Introduction

The ultimate aim of any radar experiment is of course to determine information about the structures which backscatter the radio waves, and the environment in which they exist. For example, it might be of interest to study the wind speeds associated with the scatterers, or the shape of the scatterers, or to differentiate types of scatterers or reflectors. It might be of interest to determine the radar cross-section of the scatterers, or their spatial distribution over the sky. Other desired information might include the spatial and temporal variation of the scatterer velocities as a function of time and height. If the radio scatter is due to turbulence, it might be desirable to measure the intensity of the turbulence, and/or its spatial distribution. It might be of interest to determine the relative percentages of turbulent to non-turbulent scatter. The list can go on.

In the preceding chapters, we concentrated on: (i) the principles of radar (Chapters 2 to 6); (ii) signal processing procedures (Chapters 3 to 5); and (iii) the nature of the scattering mechanisms (especially Chapter 3). Now is the time to bring all this information together and look more closely at the interaction between the radar and its scattering environment. In particular, we want to determine how the radar may be used to deduce information about the scatterers themselves. This information could include all sorts of spatial scales, right down to the radar wavelength (often indirect information at such small scales), and a wide variety of temporal scales, from fractions of a second to many years.

The purpose of this chapter is therefore to discuss ways that relevant atmospheric parameters can be determined and then interpreted, in order to give new insights into the nature of the scatterers. We will re-examine some of the parameters already discussed, like spectral characteristics, and we will also introduce new ones, like the turbulence anisotropy, amplitude distributions, phase distributions, turbulence strengths, tropopause height, and so forth. Some of the approximations used in determining these parameters are also critically examined. Some consideration will be given to experimental design, and then interpretation of the results. Studies of the parameters evaluated over long periods of time can give a considerable amount of additional information, over and above that which can be determined from a few discrete observations, but discussion of this aspect will not be considered in great detail, due to lack of space.

This book is about designing, building, and using atmospheric radars. Of course the term “atmospheric radar” covers a wide and diverse set of instruments, which can be used to study a wide range of atmospheric phenomena, and we cannot cover all radar types nor all applications. However, radars used for MST (Mesosphere-Stratosphere-Troposphere) studies employ a very high percentage of the techniques used in atmospheric studies, and cover an extraordinary range of physical processes. Therefore we have chosen this field as our focus. A reader familiar with this book should not only have developed a broad comprehension of the MST region, but should be able to diversify easily to other fields of atmospheric radar work.

While the primary targets of this book are new and advanced graduate science and engineering students working with radar to study the atmosphere, we have also aimed to make it accessible and useful to a wider audience. The extensive references and diagrams should make it valuable as a general reference resource even for more experienced workers in the field. The level of difficulty in each chapter has been adapted to suit the standards of a student with a modest background in mathematics and signalprocessing. Some level of understanding of Fourier methods, including Fourier integrals, is desirable, although not mandatory. Nevertheless, some of the chapters are pitched at a level which could be followed even by an interested amateur. Chapter 2, for example, gives a moderately detailed history of the development of atmospheric radar, examining the development of experimental radio applications for both meteorology and world-wide communication following World War II, and would be of interest to, and easily comprehenced by, an enthusiastic radar hobbyist or history buff. Yet the detail on scatter processes in Chapter 3 in regard to the refractive index of the atmosphere and ionosphere should be enough to satisfy more discerning tastes in mathematical complexity.

The layout of the chapters has been carefully developed, mixing the areas of technical detail and practical application in a way that we hope will keep the reader stimulated as we develop parallel themes of radar engineering, experimental design, application and understanding of meteorological/atmospheric physics and chemistry.

We begin with an overview of the atmosphere which can easily be comprehended by a reader with no knowledge at all of radar.

Introduction

It should be clear from the foregoing chapters that the range of applications of MST and windprofiler radar is broad and challenging. Some techniques are mature, some are under development, and some are even no doubt yet to be discovered. Measurements of wind velocities and, by extension, wave motions, wave-mean flow interactions, momentum flux deposition and turbulence, are possible. Capabilities for temperature measurements, and the possibility of humidity measurements, have been discussed. Strange echoes such as polar mesosphere summer echoes have given new insights into the plasma processes of the lower thermosphere. Studies of turbulence anisotropy are possible. We have demonstrated functional radar designs that cost as little as $100 000 up to many millions of dollars.

We will not dwell on these many achievements, however, which should be selfevident. What is perhaps of greater interest is the future of these instruments, and this will be the main focus here.

The future

The future harbors both pragmatic and curiosity-driven aspects. From the point of view of the former, networks of radars, providing data for incorporation into computer forecasting and now-casting models, offer the hope of better forecasts. They have been shown to have benefits in forecasting on time-scales from a few hours out to several days, especially with systems deployed in Japan, Europe, and Canada (see Chapter 12). At the time of writing (2015), the European Space Agency is about to launch a specialized satellite instrument (AEOLUS) for measurement of tropospheric winds from space by lidar, and the networks of windprofilers discussed will be crucial tools for validation of these data. However, since the satellite only measures winds at sunrise and sunset, the radars, with their continuous recording capability, will continue to provide valuable input to meteorological models for many years to come.

Accurate records of winds are of course valuable for large-scale forecasts. This can impact aircraft travel, allowing better flight planning. The ability of radars to make reliable measurements of turbulence strengths can also be of value from the perspective of aircraft passenger safety.

Introduction

As we have already discussed, there are many competing factors that must be taken into account in order to optimally investigate the atmosphere through radar observations. One of the more notable examples is the Doppler dilemma. Obviously one would like to select an inter-pulse period (IPP) corresponding to a sufficiently large Nyquist velocity interval. Here sufficiently large means a velocity range that encompasses most of the anticipated radial velocities to be observed. The range of Nyquist velocities is extended by decreasing the IPP. However, decreasing the IPP also reduces the maximum unambiguous range that can be measured. Ideally one would like to maintain a large Nyquist velocity (short IPP) and large maximum unambiguous range (long IPP) – hence the dilemma. Another example involves the disparity between the desire to improve range resolution and improve radar sensitivity. Range resolution can be improved by decreasing the radar pulse width; however, this means a decrease in the amount of power that illuminates a scatterer and corresponding decrease in detectability. That is, the desire to increase the detectability of atmospheric signals by transmitting longer radar pulses is at odds with the need to improve range resolution.

In many cases, techniques have been developed that allow us to work around the compromises that arise in designing radar experiments. For example, pulse compression (discussed in Chapter 4) is used to improve range resolution without compromising the signal-to-noise ratio (SNR) (Schmidt et al., 1979). By and large, such techniques are known to introduce corresponding undesirable side effects. For the case of pulse compression, either the existence of some level of range side-lobes, or a decrease in temporal resolution, are a by-product of complementary codes.

In this chapter, we discuss how the use of multiple-receiver and multiple-frequency techniques can be used in atmospheric remote sensing as a means of improving angular and range resolution respectively. Before proceeding, we should clarify one point of nomenclature. The term multiple-receiver will be used throughout this chapter to describe a radar system that is capable of receiving and recording atmospheric signals on two or more spatially separated antennas or groups of antennas. The myriad names associated with interferometric techniques were discussed in Chapter 2, Section 2.15.6: here, we will discuss in detail just a subset of these, but the points discussed will cover to some extent all the techniques.

Introduction

Wind motions in the atmosphere can cover a wide range of temporal and spatial scales. They may include variations on annual, seasonal, monthly, daily, hourly, and minute scales, down to scales of seconds. Spatially, motions may cover global scales down to synoptic (continental-sized), meso- (city-sized) and microscales (e.g., Ahrens, 1999, Figure 10.1).Windprofiler radars can study all of these scales. However, we cannot possibly cover all of them in this chapter. Larger scale motions (including planetary waves and tides) can be studied well with satellites and in-situ instruments carried by balloons and rockets, as well as numerical computer models. While profilers can also contribute here, it is at the smaller scales that windprofilers really make their best contributions. We will therefore concentrate in this chapter on synoptic, mesoscale and microscale motions, with strongest emphasis on the last two. The primary focus will be on height regions where MST radars have made a significant contribution, restricting discussion to the troposphere, lower stratosphere (below 25–30 km), and the upper mesosphere and lower thermosphere (60 to 100 km altitude). Other height regimes will be discussed primarily in their relation to these regions.

In meteorology, atmospheric mesoscales motions refer to spatial scales between a few kilometers and one or two hundred kilometers, and temporal scales of the order of minutes to a few hours. In the troposphere, mesoscale events include thunderstorms, tornadoes, and various types of local circulations like land and sea breezes and valley breezes. Typical synoptic scale events include hurricanes, high and low pressure systems, and frontal systems.

Some or all of these events may be quite familiar to many readers. In fact, these events are only really dominant in the lowest few kilometers of the atmosphere. MST atmospheric radars can be used to investigate these phenomena, and this has been done in the past (e.g. Strauch et al. (1984); Gage et al. (1991a); Webster and Lukas (1992); Teshiba et al. (2001) (and references therein); Röttger and Larsen (1990); Hooper andPavelin (2003), among others). Since this book has a special chapter on meteorology, these events will not be pursued here in any detail.

When MST radars are used for studies to heights of ten kilometers and more, and even into the upper atmosphere, a different class of mesoscale/synoptic scale motion becomes apparent. This motion is often well organized and can propagate over large distances.

Introduction

Some of the earliest applications of windprofiler radars were in regard to tropospheric and lower stratospheric studies. The radars developed at the Sunset site near Boulder, Colorado (Green et al., 1979) and in the Harz mountains in Germany (the SOUSY radar (Czechowsky et al., 1976)) were two of the earliest such instruments, and were certainly built with meteorological studies in mind. Some of these radars have already been described in Chapter 2, and the SOUSY radar was extensively discussed in Chapter 6.

The most direct meteorological studies have been in regard to wind motions, but these radars have also been usefully employed in other areas, including studies of turbulence strengths and anisotropy, tropopause height measurements, gravity wave momentum fluxes, precipitation measurements, temperature profile determinations, and various others.

It is impossible to cover all aspects of MST radar applications relating to the troposphere in just one chapter. For this reason, we will concentrate mainly on results, rather than on specific details about techniques. It will be assumed that the techniques have been sufficiently covered in earlier chapters.

The early years of tropospheric studies have been especially well covered in several excellent reviews, including those by Röttger and Larsen (1990), Gage (1990), Larsenand Röttger (1982) and Balsley and Gage (1982). Some of the early parts of this chapter will involve a recap of the main results of those publications.

Röttger and Larsen (1990) discussed the origins of VHF MST radar studies in the context of: (i) developments following the use of high-power X, S, and UHF band radars in the United States of America, as well as FMCW (frequency modulated continuous wave) techniques; coupled with (ii) the detection of tropospheric echo fading observed at Jicamarca (Peru) by Woodman and Guillen (1974); and (iii) the application of phasecoherent techniques. These events in turn led to the first dedicated VHF-ST radars being built at Sunset, near Boulder, Colorado, and in the Harz mountains of Germany (the SOUSY, sounding system radar). Phase coherent detection was especially important in the development of such systems, for without it, detection of useful tropospheric echoes with VHF systems would be nearly impossible.

Introduction

The groundwork describing the atmospheric environment and the types of flows that radar can study in the Earth's atmosphere has been laid in the previous chapter. We now turn to a brief history of how radars came to be involved with studies of this type.

While most of this book is about MST radar, it is important that MST radar be seen in a broader context. We therefore begin this section on the history of the development of MST radar by looking not at MST radar itself, but rather at the development of meteorological radar. As indicated earlier, the period following World War II saw various developments of radar. Two primary streams were (i) ionospheric studies for world-wide communication, and (ii) studies of contaminants in radar detection for military and civil applications. The first stream of development led to extensive studies of the upper atmosphere and ionosphere, and the second led to more intensive investigations of the troposphere. Only with the development of MST radar did the two streams once again really merge.

Initially, there were two main aspects to radar detection – determination of range and, if possible, direction. Directional determination was achieved by using large antennas which concentrated the radar directionally, and range was generally found using timeof- flight delays.

The atmospheric radar principle for range-detection is basically fairly straightforward. A short pulse of an electromagnetic wave of typically several microseconds duration is transmitted from the radar antenna, whereupon it eventually may strike a target. It is then scattered back from the target to the radar antenna. The receive signal is called an echo, by analogy with the sound heard when your voice echoes from a distant object. Multiple radar echoes can be detected if there are multiple targets.

Echo samples are examined at consecutive time steps. Using early radars, this was done visually, whereas with more recent ones, digital sampling is used. Since the radar signal propagates with the speed of light c, the time t elapsed from the transmission of the pulse to the reception corresponds to a given range r = ct/2. We find, as an example, that echoes from backscattering targets at a range of, say 15 km, are received 100 microseconds after the pulse was transmitted. Echo samples are taken at a series of successive delays, called range gates.

Introduction

As discussed in Chapter 2, the MST technique began in part with Woodman and Guillen (1974) after discovery of atmospheric echoes from the troposphere using the Jicamarca incoherent scatter radar in Peru. Following this, recognizing the potential for meteorological applications, several groups set about building specialized radars for low altitude (less than 20 km) application, based broadly on the Jicamarca system. Primary groups who followed this course included a NOAA group in Boulder, Colorado, and a group at the Max Planck Institut für Aeronomie in Northern Germany. The NOAA Aeronomy group developed the so-called “Sunset radar” which was installed close to Boulder, and later a larger system at Poker Flat in Alaska. The Poker Flat system was in part also designed for mesospheric studies. The Max Planck group built a radar in the Harz Mountains, near Katlenburg-Lindau. These were the first VHF instruments designed specifically for meteorological studies.

Later, similar radars were developed by other groups in the UK, Japan, Australia, and various other countries, eventually leading to large networks of such radars. The term “windprofiler” was adopted to describe such radars when used for tropospheric and lower stratospheric (meteorological) wind measurements. One notable development was the construction of the large MU (middle-upper) radar near Shigaraki in Japan. At the time this was a state-of-the-art instrument, and had many important developments incorporated into it.

In this chapter, we will describe some of the details of three radars. One will be the German SOUSY radar in the Harz Mountains, one the MU radar, and the third a lowcost radar (CLOVAR) built in the 1990s in Canada. The objective is not so much to discuss the history of these radars (that was considered in some detail in Chapter 2), but to give more detail about technical developments through the course of evolution of the MST technique. The SOUSY radar was built at a time when personal computers were just starting to be developed, but were of very slow speed. Instruments like the Data General PDP-8 and NOVA mini-computers were just under development; the PDP-8 was developed around 1965, and the NOVA came into being in the late 1960s and early 1970s.

Introduction

In this chapter, we discuss various extended, and in some cases unusual, applications of MST radar. These may be special cases of general MST techniques, or specific applications of the technique applied to special cases, or even quite unusual applications which are a substantial deviation from “normal” MST standard practices. If such a topic fits well in another chapter, it may appear there – if it is somewhat of an exception, or has a sightly unusual methodology, or is not really an operational technique, it may appear here. Polar mesosphere summer echoes are an example of an “extended” application. While the techniques used to study these unusual echoes are really the same as for other MST studies, the unusual physics associated with the scatterers that produce these echoes makes them of particular interest. Lightning study is an example of a slightly “miscellaneous” application, in that the techniques are a little unusual (high PRFs, and the events are very short lived). Meteor study is an example of a slightly non-standard application that has grown into a substantial field all of its own. Differential absorption is a technique developed early in the days of radar in the 1960s and 1970s which has had a rebirth in the last decades, and deserves a brief mention here. Precipitation study with MST radars is a relatively mature field, but is still a secondary application, so is also included here.

Each of these fields has a significant role in its own right, but extended discussion of them would simply take up too much space, and would spread the intended application of this book beyond its original goals. Hence the topics are summarized briefly in this chapter – maybe too briefly for some, but we have tried to give sufficient references that interested readers may expand their knowledge through these references.

This book is intended to concentrate on experimental and analysis techniques, and the underlying processes (both geophysical and technical) that guide the experiments and their design. Examples of the latter include the basic theory of turbulence, and the theory behind gravity waves (see the next chapter, and also some small discussion in Chapter 2).

Introduction

Many instruments have been used to study the atmosphere, both by in-situ and remote methods. From anemometers to satellites, chemical sensors to balloons and rockets, the array of tools is broad. Since the early 1900s, a key instrument for such studies has been radar. RADAR stands for Radio Detection And Ranging. Radars operating in a variety of frequency bands, from wavelengths of kilometers to wavelengths of millimeters, have all found application. They have been used to study the upper ionosphere and the neutral atmosphere, right down to ground level.

In this book, we will concentrate on a class of radar generally referred to as MST radar. In this description, M stands for Mesosphere, S for Stratosphere, and T for Troposphere, where these three “spheres” refer to different height-regimes of the atmosphere which collectively cover the region from ground level up to about 90 km altitude. More exact definitions will be given shortly. For now, consider the troposphere as the region from the ground to 12 km altitude, the stratosphere as the region from 12 to 50 km altitude, and the mesosphere the region from 50 to 90 km altitude. Under the narrowest definition, the term MST radar was originally used primarily to refer to radars operating in the VHF (very high frequency) band, with special emphasis on frequencies around 50 MHz, which could probe (at least in part) all three regions. More generally it has come to refer to any radars that can be used for studies of any of these three regions of the atmosphere. These radars include MF (medium frequency), HF (high frequency), VHF, and UHF (ultra-high frequency). They also include so-called meteor radars. Generally, precipitation radars (referred to as “Doppler radars” by the meteorological community) are not considered to be MST radars, although we will discuss them a little in this book. (As an aside, we will generally refer to these radars as precipitation radars in this book. The phrase “Doppler radar” is not a good one to describe these radars, since they are most certainly not the only Doppler radars! The term “Doppler radar” arises from the fact that these radars can measure the Doppler frequency-shift of reflected signals.

Introduction

Fundamentally, atmospheric radars are designed to transmit an electromagnetic (EM) wave and to observe the effects that the atmosphere has on the scattered wave. These interactions may take the form of bending of the radiowave path, or reflection and scattering. In the simplest case, a transmit antenna and a receive antenna are required, which may be located at separates sites. More complex systems might involve multiple receivers and even multiple transmitters. Most commonly in MST atmospheric work, the transmitter and receiver are co-located; in these cases, refraction of the ray paths is not generally significant. Reflection and scattering are the primary phenomena that need to be considered in MST studies.

The radar targets in MST studies

Atmospheric reflection and scattering occur due to the interaction of the EM wave with changes in the refractive index. As discussed in Chapter 3, these refractive-index changes may be caused by a variety of phenomena. We will quickly revisit some of these processes here, because they help us to understand the different modes of radar analysis that we will discuss. Of course, aircraft and missiles are perhaps the most obvious examples of targets that spring to mind when we talk of radar, but these are not the primary targets when it comes to atmospheric studies. One simple example that is relevant is water droplets embedded in the air. In this case, the refractive index inside the water droplets is very different to that of the surrounding air, so each water droplet may scatter a small amount of incident radiation. In this case, scatter from a large number of water droplets is required before a detectable scattered signal can be produced. Insects and birds contain water, so they too can act as radiowave scatters. Indeed some radars use insects as tracers of atmospheric motions. Another example is the ionized trail of plasma left behind when a meteoroid enters the atmosphere. Meteoroids are generally small grains of dust (with diameters from micrometers to centimeters, though larger ones can occur) which enter the atmosphere at high speed (typically 10 to 70 km/s), creating large levels of frictional heating and thereby ionizing the air around them. As a result, a long trail of plasma (typically a few km in length) exists behind the meteoroid, and this so-called “meteor trail” can reflect radiowaves.