the PPI on second and third sweeps but could be correlated with islands and mountainous terrain. Elevated layers such as this are commonly found in the southeast (northeast at S latitudes) quadrants of trade-wind anticyclonic systems.

The anomalous propagation to which such irregular refractivity conditions give rise is of considerable significance to the problem of target identification and false targets. In the first place, the whole basis of the radar technique depends upon knowing the direction in which the radar energy is propagated. For normal practice, propagation must be close to rectilinear. When the radar energy is being strongly curved, information on a target's location derived from the position of the radar antenna can thus be highly erroneous. Again, echoes may be received from the ground or from other targets that are not normally within the range of the radar or within its 'field of view' at any given antenna elevation. Ground echoes from beyond the normal radar horizon are cases in point.

An especially significant condition arises when the antenna is elevated in a direction which is near a critical angle for trapping or ducting. In this case, while much of the energy may be propagated in a direction approximating that intended, because of the finite dimensions of the radar beam, some energy may be severely refracted. This is illustrated diagrammatically in figure 10.

With such a mechanism an aircraft could be tracked fairly accurately, but in addition, echoes could be received from the ground (intermittently if the surface reflectivity or propagation conditions are variable as might be the case in areas of thunderstorms). Such echoes would be displayed as though they were due to targets seen at the angle of elevation of the antenna, and thus at heights which would depend upon their range. A great variety of such possibilities can occur depending upon the geometry involved, the refractive conditions, and the nature of the terrain.

The range of possibilities is further extended if the distribution of radar energy in the side lobes is taken into consideration. With a side lobes strength 30dB below the main beam (a factor of 1000 in power), a side lobes target will yield a return equal in strength to the main beam return of an identical target at a range 5.6 times greater (the 4^th root of 1,000). Thus a target detectable at 100 mi. in the main beam might be detected by the (first) side lobes at a range of up to 18 mi.

Anomalous propagation of the type described is also significant in determining the distribution of energy within the envelope of the main beam, particularly in broad vertical beam systems. At low angles some energy within the beam impinges on the earth's surface near the radar and is reflected, still within the envelope of the beam. Because the path followed by such energy is necessarily longer than the direct path and because of the wave nature of the energy, in-phase and out-of-phase interference will occur, leading to a vertical lobe structure in the beam envelope (see Fig. 10). Anomalous propagation conditions can readily produce variations in the normal distribution of energy within the beam due to this mechanism and thus can easily lead to unexpected variations in signal intensity from distant targets.

It is important to recognize the difficulties that are inherent in establishing whether propagation conditions are anomalous in certain cases. Where the gradient of refractivity extends uniformly over large horizontal areas, there is little difficulty in determining the situation either from conventional meteorological data or from the manifestation of the anomalous performance of the radar itself (for example, the detection of ground clutter to abnormally large ranges). In some cases it is possible to infer, with some confidence, from the meteorological conditions (especially if data on the vertical profile of temperature and humidity are available) that anomalous propagation is not present. In many cases, however, the causative conditions may be very variable in space and time, and it is then difficult to be at all confident

about the nature of propagation at any particular time or in any particular place. Even if timely radiosonde data are available from a nearby location, the information they provide on the thermal and humidity gradient is often inadequate for the assessment of the refractive conditions. In particular, special experimental observations have shown that shallow layers of abnormal refractivity commonly occur either close to the surface or at various levels aloft.

It is often possible to infer only the likelihood or improbability of anomalous propagation conditions by reference to the general meteorological conditions that prevail. Thus one would expect normal propagation in the daytime in a well-mixed, unstable airstream with moderate winds over a dry surface, while expecting marked superrefraction over moist ground during a calm clear night following the passage of a front that brought precipitation in the late afternoon.

Localized conditions favorable for superrefraction are also caused by showers and thunderstorms (Ligda, 1956). The cold downdraft beneath thunderstorms can cause colder air near the surface than aloft while evaporation from the rain and rain-soaked surface, causes locally higher humidities.

In addition to the detection of distant ground targets by refraction of the radar beam, there is the possibility of reflection or forward scatter of the beam to ground targets. Whether or not layers that would reflect the beam to-the ground would also be detected by the radar has been part of the controversy concerning the nature of invisible targets in clear air. These so-called "angel" echoes have been observed since the early days of radar (Plank, 1956; Atlas, 1959 and 1964; Atlas, 1966a). Detailed case studies of selected angel situations illustrate the difficulty of determining the nature of the targets causing the angel echoes. For example, Ligda and Bigler, (1958) discuss a line of angel echoes coincident with the location of a cloudless cold front. They discuss the likelihood that the line was due to differences in refractivity

between the two air masses or to flying debris, leaves, paper, small twigs, birds, insects, etc., carried aloft by turbulence during the frontal passage. Although surface weather instruments recorded a drop of 13°F in less than an hour, this sharp temperature change together with the change in both vapor pressure and atmospheric pressure did not appear to be sufficient to cause gradients of refractivity of sufficient strength to produce the observed echo line. In spite of this difference between refractivity gradients based on surface observations (of pressure, temperature, and moisture) and those required to explain the source of the echo) Ligda and Bigler found serious objections to any hypothesis other than that the echo was due to refractivity gradients. They mention the need for instruments capable of measuring sharp refractivity gradients.

Atlas (1959) studied in detail a situation at Salina, Kans. on 10 September 1956 where cellular and striated echoes covered much of the PPI to ranges of 85 mi. He concluded that the echoes were due to forward scatter from a patterned array of refractive index inhomogeneities to ground targets and back. Recently Hardy and Katz (1968) discussed a very similar radar pattern. They concluded that insects were responsible for the echoes and that cellular pattern of insects was due to atmospheric circulation. Atlas (1968c) agreed that insects may be responsible for some echoes but that the forward scatter explanation is valid in other instances.

Investigations of angel echoes with high-power, high-resolution radars at three different wavelengths have made it possible to learn much about the nature of targets producing various types of angel echoes. Simultaneous observations at 3 cm., 10.7 cm., and 71.5 cm. with the ultrasensitive MIT Lincoln Laboratory Radars at Wallops Island, Va. have been described by Hardy, Atlas, and Glover (1966) , Atlas and Hardy (1966a), and Hardy and Katz (1968a). They found two basic types of angel echoes: dot or point echoes and diffuse echoes with horizontal extent. The dot angels are incoherent at long ranges or when viewed with broad beams but are discrete coherent echoes when viewed by a radar with high resolution. They may occur in well defined layers and may have movements different from the wind at their altitude. Their cross-sections and wavelength dependence are consistent with radar returns to be expected from insects. Since no other explanation fits all the observations of these dot angels, it is concluded that the targets are insects.

Extensive diffuse echo layers have been noted at a variety of heights and sometimes exhibit an undulation or wave motion. The height of these layers coincides with levels at which refractive inhomogeneities may be expected, e.g., at the tropopause. It can be shown theoretically (as summarized by Hardy (1968b) that the measured radar reflectivity of such layers accords well with the theory of the scattering of electromagnetic energy by dielectric inhomogeneity due to Tatarski (1966). The reflectivity (eta) is related to wavelength (lambda) and the coefficient C² which describes the degree of refractive inhomogeneity due to turbulence, by the expression

from which it will be seen that such layers are more likely to be detected by radars operating at shorter wavelengths. Although, because this simple relationship does not apply in the dissipation range of the turbulence spectrum the largest values of n occur at about 5 cm (Atlas 1966b). These phenomena have been much studied recently in connection with the detection of clear air turbulence. (Hardy, 1968b; Ottersten, 1968: and Atlas, 1968b). It is concluded that such turbulence may be detected with ultra high performance radars but only when well marked. (Note that the significant physical feature detected, i.e., the dielectric inhomogeneities, is caused in these cases by the turbulent condition of the atmosphere.)

Radars of the type normally used for tracking and surveillance are unlikely to detect such layers. On the other hand, it has been suggested that on occasion at low levels where marked intermixing of dry and moist air is present, dielectric inhomogeneities will be sufficiently marked and be present in sufficient quantity to produce detectable echoes with radars of relatively modest performance.

Measurements made by Atlas (1953, 1959) and others indicated that atmospheric layers occasionally exist having power reflection coefficients, at normal incidence, of 10^-14 or greater (i.e., 140 db attenuation). The power reflection coefficient of such layers would be greatly magnified if the radar energy impinged on the layer at a small grazing angle. The increase is roughly proportional to the 6^th power of the cosecant of the grazing (i.e., elevation) angle. Thus at a grazing angle of about 10 mrad, the reflected signal would be as high as 10^-2 (a 20 db attenuation). Under actual atmospheric conditions the partially reflected signal of ground objects for example, would be expected to be detectable only at grazing angles (and thus, initial elevation angles) low enough to produce return signals above the noise threshold of the radar receiver. This would produce a "forbidden cone" effect, where no such anomalous signals would be detected closer than a certain range (because of elevation angle, range relation of a layer at a constant height); this has been actually observed in several cases (see Section III, Chapter 5).

It is conceivable that there could be rare occasions when only isolated atmospheric inhomogeneities existed or when the inhomogeneities were such that only the most reflective ground targets were detectable. In such situations only one or two unusual ground targets would appear on the PPI. Levine (1960), in a discussion of mapping with radar, points out how certain combinations of ground and man-made structures act as 'corner reflectors' and return a much stronger signal to the radar than is returned by surrounding features. The sides of buildings and adjacent level terrain, or even fences and level terrain, constitute such reflectors. He states that in areas where fences and buildings are predominantly oriented north-south and east-west, the 'glint' echoes from the corner reflector effect appear at the cardinal points of the compass and have therefore been called a "cardinal point effect." In addition, different types of vegetation have different reflectivities and these vary further according to whether they are wet or dry.

From the above discussion it is obvious that the identification of targets as being ground return due to forward scatter or reflection

is difficult in any but the most obvious situations. Still it should be realized that situations do occur when only very localized areas of ground return may be detected and due to the detection mechanism the location of the intersection of the radar beam with the ground may vary from sweep to sweep of the radar antenna. The problem of verifying whether the target is ground return is greatly complicated by the fact that measurements of refractivity gradients cannot currently be made in sufficient detail around the radar site to describe with precision the medium through which the radar beam is being propagated.

Radio Frequency Interference

During the past 15 years, electromagnetic compatibility (EMC) has emerged as a new branch of engineering concerned with the increasing problems of radio frequency interference (RFI) and the overcrowding of the radio frequency spectrum. The EMC problem is increasing so rapidly that considerable engineering efforts are included in the design, development, RFI testing and production of all new electronic equipment from the electric razor and TV set to the most sophisticated of electronic equipments, such as computer and radar systems. This is true for entertainment, civil, industrial, commercial, and military equipment. The problems are compounded not only because the frequency spectrum is overcrowded, but much earlier generation equipment, which is more susceptible to and is a more likely source of interference, is not made obsolete or scrapped. New generation equipment is potentially capable of interaction problems among themselves, as well as playing havoc with older equipment. Each year sees new users bringing new equipment into the frequency spectrum: such as UHF television, garage door openers, automatic landing control systems, city traffic management and control systems, and a vast array of new electronic devices being introduced into tactical and strategic defense systems.

RFI contributes to the information displayed on radar scopes. It is caused by the radiation of spurious and/or undesired radio frequency

signals from other non-associated electronic equipment, such as navigational aids, data processing computers, voice communication systems, other radars, and from more common sources, such as ignition and electric motor control systems. RFI can also be emitted from the radar system's own components, causing self-induced interference.

Much interference may be sporadic, producing only a short lived 'echo.' There may be instances, however, when the interference occurs at regular intervals that could nearly coincide with the antenna rotation rate so that the spurious echo' might appear to be in approximately the same position or close enough to it that the operator would assume there was a target moving across the scope.

Radio frequency interference can enter the radar system in many places:

1. In the transmitter where it can affect the stability and fidelity of the transmitted output pulse waveform;
2. In the receiver local signal-generating and amplifying circuitry where its effects can be similar to the transmitter perturbations;
3. In the external transmitter/receiver space link where the interference effectiveness depends upon its intensity, frequency, power level, direction of arrival and signal spectral characteristics.

External interference entering on the link through the antenna input is the most common of these possible interference sources. Plate 67 shows some of the more easily recognizable radio frequency interference patterns from other radar systems.

Plates 67a&b: Radio Interference

Click on Thumbnail to see Full-size image.

This type of interference considerably reduces the effectiveness of the radar, but this type of interference, taken alone, is usually readily identifiable by operating personnel. This might not be as true when it occurs in conjunction with extraordinary meteorological, propagation, and equipment degradation phenomena.

The photographs in Plate 66 are time exposures of the PPI. The camera shutter is left open for a full rotation of the antenna so the photograph is generated by the intensity of the cathode ray tube electron beam as it rotates with the antenna. This is in contrast to an instantaneous photograph that would be brightest where the trace was located at the instant of exposure and, depending on the persistence of the cathode ray tube, much less bright in other regions. While the interference in these photographs appears as lines it would appear as points at any given instant. The lines are generated by the time exposure as the points move in or outward along the electron beam. The photographs also show precipitation echoes. Examination of the photographs shows that the interference does not mask the larger precipitation echoes to any appreciable extent but might mask small point targets.

A radar receiver has a limited bandwidth over which it will accept and detect electromagnetic signals. In this acceptance band, the receiver reproduces the signals at the receiver output and displays them on the radar presentation display. Thus any interfering signals that fall within this band will be detected and displayed by the very sensitive receiver. In an S-band (2ghz) pulse radar, the typical bandwidth of the receiver will be 20 - 50 ghz. Any weak signals in this frequency band will be detected. Even out-of-band signals can interfere if they are of sufficient signal intensity to overpower the receiver out-of-band rejection characteristics. For instance, a very strong out-of-band signal of 10 watts might be typically attenuated by the receiver preselection filter by 60 db, reducing it to a signal of -20 db. To the radar receiver,

this can still be a powerful signal, as it might have a sensitivity of displaying signals as weak as from -50 to -80 db or less. It is also likely that the out-of-band interference will be derived from the nonlinear interaction of the desired return signal and the out-of-band interfering signal. The resulting interaction (mixing) of these signals in the receiver can generate still weaker intermodulation products that fall within the passband of the system circuits so that they are displayed. Spurious responses can occur at other than the frequency to which the radar is tuned because of inadequacies in the rejection of the unwanted frequencies in the receiver. The inadequacy is caused by insufficient out-of-band filter rejection coupled with a high level of RFI.

Increasingly more powerful transmitters and more sensitive receiver radar systems need even greater relative suppression of unwanted emission, to prevent the absolute level of out-of-band interference from rising to intolerable levels, thus causing interference to and from other electronic systems.

Even if normally operating radars are not affected by this interference most of the time, the degradation of the radar components or of nearby systems can cause the temporary increase in interference at the radar site. Radar personnel are continually concerned with this problem. Such acts as opening an electronic cabinet can cause the local RFI to increase sufficiently to create an RFI nuisance to the radar system.

Each radar system has been designed to fulfill a single class of target tracking function, being optimized to provide proper and reliable target data a high percentage of the time. However, all systems, including radar systems, have their limitations. Thus, it must be recognized that there will be times when other systems will interfere, component parts will either gradually degrade or catastrophically fail, propagation and meteorological conditions will deviate far from the normal environment, and maintenance and operating personnel will

occasionally fail to function effectively. For all radar and other electronic systems, an increasing amount of effort is expended to reduce the occurrence of these degradations or failures and to minimize their effects.

Lobes and Reflections

Because of radar engineering design limitations, it is not possible to direct all of the transmitter energy into the main antenna beam and small but measurable amounts of energy are transmitted in many other directions. Similarly, energy can be received from such directions, in what are known as the side lobes of the antenna, and can give rise to erroneous directional information. Particularly complicated situations arise when side lobe problems are associated with building or ground reflection mechanisms. For example, if a radar antenna is radiating 100,000 watts peak power in the main beam, 100 watts can be simultaneously radiated from a -30 db side lobe in another direction. Fig. 11 (adapted from Skolnik, 1962) shows a radiation pattern for a particular parabolic reflector. Note that if the main beam is radiating 100 Kw, the first side lobe, the first minor and the spillover lobe radiate about 100 watts. This 100-watt radiation will be reflected from large targets in this side lobe heading but will be shown on the PPI as having the same bearing as the main beam of the antenna. This display of a false target is called a ghost. In this particular instance two targets having identical radar cross-sections would appear as returns of equal intensity if one were in the main beam and the other in the side lobe but 5.6 times closer to the radar.

Highly reflective targets can often be detected in the side lobes. Thus a single large target detected in the numerous side lobes can be displayed in a number of places simultaneously. Since, in radar displays, target echoes are represented as being in the direction in which the antenna is pointing, not in the direction from which the energy is returning at the time of the detection, side lobe echoes from

a single target can be shown as a collection of false targets. Such target outputs from side lobe returns are generally systematically located in the display relative to the main beam return signal. Therefore, in general, side lobe return signals are readily identifiable by the operator and will tend to cause obliterati