Figure 1.- Frequency bands in the free-space microwave window, allocated to certain transmitting Radio Astronomy Surveys Radio astronomers have carried out few surveys of the sky to what are believed to be flux levels of interest here. It is even possible that one or more of the "point" sources appearing in a survey in one frequency band but not in another survey in a different band, could be an unexpectedly powerful artificial signal. Radio astronomical surveys of large areas of the sky can be divided into two classes: broadband continuum surveys (1 MHz to 5 GHz bandwidth), and spectral line surveys (102 to 106 Hz bandwidth). Large bandwidths, absence of adequate spectral resolution, customary on-line instrumental data compaction, and absence of specific interest in the existence of extraterrestrial artifacts, all have reduced enormously the chances of recognizing an artificial extraterrestrial signal. In fact, a signal so strong as to be noticed, generally would have been ascribed to man-made radio frequency interference (RFI). To illustrate these points, we describe briefly two surveys reaching low flux levels by radio astronomical standards. The Parkes 2700-MHz source survey of the sky visible from Australia employed a bandwidth of 200 MHz and a sensitivity such that a signal flux level of about 1020 W/m2 would have produced unity signal-to-noise ratio (SNR = 1). No procedures were in effect to recognize a narrow band signal at this or much higher levels, unless the absence of this "radio source" on other surveys at adjacent frequencies caught an astronomer's interest and the source was then reexamined with suitable care. Westerhout used the NRAO meridian transit 300-ft antenna to survey a region along the Galactic equator 225° in longitude by 4° in latitude. Frequency resolution was 9.5 kHz in a band about 1 MHz wide. Unity SNR corresponded to about 3X1023 W/m2, but automatic data handling algorithms and RFI subtraction procedures would generally have prevented output registration of a coherent signal orders of magnitude stronger. The radio astronomy situation (and parallel situations in other areas) can be summarized this way. Procedures in use in high data volume observations tend to discriminate against the discovery of unexpected phenomena. Scientists generally do not explore a range of phenomena merely because it is not forbidden by any known laws. They usually have more immediate objectives in mind. The polarization of starlight was observed half a century after the work of Maxwell and Hertz; and the discovery of pulsars was a fortuitous accident due to the presence and curiosity of an astute graduate student in a favorable observational situation. Analog pen and paper recording of data was in use. For a third and final example, proof of the polarization of the diffuse background radio radiations was delayed at least 5 years by an unwillingness to test for polarization just because it was an unexplored degree of freedom in nature. Signal Classes Postulating the characteristics of signals we might be able to detect from another species has enlivened many a casual moment. We have only our experience over the past 30 years or so, our electromagnetic technology in which we take some pride, and our projections of how in the near future we might exploit this technology to suit our manifold desires. Above some tens of megahertz most of the power from all our surface transmissions is dissipated in the endless reaches of outer space. For over a quarter century and with increasing intensity, we have been generating an expanding aura of fairly powerful signals about the Earth, one we could detect if we had a sufficiently sensitive radio telescope situated 20 light years from Earth. As has been suggested many times, this may be a transient phase in our development, one that may be over in a time period microscopic on the cosmic scale. Nevertheless, a reversal of the historical trend toward higher transmitter powers is not yet in evidence. It does seem likely that some of our present transmissions will vanish from our scene in favor of more efficient, more directive procedures; cable and low power satellite transmissions may replace high power TV broadcasting, for instance. The future of high power radar is perhaps less clear. Table 1 is a fairly complete tabulation of published current high power radar signals in the 1 GHz to 2.5 GHz frequency range. They generally sweep out the sky continuously in regular patterns both in time and in direction. One class of signal for which we might search is the intraspecies type of transmission. We may be fortunate in having a relatively close neighbor on whom we can eavesdrop. From our own experience, this would require not only proximity or exceptionally great sensitivity because of the modest power levels we may perhaps expect, but also a complex pattern recognition capability in the time, frequency, and modulation domains. Most of our transmissions have a strong carrier component. If this holds with our nearest neighbor (and it may because it is a simple way to This table is a summary of published radar installations. Equivalent Isotropic Radiated Power in watts. CUHF color TV (460-890 MHz provides many carriers in the 105-106 W range. They are radiated with high stability and the toroidal dSee table 3 in Section III-2 for further details. maintain coherence), it eases the pattern recognition problem since stable, monochromatic signals are relatively easy to detect. Finally, we are in total ignorance of their spectrum utilization and have little idea even how they might utilize transmitters in "space activities." We can only assume that bandwidth and propagation requirements dictate some frequency allocation scheme there as here. We have, so far, no reason to search any frequency band outside the low noise, free-space microwave window; and since signal stability so directly affects our achievable carrier search sensitivity, the water hole seems a likely place to start an eavesdropping search. Eavesdropping can be classified as a non-cooperative search situation. Alternatively, "they" may be transmitting in a cooperative, beacon mode, carefully arranged to assist discovery by others, particularly newcomers like ourselves. Any concern on their part for range would surely highlight the advantages of the low end of the free-space microwave window. Since we have conceived it and it is feasible for us today, it is quite possible (but who knows how likely) that other species are curious as are we, and indulge in their allotted share of an obvious cooperative search strategy. Beacons could exist in many modes and for many differing applications. All who study extraterrestrial phenomena should be alert to the possibilities, regardless of the portion of spectrum of immediate interest to them. It is worth noting that beacons can be of any power, isotropic or beamed, and continuous or repetitive. If I were asked how to construct a beacon (to announce our presence) to be built on a crash basis by the year 2000, I would suggest these chief properties: 1 GW continuous radiated, power, an isotropic radiation pattern, a frequency in the water hole some megahertz above 1421 MHz, frequency stability to 1014 or better for the circularly polarized carrier, and modulation by polarization reversal in three modes: (1) carrier alone, (2) a bit every few seconds of binary, "acryptic" information to assist first decoding, (3) large information transfer at a bandwidth up to perhaps 104 Hz, and the transmitter and all that it requires in Earth's solar orbit on the other side of the Sun in order to provide a low Doppler drift rate and to minimize pollution of the local terrestrial electromagnetic environment. Modulation modes (1) and (2) would be present over 90 percent of the time in order to assist first detection. Such a beacon installation would be at about the limit of our own technology, and it would be detectable by a system equivalent to a modest Cyclops at a distance of 1000 light years or more. Signal Propagation Paths The interstellar medium, the interplanetary medium, and the Earth's atmosphere and ionosphere can all affect the coherence of signals passing through them. The physical processes, dispersion, scatter, and multipath transmission, are well understood theoretically, but our observational knowledge leaves much yet to be learned. In all these media there are systematic and turbulent motions of matter and free-electron density and magnetic field intensity which all show variations in direction, distance, and time. In the lower atmosphere, corresponding variations in the water vapor density add their contributions to the total possible coherence loss. Present information suggests that in the water hole and over distances as great as 1000 light years, the coherence of interstellar intelligent signals may suffer an appreciable loss unless the initial bandwidth (B) is limited to about the range 103 Hz <B<106 Hz These limits are uncertain by perhaps a factor of 10 and, since bandwidth is an important search dimension, point source scintillation and pulsar pulse-shape studies should be encouraged. Radio Frequency Interference (RFI) Solving the RFI problem is crucial to SETI. Section III-8 discusses RFI extensively and Section III-9 presents the unanimous Science Workshop resolution on the matter. Here, we mention briefly only certain salient points. A SETI system, regardless of antenna area, requires protection from man-made radiations down to a level which is 50 to 100 dB more stringent than the usual communication system requirements (see Section III-8). The protective measures required depend on the location of the SETI system and upon the frequency band or bands being searched. Figure 1 shows that most of the terrestrial microwave window below 22 GHz has been allocated to services such as radar and satellite communications, and therefore is generally difficult for SETI. It is fortuitous that, so far, the 1400 to 1727 MHz "water hole" band is used chiefly by a multiplicity of low and very low power services with which, in the main, SETI is compatible. SETI does not require an exclusive frequency band allocation. Only a moderate degree of worldwide and local cooperation is needed in order to preserve the terrestrial water hole for SETI. We should: 1. Choose a site for SETI over the horizon from centers of high population density, and out from under heavily used commercial aviation lanes. 2. Avoid the use of water hole frequencies by even low power services within some 100 to 200 km of the SETI site. 3. Obtain worldwide agreement to keep the water hole band significantly free of interference at the one or more SETI sites agreed upon. 4. Allow the present, mild use of the band by satellite services to come to a natural, perhaps hastened, termination over the next 5 or 10 years. Item 3 involves more than just setting limits on EIRP. All transmitters radiate some power into adjacent bands and into harmonic bands. The official national and international standards on spurious emissions are old and well behind the state of the art; and the actual situation is often worse because of the effort required to challenge effectively operations believed to be below even the standard requirements. Some civil and federal communications groups are already trying to improve both the standards and the practice to levels much closer (at least several powers of ten) to the knee in the performance/cost curve of current first-rate technology. To summarize the Earth-based search site situation, only relatively mild allocation problems are foreseeable with respect to sharing frequency allocations with many kinds of Earth-based transmitters. Searches cannot compete with line of sight satellite down transmissions. Any overall search strategy should contain an element that proposes actively to support other groups trying to bring transmission practice closer to that permitted by the state of the art. If adequate RFI protection for Earth-based search cannot be provided, it will be necessary to develop space search systems shielded from Earth. Space-based search systems are without the natural and effective shielding properties of the Earth. In descending order of estimated relative cost, we list below four desirable and feasible (before the year 2000) off-Earth sites for long-term search systems - long-term because much useful initial search and research could be carried out on Earth and in low Earth orbit while |