CRAF Input to CEPT Detailed Spectrum Investigation - phase III

The European Science Foundation is an association of its 65 member research councils and academies in 22 countries. The ESF brings European scientists together to work on topics of common concern, to co-ordinate the use of expensive facilities, and to discover and define new endeavors that will benefit from a co-operative approach
The scientific work sponsored by ESF includes basic research in the natural sciences, the medical and biosciences, the humanities and the social sciences.
The ESF links scholarship and research supported by its members and adds value by cooperation across national frontiers. Through its function as coordinator, and also by holding workshops and conferences and by enabling researchers to visit and study in laboratories throughout Europe, the ESF works for the advancement of European science.

On behalf of European radio astronomers, the ESF Committee on Radio Astronomy Frequencies, CRAF, coordinates activities to keep the frequency bands used by radio astronomers free from interference.

COMMENTS to the Detailed Spectrum Investigation, Phase III

by the European Incoherent Scatter Association (EISCAT)

Kiruna, October 13, 1998 - Gudmund Wannberg, Deputy Director

1. Introduction

Incoherent scattering of radio waves from plasma is a physical process, which has been understood in principle ever since the classical work on the properties of the electron by J.J. Thomson in the late 19th century. It was realised that free electrons immersed in an electromagnetic field would be accelerated by the electric component of the field and thus reradiate a part of the energy contained in the field. Once the mass and charge of the electron were determined, it became possible to compute the electron scattering cross section. This proved to be so small, of the order of 10-24 m2, that the process was initially believed to be of no practical significance.

However, the development of high power VHF and UHF radars during WW II pushed technology to a point where physicists and radar engineers realised that, given a large enough radar, it should be possible in principle to detect incoherently backscattered radar returns from the plasma in the ionosphere. One should then be able to determine the electron density at altitudes above the F-layer peak, which normal ionosondes cannot reach. Tests were performed at different laboratories during the early 1960s and backscattered signals were eventually received on the ground. It came as a pleasant surprise that the initial predictions of signal strength and information content were overly restrictive, and that in fact the radar echo contained a wealth of information about the scattering plasma. This led to the establishment of a number of dedicated incoherent scatter radar (ISR) facilities in many areas of the world during the 1960's and 1970's.

2. International status of ISR based research

The ISR method is well established as the technique of choice for a multitude of ground-based studies of the ionosphere, both as regards its basic physical and chemical properties as well as certain aspects of ionospheric HF radio propagation. There are at present four ISR systems in operation in the American sector (the most well known located at the Arecibo observatory in Puerto Rico) and one in Japan. The American radars are all financed by the U.S. National Science Foundation and managed under contract with different research institutions. The European situation will be addressed below.

ISR based science enjoys strong support from URSI, where incoherent scatter users are well represented in Commissions G and H. At its 1990 General Assembly, the URSI Council passed a resolution, U.22 (1990), highlighting the spectrum allocation and interference problems common to all ISR installations and urging the relevant ITU bodies to address these in a constructive manner.

Incoherent scatter is only applied for purposes of basic research and derives no economic benefits whatsoever from its spectrum usage; the information gathered by ISR systems is on the other hand of general interest to the society, and may in the future become very important for environmental research.

3. The European ISR scene.

In Europe in the 1970s, one ISR system was operating in France and another in the U.K., both of which have now ceased operation. There is currently an ISR system at the Polytechnic Institute in Kharkov in the Ukraine, about which little is known. The Ukrainians have recently started collaboration with U.S. research groups and more information about their activities should be forthcoming soon.

In 1981, the European Incoherent Scatter Association, EISCAT, began ISR operations in northern Scandinavia. EISCAT is a joint undertaking of the science research councils (or equivalent) of six European countries, viz. Finland, France, Germany, Norway, Sweden and the U.K. Two years ago, the Japanese Institute of Polar Research entered the Association as a full member. EISCAT owns and operates a dual radar system in Tromsö, Norway, with additional receiver sites in Kiruna, Sweden and Sodankylä, Finland. This mainland system is currently in full operation, accumulating about 1500 operating hours annually. EISCAT has recently added a fourth radar site, located on the island of Spitzbergen in the Svalbard archipelago. The EISCAT customer base comprises research groups and individuals from the member countries, totaling about 200 European scientists. EISCAT data is regularly submitted to international ionospheric science databases, where the worldwide scientific community can access it.

The demand for interference-free reception is as stringent in ISR work as in radio astronomy. Due to this commonality of interests, EISCAT is a full member of the European Science Foundation subcommittee for the protection of radio astronomy frequencies, CRAF.

4. Physical and technical aspects of incoherent scatter radars

4.1 Operating principle

In a typical ISR, short, high power RF pulses, modulated by some pulse compression sequence, are transmitted from a directional antenna upwards into the ionosphere. Nearly all the transmitted power escapes to space, but a minute fraction is scattered by the plasma, and a part of that scattered field is intercepted by large aperture receiving antennas on the ground, amplified and processed into physical parameters.

The signal scattered from the plasma back to the radar receiver is the sum of scattering contributions from all free electrons illuminated by the radar pulse. Each electron imparts a Doppler shift to the part of the incident wave scattered by it, proportional to its velocity. The electron velocities are governed by the laws of thermodynamics and electromagnetism and by the internal state of the plasma. While it is the electrons that scatter the signal, the motion of the ion component of the plasma strongly couples to the motion of the electrons through electromagnetic interactions. The Doppler spectral distribution of the received signal therefore contains implicit information about the electron temperature and density, as well as the ion temperature, the ionic composition and/or the ion-neutral collision frequency. Additionally, the mean Doppler shift of the received signal provides a direct measurement of the bulk velocity of the target.

As the scattering process is weak, the transmitted pulses propagate almost unattenuated through the whole ionosphere and illuminate the plasma all along the propagation path. Thus, scattered signals can be received from a range of altitudes from approximately 80 km up to a maximum altitude, which depends on the system sensitivity, typically at least 800 km. Some radars are sensitive enough to reach up to more than 1500 km altitude, albeit only after integrating the radar returns over time for up to several minutes.

By analysing the spectral shape of the radar return as a function of range and time, information about the temperature, composition and state of motion of the ionosphere can be derived. This assumes that the spectral measurement has been done with a strictly amplitude-linear system, because overload and intermodulation in the radar receiver can severely distort the results. Incoherent scatter is thus an extremely powerful tool for ground-based remote sensing of space plasma. It has provided the scientific community with more information about the topside ionosphere than any other technique. ISR is now being applied to the study of new important areas of research, such as solar-terrestrial interaction in the Earth's high latitude ionosphere and magnetosphere, and related problem areas such as middle atmosphere dynamics and the ozone problem. ISR systems are also often used for ground-based support of major satellite projects.

4.2 Modulation bandwidth

In contrast to the situation prevailing in conventional radar applications, where targets are hard and compact, and where the modulation is chosen to obtain some acceptable compromise between range and Doppler resolutions, ISRs receive scattered returns from all the plasma along the instantaneous line of sight. Over the altitude range from 80 km to, say, 1000 km, the scale length (i.e. the distance over which typical atmospheric and plasma parameters such as pressure, temperature or composition vary by 1/e) changes from a few hundred meters at 80 km to several hundred kilometres at 1000 km. As a consequence, to resolve the ionospheric target properly at all altitudes one requires several different modulations spanning a 1000:1 range in modulation bandwidth (from about 2 MHz instantaneous bandwidth for the lowest altitudes to about 2 kHz for the highest altitudes). Since the modulation bandwidth is determined by the required range resolution, it is independent of the radar carrier frequency. To improve the measurement statistics, pulse patterns are often repeated on a number of transmitter frequencies in sequence. The bandwidth required for full utilisation of an ISR transmitter is therefore at least 2 MHz for a moderate duty cycle system, increasing to about 4 MHz for a 10 % or more duty cycle system.

4.3 Detection bandwidth

The dominant feature of the scattered signal is centered on the transmitted carrier frequency. The spectral distribution of this signal contains information about the physical state of the target, and its bandwidth is to first order directly proportional to the radar carrier frequency. For an ISR operating at 900 MHz, the scatter bandwidth is of the order of 40 - 50 kHz. The full bandwidth of the received signal is equal to the convolution of the modulation bandwidth with the scatter bandwidth. It follows that for narrow-band modulations the required detection bandwidth is almost equal to the scatter bandwidth, whereas for wide-band (i.e. high range resolution) modulations the detection bandwidth is almost equal to the modulation bandwidth.

In addition to the main feature described above (the so-called ion line echo), the incoherent scatter process also produces a number of weak, narrow-band returns at frequencies offset from the transmitter carrier by varying amounts up to as much as +/- 10 MHz. These plasma lines and gyro lines contain additional information about the plasma state. Under favourable conditions it is possible to use them to measure e.g. the vertical electric currents flowing in the ionosphere. However, successful operational use of plasma and gyro line measurements requires a reasonably clean reception band, about 20 MHz wide, centred on the transmitter frequency.

4.4 ISR operating frequencies

To first order, the electron scattering cross section is independent of radar frequency, so in principle an ISR can operate in any convenient frequency band and still provide comprehensive information about the target plasma. In practice, the situation is more complicated. The collective plasma-physical effects, which make incoherent scatter such a powerful diagnostic, gradually become ineffective at radar wavelengths shorter than approximately 30 cm, thus limiting the operating frequency of an ISR to something less than about 1 GHz. When boundary conditions like the cosmic noise spectral distribution and the physical sizes of ISR antennas are also considered, there is a broad optimum sensitivity region stretching from about 350 MHz to 500 MHz. Frequencies above and below this region have been applied occasionally for special reasons. The lowest frequency ISR system operating today uses frequencies around 50 MHz, while one system in Greenland operates at 1.3 GHz, however with greatly limited altitude coverage.

4.5 ISR system parameters

Because the incoherent scatter process is so weak, ISR systems must use very high transmitter peak powers, large antenna apertures and very quiet receivers to achieve the sensitivity required to successfully detect and process radar returns from ionospheric altitudes. Typical transmitter peak powers range from 1 to 3 MW. The antennas are often fully steerable reflector systems, with surface areas in the 500 - 1000 m2 range. System noise temperatures are kept well below 100 K at all radars operating above 400 MHz. The average duty cycle of an ISR transmitter is of the order of some 5 - 10 % (cf. the much lower duty cycles of air traffic control radars...), and some systems approach the 25 % mark. Thus, the average transmitted power of an ISR installation can be in the order of several hundred kilowatts, representing a substantial interference potential to other services.

At the same time, the ISR receiver has to process backscatter signals with typical power densities of - 190 dBm/Hz or less, referred to the receiver input. With careful engineering, this can be done with preserved linearity over a dynamic range in the order of 70 dB, but in-band interference from communications services (e.g. NMT900 or GSM) is almost certain to saturate the system. Co-channel interference is completely disastrous; incoherent scatter is a linear remote sensing technique with nearly no co-channel interference margin!

4.6 The EISCAT UHF system

Of the three EISCAT radars, only the UHF system uses frequencies falling in the DSI III range. The UHF transmitter site is located near Tromsö, Norway. The transmitter can operate at 931.5 +/- 4 MHz with a peak power of 1.5 MW into a 32-m diameter steerable parabolic dish antenna. Pulse lengths vary from 1 (s to 10 ms. Reception of signals in the 916.5 - 935.0 MHz range is done both in Tromsö as well as at two receive-only sites at Kiruna, Sweden and Sodankylä, Finland. The UHF transmitter power amplifier uses a specially designed internal cavity klystron; any centre frequency change greater than approximately plus or minus 1 MHz will require rebuilding the klystron.

4.8 Technical summary

For fundamental physical reasons, ISRs must operate in the (50-1000) MHz range

  • Precise frequency is largely unimportant over the (200 - 1000) MHz range
  • The transmitted spectrum can cover up to 2 MHz per carrier, up to 8 MHz for high duty cycle frequency-hopping systems
  • The receiving bandwidth is mainly determined by the transmitted spectrum. To recover all information in the scatter signal, up to 20 MHz bandwidth is required
  • The scattering process is very weak, low noise receiving systems are required
  • The method is very sensitive to interference. Quiet frequency assignments are required

    5. Frequency allocation and licensing issues

    5.1 Formal ITU status of ISR systems

    Incoherent scatter radars have never been officially recognised by the ITU as belonging to any particular Service. In a formal sense, though, ISRs fall under the definitions of radiodetermination and radiolocation systems as given in the ITU Radio Regulations, Chapter I, Section I, articles 1.8 and 1.10, and Section III, article 3.27. Licensing authorities should therefore be within their full rights to assign frequencies in any band designated for radiolocation to ISR systems. However, the stringent need for interference-free reception makes it unattractive, and often impossible, to operate ISRs in allocations that must be shared with other radar systems.

    5.2 Current EISCAT UHF frequency allocations

    Since all EISCAT transmitters are located in Norway, it is the Norwegian frequency planning authority, Statens Teleforvaltning, which assigns transmitter frequencies and issues transmitting licenses. The UHF system is currently licensed to operate on a non-interference basis in the 927.5 - 935.0 MHz band. The Swedish and Finnish receiving sites are not currently subject to licensing requirements and are therefore operated without any formal permits. In Finland, a degree of protection of the receiving spectrum is offered through a ruling of the Telehallintokeskus, prohibiting the use of the 935.0 - 936.5 MHz band by NMT and/or GSM base stations within a 100 km radius from the EISCAT station in Sodankylä. So far, no corresponding arrangement has been reached with the Swedish PTS, although the issue has recently been re-opened by us.

    6. Comments and conclusions in the context of DSI III

    At the present time, EISCAT channels all Western European scientific use of incoherent scatter radars. It is likely that this scenario will remain unchanged in the short term, or at least until the year 2008. Our comments to the DSI III are therefore mainly concerned with safeguarding the operation of our UHF installations.

    The EISCAT UHF system operates in a band which in the harmonised European frequency plan is designated exclusively for fixed-to-mobile services, i.e. NMT900 (which is being phased out) and GSM. However, in the RR, the band 890 - 942 MHz is allocated to radiolocation on a secondary basis in Region I, thereby formally justifying the EISCAT UHF assignment. Before the mobile services were introduced, successful UHF observations were regularly made up to 940 MHz, but the congestion above 935 MHz is now so massive, that very little useful EISCAT work can be carried out there. On the other hand, the 916.5 - 935 MHz range is still useable at all sites, albeit with occasional in-band interference from, among others, the military. Provided that we are not forced to share this spectrum with any other users, we can live with this situation, even though it artificially limits our ability to retrieve all the information the radar could generate. Also, because of the high cost and complexity associated with any major frequency change, EISCAT prefers a status quo ante to any other scenario.

    Unfortunately, the actual situation is different. Our core 931.5 +/- 4 MHz frequency assignment is now under potential threat. Following the completion of the DSI Phase II, the ERC passed a Decision, no. (97)02, designating the 925 - 935 MHz range as a common European GSM extension allocation. Finland and Norway have already committed themselves to the terms of this Decision. We have noted, however, that there is currently no type approved GSM equipment available for this frequency band. Instead, industry seems to put all its efforts into developing "dual-band" terminals for GSM and GSM1800.

    EISCAT hopes, therefore, that business pressure will prevent the GSM extension band from being occupied by the GSM service in the next six to eight years, which by extrapolation probably means forever.

    EISCAT strongly advocates that all future expansion of the GSM service should take place in the 1800 MHz bands, and that third generation systems (UMTS) should be introduced as soon as practical.

    If however the 925 - 935 MHz band must be taken in use by GSM anyway, EISCAT will need to act on the national level. We will then ask the Finnish, Norwegian and Swedish administrations to delay the use of 925 - 935 MHz as much as possible in the affected areas of the three Nordic countries, or to avoid it altogether. A reasonable starting point would be to ask the administrations to set up co-ordination zones in the Tromsö and Kiruna areas, similar to the one already established in around Sodankylä.

    Military systems in the 930 MHz region represent another potential risk to the EISCAT operation. The band is used in the Sodankylä area by the Finnish military, and several cases of severe, even blocking, interference have been registered. We also note with concern that the section on NATO requirements in the DSI Phase III Workshop document proposes to harmonise the use of 915 - 918 and/or 925 - 929 MHz, Europe-wide, for so-called CJTF tactical radio relay links. If the proposal is accepted, military systems operating in these bands may start to show up also in Norway and/or Sweden in the near future.

    EISCAT proposes that problems in this context could best be avoided by the observation of co-ordination zones around the EISCAT UHF sites also by the defence forces of the three Nordic countries. It is unclear to us if the frequency management authorities can enforce such co-ordination, so we would appreciate to be kept informed of any developments in this respect.

    7. Proposal for a co-ordinated ISR frequency band at 915 - 920 MHz

    Considering the recent accession of Japan to EISCAT and the potential importance of ISR measurements of the mesosphere and stratosphere for environmental research, it is likely that scientific ISR operation will continue in Europe for the foreseeable future, and that EISCAT will constitute the backbone of this activity.

    We realise that it may be impossible to retain the present EISCAT frequency assignments in the period beyond 2008. To make the best possible continued use of existing investments it is highly desirable to obtain a permanent, 5 MHz wide, European primary frequency allocation for ISR systems somewhere in the 915 - 925 MHz band. The draft table of European allocations that served as a baseline for the DSI part II lists 915-920 MHz as allocated to ISM and MetAids. In the EISCAT response to DSI II, it was proposed that also incoherent scatter radars could use this band. We now reiterate this proposal.

    While a 5 MHz allocation is not wide enough to allow full plasma line reception, at least it would be sufficient to continue the "core" ion line based ISR work effectively. At 915 MHz, the existing EISCAT antennas, transmission line systems and receivers could be used with relatively minor adjustments and only the transmitter would require extensive redesign. Such an allocation would also simplify the problem of getting formal protection for the remote receiving sites in Sweden and Finland.

    If this proposal were implemented, it would give formal recognition to the existence and significance of ISR systems and open up the possibility for establishing new ISR facilities elsewhere in Europe, on much firmer grounds than has so far been the case.

    Appendix:

    URSI Resolution U. 22 (1990):
    Incoherent Scatter and MST Radars

    The URSI Council,

    considering

    (a) that IS (incoherent scatter) and MST (mesosphere-stratosphere-troposphere) radars are valuable tools for studying the atmosphere;
    (b) that due to the weak scattering processes used, the radars must transmit high powers on high gain antennas and detect small returns in a high noise environment;

    noting that increasing spectrum congestion, resulting from the expansion and use of radio services, may cause harmful interference to the radars;

    resolves

    1. to bring this problem to the attention of the International Telecommunications Union (ITU) and its technical bodies, the International Frequency Registration Board (IFRB) and the International Radio Consultative Committee (CCIR) and, in particular, to the attention of the national telecommunications authorities, through URSI Member Committees;

    2. to urge them to take reasonable measures to minimise interference problems at IS and MST radar sites;

    expresses the hope that frequency allocations will continue and will be issued to these radars on an interference-free basis and that, if possible, interference protection zones will be established around these radar facilities.

    Last modified: August 30, 1999