Contents lists available at ScienceDirect

Acta Astronautica

Acta Astronautica 68 (2011) 362–365

0094-57

doi:10.1

� Tel.

E-m

journal homepage: www.elsevier.com/locate/actaastro

Short-pulse SETI

Seth Shostak �

SETI Institute, 515 N. Whisman Road, Mountain View, CA 94043, USA

a r t i c l e i n f o

Article history:

Received 27 February 2009

Accepted 21 September 2009Available online 17 October 2009

Keywords:

SETI

Strategies

Short-pulse

Pulse

Optical

65/$ - see front matter & 2009 Elsevier Ltd. A

016/j.actaastro.2009.09.022

: þ1650 960 4530; fax: þ1650 9617099.

ail address: seth@seti.org

a b s t r a c t

While most optical SETI experiments are configured to detect nanosecond pulses, the

majority of their counterpart radio searches integrate for seconds to minutes, looking for

unchanging narrow-band carriers or slowly pulsed modulation. The former approach is

suggested as an effective way to stand out against stellar photon noise, while the latter

approach is dictated by the dispersive effects of the interstellar medium as well as the

high visibility of narrow-band signal components.

In this paper, we consider effective signal strategies for those that produce, rather

than simply search for, optical and radio beacons—signals that are designed to elicit

responses from technological civilizations. By considering the communication problem

from the point of view of the transmitters, rather than the receivers, we deduce some

likely signal characteristics for beacons, and concommitant new strategies for SETI.

& 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Current SETI experiments, both radio and optical, sharea common characteristic of observing strategy: theyexpend short dwell time (sometimes called ‘‘integrationtime’’) on targets. This is inevitable for sky surveyexperiments, which sweep entire hemispheres or more,but it’s also the case for targeted searches. Typical dwelltimes in the radio are 1–2 s for surveys, e.g. SERENDIP [1],and 100–200 s for targeted searches, e.g. Project Phoenix[2]. Optical SETI dwell times are Z48 s for the Harvard skysurvey [3], and 600 s for the Lick Observatory targetedsearch of nearby stars [4].

In order for such efforts to have any chance at detectingextraterrestrial intelligence, the transmitted signal must bepersistent. This is, of course, largely due to the fact thattargets (either stellar, or simply patches of sky examined bysurveys) are looked at only a few times at best; in the case ofthe targeted searches, usually only once. This is the basis forthe assumption that detectable societies must send signalsour way for thousands of years or more. Aside from this

ll rights reserved.

onerous synchronicity requirement, there’s also the fact thatany candidate signal would only be considered a detection ifit could be confirmed, and the current time required forconfirmation is at least days, and in the case of the searchbeing made using the SETI home screen saver, years.

The assumption of a persistent signal is a neccessity,and is mandated by our own receiving strategies. How-ever, would persistance be integral to an alien society’stransmission effort? If not, could we accommodate ourSETI experiments to still offer us a reasonable chance for adetection? In this paper we consider signaling strategiesfrom the point of view of the sender, and the consequentimplications for SETI.

2. Why would they transmit?

We assume in this discussion that the sender istransmitting a deliberate signal. As many authors havenoted [5], the emission of high-power leakage radiationmight be only a transitory phenomenon for any techno-logical society, and therefore an improbable find(although incidental high-powered transmitters, such asradars that are used to determine the position and rangeof long-period comets, might weaken this assumption).

S. Shostak / Acta Astronautica 68 (2011) 362–365 363

The possible motives for transmitting deliberatesignals are manifold, and obviously we cannot know forsure why a society would instigate a broadcast. In thispaper, we will presume that their rationale is to establishcontact with other societies, although given interstellarlight travel times this might not be with the intentionof establishing two-way communication. Indeed, sincetransmissions might, simply as a practical matter, be one-way only, we can reasonably surmise that easily detect-able signals would either be information-rich themselves,or be ‘‘pointers’’ to other signals that are.

There are two obvious incentives for a society toinitiate deliberate transmissions.

(1) Signals have been received from some other world.It’s possible, of course, that Earth might be a new-foundobject of attention for some society that, despite the odds,has noticed our leakage radiation, and is undertaking adirected response. This suggests one straightforwardstrategy for our SETI projects: observe all stars that areno more than half as far as our signals have reached. Thisis of order 30 light-years. Unfortunately, the number ofstars within this radius is only �103, and it would behighly fortuitous if such a small sample were to contain atechnically competent society. As example, if the numberof transmitting galactic sites is N=10,000 (a strawmannumber we will use throughout this paper), then theaverage separation between any two societies is �1000light-years. This suggests that seeking replies from thosewho have noticed us will make sense beginning twomillennia from now.

(2) The desire to ‘‘fish’’ for new technological societiesby transmitting to worlds known to have life. Within adecade, telescopes such as the Terrestrial PlanetFinder and Darwin will allow us to both image, and insome cases spectrally analyze, the atmospheres of extra-solar planets. Presumably then, a society that is but a fewcenturies ahead of ours will have at their disposal—assuming life is commonplace—long lists of ‘‘bio-planets’’.Since leakage might be an improbable find, or may notyet have reached them, it’s conceivable that such asociety might undertake a fishing expedition, pingingplanets showing evidence of life on the assumptionthat some fraction of them will have produced intelligentlife. The problem here is that the list of bio-planet targetsmay have to be quite long in order for the senders tohave a reasonable chance of pinging a planet withtechnically sophisticated inhabitants. Earth’s atmospherehas been enriched with oxygen for �2 billion years;more than long enough for our status as a world with lifeto be known by any culture in the Galaxy (and far beyond)that wields the instrumentation necessary to find it. Atechnologically sophisticated society able to receiveradio signals from space has existed on Earth for only60 years. If we assume that technological societies such asour own have a lifetime of L years, then the fraction oftime during which a bio-world similar to Earth couldrecognize a fishing signal is roughly L/2�109. For L

�10,000 years, this would mean that a fishing experi-ment would have to ping �2�105 bio-planets in order tohave a reasonable probability of reaching anyone. Andnote that implicit in this computation is the optimistic

assumption that every planet with life eventually pro-duces intelligence.

3. Transmission strategies

It seems overwhelmingly likely that scenario (2) aboveis the most germane to our current SETI effort; we shouldexpect to be targets of fishing experiments. However, fromthe transmitter’s point of view, a successful effort todeliberately send information to unknown beings willrequire pinging large numbers of star systems, asdescribed. Doing this sequentially could be daunting. Forexample, if each world requires a one week transmission(which, as we’ve noted, would suit the confirmation timerequired for many contemporary SETI experiments), and2�105 worlds need to be pinged, then the repeat intervalis nearly 4000 years. Not only does this require consider-able patience on the part of the transmitting society, italso makes it extremely unlikely that the ping will ever beseen, and there will be virtually no chance of thishappening unless the receiving society is able to observeall-sky, all-the-time.

The obvious way to ameliorate this dismal prospect, atleast from the transmitter’s standpoint, is to eitherbroadcast to many (or all) star systems at once, or toshorten the length of the pings, thereby reducing therevisit time. We consider both strategies:

Omnidirectional beacons, either broadcasting to theentire Galaxy, or at least to the disk plane, would clearlyobviate the necessity of knowing which worlds had life, orfor that matter anything else about possible targets. Theproblem with this ‘‘lighthouse’’ approach is the enormousenergy requirement. Project Phoenix, which used theArecibo 305 m antenna and integration times of �200 s,had a flux density sensitivity of �10�25 W/Hz in a 1 Hzchannel, and would have been able to detect signals withan EIRP of �1014 W at 1000 light-years. Unless Nb10,000,a beacon only capable of reaching societies at this radius isnot particularly interesting. The desired reach for arandomly situated galactic beacon able to signal anywherein the Milky Way would be 70,000 light-years, necessitat-ing an EIRP of 5�1017 watts for the sensitivity limitreached by Project Phoenix using the Arecibo antenna.This is tens of thousands of times greater than the totalenergy consumption of humanity today, and somewhatgreater than the terrestrial insolation. However, while thisis a truly prodigious energy requirement, one could arguethat it is not an inconceivable one for an advanced society(and, of course, the number is predicated on our currentreceiver sensitivity limits, not those that might obtain forus in the future). These considerations should caution usagainst dismissing the possibility of such beacons.

However, irrespective of the actual power level of anomnidirectional beacon, a worthwhile reduction in theenergy demand, amounting to several orders of magni-tude, could be simply achieved by broadcasting shortpulses, with relatively long interpulse periods. This worksbecause the receiver sensitivity scales with t1/2, where t isthe integration time, but the power requirement forthe transmitter scales linearly with t. Therefore if instead

Fig. 1. The diameter of the primary mirror or lens for optical telescopes

constructed over the last century, including large-aperture instruments

planned for the coming few decades. The doubling time is �20 years.

S. Shostak / Acta Astronautica 68 (2011) 362–365364

of a continuously ‘‘on’’ beacon, the sender opts for aone-second flash with a repeat interval of 3 h, then theaverage power requirement for the beacon to achieve thesame detectability at the receiving end drops by twoorders of magnitude. Shorter flashes simply increase thispower saving, although scattering may limit the pulselengths to Z1ms [6].

Infrequently spaced one-second flashes might seemuseless, as they would, by themselves, convey littleinformation. The flashes could, of course, be comprisedof short pulse trains, with coded information, althoughthis introduces a spreading of the signal over more of theband, and the shortness of the pulses will be limited bythe dispersive effects of the interstellar medium. But evensingle, unmodulated pulses, simply by their presence,define a location on the sky and a frequency on the dial.Consequently, they could serve as the highly visiblehailing component of a two-tier transmitting setup. Anysociety finding this flashing beacon would surely searchlong and hard for a weaker, modulated component: themessage. A two-tier transmitting strategy—composed of ahighly visible, short-duty-cycle pinging signal, and a lowerpowered, persistent message—would simultaneouslymaximize detectability while reducing the energy costsof sending information.

The second approach by which a sender might dealeffectively with the large number of targets inherent inany fishing experiment is to abandon an omnidirectionalbeacon in favor of sequentially targeted transmissions. Thesender could illuminate bio-planets serially with pings atshort repeat intervals. Again, the practicality of using veryshort pings or pulse trains will be limited by theinterstellar medium [7], but a sequential targeting schemeusing millisecond pulses and addressing 2�105 stellarsystems would allow repeat intervals as short as 30 min.Once again, although the individual pings are severelylimited in the amount of information they can convey,they could serve as a pointer to a second, weaker signalwith message content. Of course, we have assumed thatthe detection system is able to detect at least two pings,for otherwise the signal would not be consideredsignificant.

Note that, given the right transmitting hardware, atargeted fishing experiment can operate at enormouslylower energy levels than the omindirectional beacon eventhough, unlike the beacon, it would be continuously ‘‘on’’.As a concrete example, imagine a phased array ofdimensions �106 wavelengths (assumed to compriseenough elements to synthesize a beam without significantsidelobes), used to target stellar systems from 1000 to30,000 light-years distance. The size of the beam wouldrange in diameter from 70 to 2000 AU, thereby easilyencompassing all of the presumedly interesting regionsabout these stars. The required power to ensure anaverage received flux density of S is simply given by

P ¼ Sd2ðD=lÞ2

where d is the distance and D/l is the array dimensionin wavelengths. If we stipulate a terrestrial detectionthreshold (e.g., Project Phoenix at Arecibo) of Smin¼10�24

W/Hz (assuming a single, one-second pulse. Note that this

is somewhat worse than the continuous sensitivity quotedabove.), then this could be achieved with 102oPo105 W.These powers are so low, one could easily envision thatfishing arrays would synthesize multiple pixels on the skyand ping many of the targeted bio-planets simultaneously.This would shorten the revisit times. Of course, it wouldalso require a power increase in proportion to the numberof simultaneous targets, but at kilowatts per target, thisdoesn’t seem to be a severe limitation. We’ve remarkedthat an omindirectional beacon capable of pinging theGalaxy at a level Smin=10�24 W/Hz with 1 s pings at 3 hintervals would require 51015 W, on average. At this sameenergy level, a focussed array could ping every star in theGalaxy with a 10% duty cycle (e.g., one 1 s pulse every10 s). If the sample of bio-world targets numbers in thehundreds of thousands, or even millions, the arrayapproach is enormously preferable to the all-sky beacon.

Our example uses D/l=106, or 200 km at 1420 MHz.While this is two orders of magnitude greater than themaximum dimension of the planned Allen TelescopeArray, the historic growth in the size of astronomicalinstrumentation is dramatic and, in the case of opticaltelescopes during the preceding century, exponential(Fig. 1). There can be little doubt that an imaging arrayhundreds of kilometers in size would be within thecapabilities of a society only slightly more advancedthen our own, and in any event the dimensions couldalways be lessened by transmitting at higher frequencies.

4. Implications for SETI experiments

By considering the strategies of sending societies,we’ve noted that, unless we are within 30 light-years ofa technically sophisticated culture, deliberate transmis-sions to us will likely be fishing expeditions, and ourworld will be one of many being illuminated. The sendersmight do this indiscriminantly using an omnidirectionalbeacon, although in order to reduce the enormous powerrequirements of such a device, the transmissions willbe short (r1 s) and sent at intervals that might beminutes or hours.

A far less energy intensive scheme envisions that thesenders ping selected targets (bio-worlds) either serially

S. Shostak / Acta Astronautica 68 (2011) 362–365 365

or to some degree in parallel, using arrays to confine theirtransmissions to the inner regions of the stellar systemsbeing targeted. While it is conceivable that multiple arrayscould cover every target of interest continuously, therewould again be a strong temptation to take advantage ofthe energy savings to be made with a two-tier strategy:attention-getting, short bright pings, and persistent, low-level messages. This assumes that once a ping hasbeen detected, and a location on the sky and a frequencyon the band have been pinpointed, a receivingsociety would expend the effort to construct a device thatwould eventually find a lower power, continuously-onmessage.

5. Conclusions

Today’s radio SETI strategies are frequently predica-ted on the existence of persistent (days to weeks) CWcarriers that can be found by integrating observationsover minutes of time and confirmed over the courseof days. A better strategy, in light of the inevitablebenefit to the sender of short attention-getting transmis-sions, is to emphasize the search for pinging pulses whichmight repeat on intervals of minutes or hours. These pings

could be the visible component of a two-tier fishingstrategy, composed of simple, highly visible intermittentsignals pointing to an information-rich transmission atlower signal levels. SETI might well benefit from morepatience on targets, and less time averaging.

References

[1] D. Werthimer, S. Bowyer, J. Cobb, M. Ledofsky, M. Lampton, TheSERENDIP IV Arecibo sky survey, In: G.L. Lemarchand,K. Meech (Eds.), Bioastronomy 99, A New Era in Bioastronomy, ASPConference Series, vol. 213, 2000, p. 479.

[2] Project Phoenix, /http://www.seti.org/site/pp.asp?c=ktJ2J9MMIsE&b=179177S.

[3] A. Howard, P. Horowitz, C. Coldwell, An all-sky optical SETI survey,/http://seti.harvard.edu/oseti/allsky.pdfS.

[4] Lick Observatory Optical SETI Program, /http://seti.ucolick.org/optical/oseti_observ.htmS.

[5] R.D. Ekers, D.K. Cullers, J. Billingham, L. Scheffer (Eds.), 2002, SETI2020: A Roadmap for the Search for Extraterrestrial Intelligence, SETIPress, Mountain View, CA.

[6] J.H. Taylor, J.M. Cordes, ApJ 411 (1993) 674.[7] S.G. Shostak, 1995, SETI at wider bandwidths?, In: G. Seth Shostak

(Ed.), Progress in the Search for Extraterrestrial Life, ASP ConferenceSeries 74, Astronom. Social Society of the Pacific, San Francisco,p. 447.