Detection and Discovery of Near-Earth Asteroids by the LINEAR … · 2019-02-15 · vere local...

• EVANS, SHELLY, AND STOKESDetection and Discovery of Near-Earth Asteroids by the LINEAR Program

VOLUME 14, NUMBER 2, 2003 LINCOLN LABORATORY JOURNAL 199

Detection and Discovery ofNear-Earth Asteroids by theLINEAR ProgramJenifer B. Evans, Frank C. Shelly, and Grant H. Stokes

■ The Lincoln Near-Earth Asteroid Research (LINEAR) program, whichapplies space surveillance technology developed for the U.S. Air Force todiscovering asteroids, has been operating for five years. During that timeLINEAR has provided almost 65% of the worldwide discovery stream and hasnow discovered 50% of all known asteroids, including near-earth asteroidswhose orbital parameters could allow them to pass close to the earth. Inaddition, LINEAR has become the leading ground-based discoverer of comets,with more than one hundred comets now named “LINEAR.” Generally,LINEAR discovers comets when they are far away from the sun on theirinbound trajectory, thus allowing observation of the heating process commonlymissed previously when comets were discovered closer to the sun. This articleprovides an update to recent enhancements of the LINEAR system, details theproductivity of the program, and highlights some of the more interesting objectsdiscovered.

A in effortsto find and catalog asteroids for the pasttwo hundred years. Initially the searches were

inspired by scientific curiosity and a desire to under-stand our solar system. More recently, however, thesesearches have also been motivated by the desire to un-derstand—and possibly react to—the threat of a po-tential collision between the earth and certain aster-oids near the earth’s orbit.

The search for a missing planet believed to be lo-cated between the orbits of Mars and Jupiter led tothe discovery of the first asteroid, Ceres, by GiuseppePiazzi in 1801. Ceres, like the vast majority of aster-oids, populates the main asteroid belt between Marsand Jupiter. The material spread across the main beltis really a planet that never coalesced properly becauseof the disruptive effect of Jupiter’s gravity. Asteroidsin the main belt represent no direct collision threat tothe earth. Over millions of years, however, asteroids

in the main belt can be perturbed into orbits that cancome close to that of the earth [1, 2]. These asteroidsare called near-earth asteroids (NEA). The firstknown NEA, designated (433) Eros, was discoveredin 1898, nearly a century after the discovery of Ceres.

Ceres was discovered through the laborious processof making direct visual observations with a telescopein order to create a hand-drawn star chart. After ob-servations were made over a period of days, Ceres wasfound to move in the sky, leading to the conclusionthat it was a planet-like object in orbit around thesun. The visual observation process employed to dis-cover Ceres ultimately led to the discovery of the firstfew hundred asteroids in the nineteenth century. Thisoriginal method of discovering asteroids by differenti-ating their motion from that of stars has not changedsignificantly since 1801, but the technology used forthese observations has progressed, especially in thelast decade. The visual observation process was used


200 LINCOLN LABORATORY JOURNAL VOLUME 14, NUMBER 2, 2003

until the late 1800s, when more precise and capablephotographic methods were introduced, which led tothe discovery of the asteroid known as (323) Bruciaby Max Wolf in 1891. Film was used to record a pairof images of the same area of the sky separated by acouple of hours. The images were typically blinkedback and forth later to allow an analyst to see the mo-tion that distinguished the asteroid from the back-ground stars. The advent of photographic technologyled to an explosion of discoveries of both main-beltasteroids and NEAs.

Photographic search techniques employed duringthe era from 1890 to 1990 obviously had importantadvantages over direct visual observation techniques.Large areas of the sky were photographed with muchless observer fatigue, and a permanent record of theobservations was maintained. Photographic searchsystems, however, had two limitations that were notovercome until the introduction of solid state detec-tor technology of the 1990s. First, the sensitivity ofphotographic film was poor, with a solar-weightedquantum efficiency of approximately 1%. Second,the process of detecting moving objects relied on theclose and careful attention of scientists and techni-cians who were subject to fatigue and nonuniformperformance.

The conversion to solid state charge-coupled-de-vice (CCD)–based detection systems and computerprocessing of the data was first demonstrated by theSpacewatch search system in 1984 [3]. At the time,the readout rate of astronomical-quality CCD detec-tors was very slow (about twenty-five kilopixels persecond for optimal sensitivity), and the computationcapability that could be applied to process the datawas quite limited by today’s standards. T. Gehrels andR. McMillan developed a search system that accom-modated the available technology in a clever way.They developed a drift-scan technique in which thereadout rate was clocked to the sidereal drift rateacross the CCD. As a result, the long scans producedimages with minimal pixel-to-pixel (flat field) varia-tion, and readout overhead was limited to ramp-upand ramp-down fields at either end. The data rate waswell matched to the capabilities of the processingequipment at the time. The use of CCDs resulted insignificantly more sensitive search systems because of

their ability to achieve a solar-weighted quantum effi-ciencies as high as 65%, a result which is almost twoorders of magnitude greater than that of film. TheSpacewatch system achieved a faint limiting magni-tude because of the long integration time as a stardrifts across the CCD. Spacewatch tailored the sys-tem to the existing telescopes and available technol-ogy, and became the most productive asteroid searchsystem for over a decade.

Current state-of-the-art CCD-based detection sys-tems, with faster readout rates and greater computingcapacity, now use the step-stare approach, which wasfirst implemented in 1995 with the Near-Earth Aster-oid Tracking (NEAT) program [4], a joint project be-tween NASA’s Jet Propulsion Laboratory and the U.S.Air Force. A step-stare system works much like a tra-ditional film-based search, with multiple imagestaken over an hour or two. The NEAT system wasable to search more sky than Spacewatch by using aCCD array with four times as many pixels and with alarger angular pixel scale, but also by using shorter in-tegration times, which allows for greater sky coveragealbeit at decreased sensitivity and depth of search.

Asteroid Collision Effect

As indicated earlier, the primary motivation forsearching for asteroids has moved from scientific curi-osity to a desire to understand and possibly react tothe threat of a collision with an asteroid. As ourknowledge of the population of known asteroids con-tinues to increase, we gain a better understanding ofthe effect of collisions. A collision between the earthand an asteroid is essentially a problem in dissipatingkinetic energy. Asteroids with diameters smaller thanapproximately fifty meters typically dissipate their en-ergy in the atmosphere, although they can cause se-vere local damage from air blast effects (e.g., theTunguska collision in 1908, which flattened overeight hundred square miles of forest in Siberia, is be-lieved to have been caused by an asteroid that wasfifty to seventy-five meters in size [5, 6]).

Asteroids exceeding a hundred meters in diametercan reach the surface of the earth and cause consider-able regional damage in a collision; large asteroids—typically defined as those with diameters exceedingone kilometer—may cause devastating global effects



[7, 8]. Figure 1 shows images of impact craters inwestern Australia and in Quebec, Canada, made bytwo relatively small asteroid collisions. The globalreach of the devastation from large asteroids isthought to be caused by the large mass of materialsplashed from the impact site, which then spreadsthrough the atmosphere and falls around the globe.Damage from ocean impacts may be enhanced incoastal areas by the creation of tsunamis [9].

Inspired by the growing understanding of the as-teroid threat and the recent collision of comet Shoe-maker-Levy 9 with Jupiter, the U.S. Congress beganseriously considering the threat of an earth impact bycomets and NEAs. A report, called the SpaceguardSurvey, was generated by a group of well-known as-tronomers who assessed the risk to earth and maderecommendations as to how to best address the threat[10]. In 1998, Congress issued a mandate to NASA

requiring that 90% of the NEAs with diametersgreater than one kilometer be discovered and cata-loged by 2008.

At the time the Spaceguard report was written, thenumber of known NEAs was 450, of which 223 werelarge. The total number of large NEAs was estimatedto be near 2200. Because of the significant increase inasteroid discovery rates in the last five years, the aster-oid community is currently better able to assess therisk presented by NEAs. Today there are 2606 knownNEAs, of which 691 are large; the population ofNEAs with diameters exceeding one kilometer is cur-rently estimated to be 1090 ± 180 [11, 12]. Thepopulation of such objects with diameters between ahundred meters and one kilometer is estimated to bearound 56,000 [13]. Note that the more accuratepopulation estimates are derived primarily from theLincoln Laboratory search data described below.

Lincoln Near-Earth Asteroid Research Program

The Lincoln Near-Earth Asteroid Research program(LINEAR) began regular operations in March 1998,just as NASA formally embraced the task of catalog-ing 90% of the largest NEAs, and quickly became themost productive asteroid survey program in history.LINEAR is an outgrowth of space surveillance tech-nology developed by Lincoln Laboratory for the U.S.Air Force; searching large areas of the sky for faintmoving objects is common to developing a catalogueof earth’s orbiting satellites and a catalogue of aster-oids. Applying the highly refined Air Force space sur-veillance technology to the asteroid search task hasprovided an order-of-magnitude increase in capabil-ity to the worldwide asteroid search effort.

The LINEAR program and technology have beenreported previously, both in the Lincoln LaboratoryJournal [14] and in the open literature [15]. This ar-ticle updates these reports, with a focus on recent sys-tem developments and enhanced productivity.

The LINEAR System

Since the start of routine operations in March 1998,LINEAR has provided 65% of the worldwide discov-ery stream for NEAs, and has now discovered morethan 50% of the known population of these objects.LINEAR has accomplished this productivity by using

FIGURE 1. Examples of impact craters caused by asteroidcollisions with the earth. The upper image shows the WolfCreek crater in western Australia, a well-preserved craterpartly buried under windblown sand. This crater is 0.85 kilo-meters wide, and is estimated to be 300,000 years old. (Aerialphoto courtesy of V.L. Sharpton.) The lower image showsthe larger New Quebec crater in Quebec, Canada. This cra-ter is 3.4 kilometers wide, with a 250-meter-deep circularlake, and is estimated to be 1.4 million years old. (Imagecourtesy of George Burnside, Manotik, Ontario, Canada.)



two one-meter Ground-Based Electro-Optical DeepSpace Surveillance (GEODSS) telescopes, located atthe Lincoln Laboratory Experimental Test System(ETS) on the White Sands Missile Range nearSocorro, New Mexico. Figure 2 shows one of theseGEODSS telescopes, and Figure 3 shows ETS, whichis adjacent to the U.S. Air Force GEODSS site. Thetelescopes are equipped with Lincoln Laboratory–de-

FIGURE 2. One of two one-meter Ground-Based Electro-Optical Deep-Space Surveillance (GEODSS) search tele-scopes at the Lincoln Laboratory Experimental Test System(ETS) at White Sands Missile Range, New Mexico. This tele-scope, which previously was used for space surveillance, isnow a component of the Lincoln Near-Earth Asteroid Re-search (LINEAR) asteroid search program.

FIGURE 3. The ETS at White Sands Missile Range, near Socorro, New Mexico, adjacent to the U.S. Air Force GEODSS site.The LINEAR program operates two search telescopes, called L1 and L2, and one follow-up telescope, called L3.

L1 L2L3

U.S. Air Force GEODSS siteLincoln Laboratory ETS

veloped CCD focal planes, and utilize a step-stareprocess rather than a drift scan [16, 17].

The CCD focal plane, shown in Figure 4, containsan array of 2560 × 1960 pixels and has an intrinsicreadout noise of only a few electrons per pixel. TheCCDs are constructed with a back-illumination pro-cess, which provides peak quantum efficiency exceed-ing 95% and solar-weighted quantum efficiency of65%. A frame-transfer feature produces a quick im-age transfer time from imaging area into frame bufferof only several milliseconds, which allows fields to beacquired as fast as the telescope can step and settle.This advanced CCD, in combination with agilewide-field-of-view GEODSS-type telescopes, rapidprocessing capability, and sophisticated moving-ob-ject detection algorithms, forms a unique and power-ful asteroid search system that can survey essentiallythe entire available sky each month.

Figure 5 shows a diagram of the LINEAR systemas it exists at the ETS today. The core of the LINEARdetection system has not changed significantly in re-cent years, and thus is quickly summarized here. Fiveframes of data corresponding to the same part of thesky are collected by a LINEAR telescope at approxi-mately thirty-minute intervals (to give the asteroidmotion enough time to become apparent betweensuccessive frames). The frames are aligned and regis-tered, and then background noise is suppressed via afixed threshold. The resulting binary quantized dataare further processed for candidate streaks, and thevelocities and astrometric positions of these streaksare determined. Objects moving faster than 0.5° perday and objects with an unusual motion, as deter-mined by their rate and direction of motion as a func-



tion of their location in the sky, are flagged as poten-tial NEAs [18]; these are visually confirmed by ob-servers and are deemed interesting. All remaining ob-jects are more likely to be main-belt asteroids; theseare too numerous to be confirmed visually and aredeemed uninteresting.

After comparison to a star catalog for position and

magnitude information, all observations for both in-teresting and uninteresting moving objects are sent tothe Minor Planet Center (MPC) at the Harvard-Smithsonian Center for Astrophysics, in Cambridge,Massachusetts, where both sets of data are used for lo-cating potential NEAs or for updating known aster-oid positions. The MPC acts as the central repositorycharted by the International Astronomical Union tocollect and publish all observations of asteroids [19].In addition, the MPC maintains the catalogue ofknown minor planets (the formal name for asteroids),and issues formal notification of new discoveries. Thetwo object categories—interesting and uninterest-ing—allow the MPC to prioritize their data process-ing so that potential NEAs receive appropriate atten-tion in a timely manner. On a long dark winter night,LINEAR typically detects 10,000 to 12,000 movingobjects (mostly main-belt objects) and sends the re-sulting 50,000 to 60,000 observations (one for eachframe) to the MPC. Perhaps half of these objects rep-resent previously uncataloged bodies.

Recent Improvements

The performance of LINEAR has been considerablyimproved since 1998, when a previous article on

FIGURE 4. The charge-coupled device (CCD) developed byLincoln Laboratory is a large-format, back-illuminated, low-noise, high-quantum-efficiency, fast frame-transfer device.

FIGURE 5. The LINEAR detection and processing system at the ETS. Each lunation LINEAR searches the entire 15,000 to 18,000square degrees of unique sky available to it. The automated telescope and data acquisition system receives input fromschedulers, collects five frames of data, processes the data, and sends output detection lists to the Minor Planet Center (MPC)at the Harvard-Smithsonian Center for Astrophysics in Cambridge, Massachusetts, where the data are collected and cataloged.

Input data

Frame-frameregistration

Background suppression normalization

Clustering andstreak detection

Photometry andastrometrics

Detection lists sent to MPC

Searchscheduler

Follow-upscheduler

Mount andcamera control

Search telescope LINEAR control room



LINEAR appeared in the Lincoln Laboratory Journal[14]. Five system improvements are responsible formost of the enhancements in the number of detec-tions. These improvements, in chronological order,are (1) the unbinning of pixels read from the CCD,(2) the addition and continued improvement of asearch scheduling system, (3) the advent of a follow-up scheduler, (4) the addition of a second search tele-scope to the system, and (5) the development of morefinely tuned detection algorithms.

The first of these system changes occurred in June1999. Prior to that date, all data were collected in a2 × 2 binned mode in order to save a factor of four indata processing, even though this mode incurs a de-tection loss by mixing background noise from a largerarea of the sky with the signal. After a data-processinghardware upgrade, it became feasible to use anunbinned mode for data collection. The result of thischange is an improvement in sensitivity of almost onevisual magnitude.

The search scheduler and follow-up scheduler, in-dicated in Figure 5, were the second and third notableimprovements to LINEAR. The search scheduler al-gorithm translates a description of the area of sky tobe searched into a sequence of right-ascension anddeclination field center positions that can then be ac-quired in an automated manner by the telescope anddata-acquisition systems. The search area is describedin terms of degrees from opposition (i.e., the pointopposite from the sun), and northern and southernlimits. The pattern then automatically minimizestelescope step and settle times, and also efficiently or-ders the fields to account for the rotation of the earththroughout the course of a night.

The follow-up scheduler performs a similar task byusing the predicted right-ascension and declinationpositions of previously detected, unidentified, inter-esting moving objects and MPC-determined NEAcandidates as the input. Since LINEAR on a goodobserving night finds many more NEA candidatesthan can be posted by the MPC on the web (via theNEA confirmation page) for follow-up by others inthe astronomical community, LINEAR itself musttake some responsibility for following up the poten-tial NEA objects it discovers. An orbit predictor esti-mates the right-ascension and declination positions of

each follow-up object for the approximate time whenthe follow-up observing will be done. Next, the pre-dicted positions of the NEA candidates on the MPCconfirmation page are automatically downloaded.The interesting objects and NEA candidates are com-bined and follow-up observations of these positionsare scheduled, on the basis of their locations in thesky, to minimize telescope motion. An output file iscreated and fed directly into the telescope controlsoftware.

Objects from the follow-up scheduler take prece-dence over the search scheduler when search tele-scopes are employed for follow-up operations. His-torically, follow-up has been done by the LINEARsearch telescopes, which takes time away from searchoperations. Recently, a third telescope dedicated tofollow-up efforts has started operations for LINEAR,as discussed in the sidebar entitled “Search versus Fol-low-Up: The LINEAR 3 Story.”

Figure 6 shows a histogram of the amount of datacollected and the amount of sky searched from March1998 through September 2003. In May and October1999, the large relative increase in square degrees ofdata collected was due to tests of the second LINEARtelescope, called L2, which entered routine operationsin February 2000. This second LINEAR search tele-scope represents the fourth major system improve-ment in the search capability of LINEAR. Approxi-mately 600 to 700 fields of data are collected pernight (1200 to 1400 square degrees) with each tele-scope, with up to 45,000 square degrees of data perlunation. Since some portions of the sky near opposi-tion are covered more than once during a singlemonth, this search strategy yields about 15,000 to18,000 square degrees of unique sky visited duringeach month.

The fifth significant system enhancement is in theimage processing algorithms, which have improvedincrementally over the years. Some of the improve-ments are due to increased computing power, andothers are due to improved understanding of the sys-tem characteristics and noises. An example of a pro-cessor-driven modification is that the processor-in-tensive plate modeling is now performed on all fiveinput data frames, and not on just the first frame in afive-frame set. This processing upgrade allows for de-



FIGURE 6. The amount of data collected and the amount of sky searched from March 1998 throughSeptember 2003. Seasonal variations in the data are clearly apparent. The large relative increase indata collected in May and October 1999 was due to testing of the second LINEAR telescope, whichentered routine operations in February 2000.

tections in frame sets in which the first one or twoframes of the night may be unusable because ofclouds or arcs of light from the moon.

Improvements in sensitivity resulting from learnedsystem characteristics include establishing eight con-stant false-alarm-rate thresholds, corresponding tothe eight channels of the imager, and setting thesethresholds as a function of object velocity. The result-ing lower false-alarm rate allows the detection thresh-old to be lowered, thereby increasing overall systemsensitivity while maintaining the same probability offalse alarm.

LINEAR Observing Strategy

Much of the initial development effort dedicated toLINEAR was focused on constructing an automateddetection system that could cover a large amount ofsky with good sensitivity for detecting asteroids.Once that goal was largely accomplished, the focusshifted to developing a search pattern that resulted inthe best productivity over the period of a lunation(i.e., one cycle of the moon’s phase). When full, ornearly so, the moon is a large contributor to the skybackground noise during observations. During thebright phase of the moon, high-sensitivity observa-tions are not possible, and the best results are ob-tained far from the sky position of the moon.

Early in the monthly observing cycle, the moon is

0

5000

10,000

15,000

20,000

25,000

30,000

35,000

40,000

45,000

50,000

Mar 9

8

Jun 98

Sep 98

Dec 98

Mar 9

9

Jun 99

Sep 99

Dec 99

Mar 0

0

Jun 00

Sep 00

Dec 00

Mar 0

1

Jun 01

Sep 01

Dec 01

Mar 0

2

Jun 02

Sep 02

Dec 02

Mar 0

3

Jun 03

Sep 03

Sea

rch

area

(squ

are

deg)

Total search areaUnique search area

still bright during much of the night; images are thuscollected well away from the ecliptic to avoid themoon. As the month progresses toward new moon,the search shifts closer to the ecliptic, the part of thesky where the solar phase angle is optimal and aster-oids appear at their brightest.

Figure 7 illustrates the sky coverage typicallyachieved by LINEAR. Normally the best search expe-rience is during the fall and winter months, when thenights are long and the sky is clear. Figure 7(a) showstypical coverage during a fall or winter month. Theoval graph represents the entire sky as seen from theearth. Only about half of the sky is available forsearch in a given month; the rest is above the horizononly when the sun is up. During the spring and sum-mer time, shown in Figure 7(b), the nights are not aslong, the weather is less conducive to clear skies, andthe galactic plane (the Milky Way) is above the hori-zon. The Milky Way contains a much larger back-ground of stars, which increases the sky brightness,thus making it harder to detect asteroids. The areascontaining the Milky Way are shown by the darkercolors in Figure 7(b).

Figure 7(c) displays the composite coverage of theLINEAR system during the year 2002. These plotshave been scaled to show a good-weather, back-ground-corrected, single-frame-equivalent integra-tion time, with the lighter colors displaying increased



S E A R C H V E R S U S F O L L O W - U P :T H E L I N E A R 3 S T O R Y

, LINEAR observa-tions are sent to the Minor PlanetCenter (MPC) in Cambridge forprocessing, and a small fractionof the objects are identified forposting on the MPC confirma-tion page accessible via the web.The posted objects are thoughtto be likely candidates for near-earth asteroids (NEA), and aworldwide network of amateurand professional astronomerscollect observations to improvethese objects’ orbits. Because of alimited follow-up capacity, only afew dozen of the tens of thou-sands of objects detected by LIN-EAR each month are posted onthe confirmation page.

Therefore, as many as a hun-dred interesting objects, and per-haps thousands of main-belt ob-jects discovered by LINEAR in agood observing night, are left tochance follow-up. In addition,objects not posted to the confir-mation page require a secondnight’s observations linked to thefirst night’s to obtain discoverycredit in the form of provisionaldesignation. Thus LINEAR isleft with a quandary as to how tohandle the potentially interestingobjects it discovers. Some ofthese objects are unrecognizedNEAs, but without a secondnight’s data they are not creditedto LINEAR. On the other hand,using the unique LINEAR search

telescope to conduct directed fol-low-up is a poor use of these re-sources, since follow-up opera-tions reduce the time availablefor searching new sky.

In the past, a compromise wasreached whereby between 10%to 20% of LINEAR telescopetime was devoted to directed fol-low-up of potentially interestingobjects not initially posted to theconfirmation page. These opera-tions have been responsible for anumber of credited discoveries,and over time we have gottenbetter at choosing the objects forfollow-up, based on their motion

relative to main-belt objects. Fig-ure A shows an example of a scat-ter plot that is used to identifypotentially interesting objectsfrom thousands of main-belt de-tections. However, many poten-tially interesting objects are leftto serendipitous follow-up dur-ing standard search operations.

Over the past year, LINEARhas begun operations of a newtelescope system dedicated to au-tomated follow-up of objects dis-covered by the LINEAR searchtelescope, thus allowing a greaterpercentage of the potentially in-teresting objects to be addressed.

FIGURE A. A sample scatter plot showing the rate of motion versus angle ofmotion expected of various main-belt asteroid families as well as interestingout-of-boundary objects such as potential near-earth asteroids (NEA) that de-serve a second look. More than a dozen similar plots are used by LINEAR, witheach plot pertaining to a specific part of the sky (e.g., 5° ecliptic longitude).

0

30

60

90

120

150

180

0 0.20.1 0.3 0.4 0.5 0.6

Rate of motion (deg/day)

Ang

le o

f mot

ion

from

0° l

atitu

de (d

eg)

Main-belt asteroidsNEAsBoundary

5° ecliptic longitude



FIGURE B. The LINEAR 3 telescope and its dome at the Lincoln Laboratory–operated Experimental Test System in NewMexico. It is a Cassegrain telescope with a thirty-inch aperture and a 0.5 square-degree field of view. It can routinely de-tect objects with a visual magnitude of 19.0 with forty seconds of integration. (Photographs courtesy of Peter Trujillo.)

Figure B shows this telescope sys-tem, called LINEAR 3, or L3 forshort. It enhances the amount ofsky that LINEAR can search by10% to 20%, and captures sec-ond-night data on 100 to 200objects each night.

The L3 system consists of a30-inch telescope with a 1024 ×1024 Lincoln Laboratory CCDsensor in a camera system similarto the one in the Moron Optical

Space Surveillance satellite track-ing system. This camera was usedin the LINEAR pre-prototype in1996 and 1997. The L3 systemsearches the sky at one-fifth therate of the other LINEAR sys-tems, but it has a field of viewlarge enough to find multiple fol-low-up objects in a single field.

A month after L3’s inauguraloperations, the system proved itsworth by gathering second-night

data on two asteroids, both ofwhich were determined to beNEAs but were not initiallyposted to the confirmation page.In fact, both objects turned outto be potentially hazardous aster-oids, and one, 2001 NT7, pre-cipitated considerable press inter-est when an announcement wasmade (since retracted) of a poten-tial collision with the earth, pre-dicted for 1 February 2019.

LINEAR Search Results

Table 1 summarizes the productivity results of theLINEAR search system, showing the number of LIN-EAR observations from 1997 through 2003. The datareported include the number of observations ac-cepted for publication by the MPC, the number of

performance. Note that the LINEAR system is cover-ing nearly the entire sky visible above our site’s effec-tive southern declination limit of –35°, althoughsome of those areas are being covered at shorter inte-gration times and to shallower depths than others.The longest and deepest searches are concentratedalong the ecliptic.



NEAs and comets discovered, and the total numberof discoveries credited to LINEAR, including main-belt asteroids and other miscellaneous asteroid classes.The total database of observations published by theMPC was 21,002,723 observations at the end of2003. Thus LINEAR in five years of operations hascontributed over half of the total number of asteroidobservations worldwide.

Figure 8 puts these numbers in perspective by illus-trating the population of asteroids in the solar systeminterior to Jupiter. Main-belt asteroids are shown inyellow, NEAs are in red, comets are blue squares, andJupiter trojans are blue dots (trojans are resonant as-teroids approximately ±60° from the major planet).In addition to its discovery of nearly 200,000 aster-oids, LINEAR has also discovered more than a hun-dred comets, which makes it the most prolificground-based discoverer of comets as well. The ap-pendix, entitled “Contribution of LINEAR to CometScience,” discusses some of the ways LINEAR hasfundamentally advanced the study of comets.

Table 2 summarizes the percentage of worldwideasteroid discoveries made by LINEAR. Potentiallyhazardous asteroids (PHA) are NEAs whose size andpotentially close approach to the earth make them themost threatening objects in the NEA population (thelist of PHAs is maintained by the MPC). As previ-ously noted, LINEAR currently accounts for nearly65% of all NEA and PHA discoveries made duringthe five years of operation from March 1998 through2003. Overall, LINEAR accounts for 53% of all NEAand PHA discoveries ever made, dating back to thefirst discovery over a hundred years ago.

Responding to a Congressional mandate in 1998,NASA set a goal of discovering, by 2008, 90% of allNEAs larger than one kilometer. At the time the goalwas set, there were 223 known large NEAs, and theestimated number of large NEAs ranged from 500 to2200 [10, 20]. Since that time another 468 largeNEAs have been discovered—323 of them by LIN-EAR—bringing the total to 691. In addition to in-creasing the discovery rate of NEAs, LINEAR hasprovided significant quantities of search statistics andsky coverage information. This information has beenused to improve the population estimate of large as-teroids [11, 12]. Figure 9 illustrates the discovery rate

FIGURE 7. The area of sky searched by LINEAR is shown for(a) October 2002, (b) May 2002, and (c) composite coveragefrom January to December 2002. The depth of search shownis the good-weather, background-corrected, single-frame-equivalent limiting magnitude.

17.00 17.50 18.00 18.50 19.00 19.50 20.00

–30

–60

–90

0

30

60

90

180135 –135

–30

–60–90

0

30

60

90

9045 135

–30

–60

–90

0

30

60

90

(a)

(b)

(c)



of large NEAs during the last thirty years for bothLINEAR and the rest of the world. The figure alsoshows the current estimate of the number of largeNEAs—about 1090 ± 180, which indicates the rela-tively small range of uncertainty in characterizing thepopulation of large asteroids.

Interesting Discoveries

Not surprisingly, LINEAR’s discoveries of nearly200,000 asteroids and comets include some interest-ing and unique objects. The most notable discovery isa new class of inner-earth-orbit asteroids in February2003; that is, an asteroid—2003 CP20—whose orbitis entirely interior to the earth’s orbit. The existence ofsuch objects had been theorized for years, but notproven until the discovery of 2003 CP20. LINEARhas also discovered two objects in resonance with theearth, both with unique horseshoe-type orbits, desig-nated 2000 PH5 and 2002 AA29. While 2000 PH5will maintain its horseshoe-type appearance onlythrough the year 2006, 2002 AA29 will likely be theearth’s companion for at least another hundred years.

In January 2000, LINEAR discovered a sun-graz-ing asteroid—2000 BD19—with the closest knownapproach to the sun. Even though no cometary activ-ity has been spotted, some astronomers suggest that2000 BD19 is an extinct comet. In November 2003LINEAR discovered an object—2003 WT42—withthe largest known aphelion (distance away from thesun). LINEAR also discovered the first-known retro-grade asteroid in June 1999, which is now numberedand named (20461) Dioretsa. A week later a secondretrograde asteroid was found, and a year later a thirdwas found. LINEAR is credited with discovering fourof the six known retrograde objects.

Table 1. LINEAR Observations and Discoveries

Observations NEAs Comets Discoveries

Pre-1998 69,092 18 0 2135

1998 553,518 135 16 18,268

1999 1,056,684 161 22 29,207

2000 2,016,162 258 17 52,642

2001 2,992,192 277 19 48,226

2002 2,914,715 286 26 31,636

2003 2,431,489 235 25 15,048*

Total 12,033,852 1370 125 197,107

* statistics not yet available

FIGURE 8. The asteroid population of the solar system inte-rior to the orbit of Jupiter. Main-belt asteroids between Marsand Jupiter are shown in yellow, near-earth asteroids (NEA)are shown in red, comets are shown as blue squares, andJupiter trojans are shown as clusters of blue dots. (Imagecourtesy of the Minor Planet Center.)



T H E C E R E S C O N N E C T I O N :N A M I N G A S T E R O I D S I N H O N O R O F

E X C E L L E N C E I N S C I E N C E

FIGURE A. An educational brochure is given to each honoree to describe thehonor and explain the art and history of discovering asteroids.

of the Interna-tional Astronomical Union, thediscoverer of an asteroid eventu-ally obtains the right to suggest aname for it. In order for an aster-oid to be formally numbered,and thus eligible for naming, itsorbit must be well determined sothat the asteroid will not be lostin the future. Developing a goodorbit normally takes a few appari-tions, or perhaps five years for amain-belt object. LINEAR hasbeen observing continually sinceMarch 1998 and has accrued dis-covery credit for nearly 200,000objects, of which more than30,000 of them have been num-bered and are available to benamed. Each month several hun-dred more LINEAR discoveriesare numbered, thus continuouslyadding to the total.

By 2001, LINEAR had ac-crued enough naming rights toprecipitate serious thought onhow to employ these rights togreatest benefit. Because the In-ternational Astronomical Unionforbids the use of naming rightsfor financial gain, operating thesearch by selling names is not anoption. LINEAR is discoveringso many asteroids that the teamfelt an obligation to avoid de-valuing the honor of an asteroidname.

After careful consideration, itwas decided that the highest andbest use of the honor of namingan asteroid was to invest it in pro-moting science education in theinternational community. We de-cided to name LINEAR-discov-ered asteroids in honor of juniorhigh school and high school stu-dents who have demonstrated ex-cellence in select science compe-titions. The name chosen for theasteroid-naming program is theCeres Connection, since Cereswas the first minor planet, dis-covered by Italian astronomerGiuseppe Piazzi in 1801. TheCeres Connection program fits

in well with the objectives of Lin-coln Laboratory and MIT, andwith the educational outreachobjectives of NASA.

The Ceres connection was de-veloped in cooperation with Sci-ence Service, Inc., an organizerand administrator of several na-tional and international competi-tions. It was inaugurated on 23October 2001, with an awardspresentation in Washington,D.C., to the forty finalists andtheir teachers in the DiscoveryScience Challenge Competition.Each student and each teacher re-ceived a certificate denoting anofficially numbered minor planet

Ceres ConnectionThe

How Minor PlanetsAre Discovered and Named



Besides discovering asteroids with unique orbits,LINEAR has also found a number of asteroids withunique light curves. Radar observations have shown1999 KW4 and 2000 DP107 to be binary objects,i.e., a pair of asteroids orbiting each other while orbit-ing the sun. The first binary asteroid pair was foundin 1993 by the Galileo spacecraft, and these two bi-nary objects account for the second and third knownpair. Finally, an early LINEAR discovery—(25143)Itokawa—was chosen by the Japanese as the targetdestination for the Hayabusa mission to an asteroid.A rendezvous is expected in September 2005, with areturn of a collected sample expected in late 2007.

The prolific discovery rate of the LINEAR pro-gram has enabled another project, entitled the CeresConnection, whereby Lincoln Laboratory promotesscience education worldwide by naming minor plan-

ets in honor of successful science students and theirteachers and mentors. This award program, whichwas awarded the 2002 MIT Excellence Award for bet-tering our community, is described in the sidebar en-titled “The Ceres Connection: Naming Asteroids inHonor of Excellence in Science.”

LINEAR Search Calibrationand Future Evolution

The productivity of LINEAR has been impressiveover the past half decade, but further evolution ofLINEAR and other asteroid search systems dependson having a better understanding of the most efficientmethods of search, and a good quantitative calibra-tion of the capability of the search. To that end, con-siderable effort is being applied to the following threeanalysis tasks.

named in their honor, along withexplanatory material, as shown inFigure A. The minor planet nameis either their last name or, if anasteroid was already named usingtheir name or a similar soundingname, a name is derived from acombination of their first and lastnames.

During the 2001/2002 aca-demic year, the Ceres Connec-

tion awarded additional naminghonors to the forty finalists andtheir teachers in the Intel ScienceTalent Search, and to 105 stu-dent winners at the InternationalSciences Fair held in Louisville,Kentucky.

In addition to rewarding thespecific achievements of thesestudents, the Ceres Connectionis intended to promote interest in

science education in the broadercommunity by popularizing sci-ence. Since the inauguration ofthe Ceres Connection in October2001, a total of 828 top-rankingscience students and teachershave returned to the classroomwith the message that excellencein science can result in a part ofthe solar system being officiallynamed in their honor.

Table 2. Worldwide Discoveries of Asteroids

March 1998 to December 2003 Total Known

LINEAR/Total Percentage LINEAR/Total Percentage

NEAs 1352/2146 63% 1370/2606 53%

Large NEAs 320/468 69% 323/691 47%

PHAs 271/418 65% 274/529 52%

Atens 126/175 72% 126/203 62%

Apollos 648/974 67% 659/1230 54%

Amors 581/996 59% 588/1173 50%



Search Experience Database. The first task is tomaintain detailed search experience information forLINEAR. This information will give a better under-standing of how LINEAR functions, and it will allowcomparison of LINEAR with other asteroid searches.This task has been accomplished by defining a data-base containing look-by-look standard measures ofthe seeing and sensitivity achieved by each telescopeas determined from star measurements.

Search Strategy Analysis. The second task is to de-velop a detailed understanding of the most effectivestrategies to search for asteroids one kilometer in di-

ameter and larger. It is necessary to decide how to dis-tribute observation efforts across the sky for optimalproductivity. This task was initially approached byplotting the LINEAR detections of all NEAs and alllarge NEAs (absolute magnitude less than 18.0), rela-tive to the opposition in ecliptic coordinates. Onlythe first detection during a lunation was included; de-tections resulting from directed follow-up activitieswere excluded. Figure 10 shows the results of theseplots, which indicate that LINEAR detects asteroidsat all declinations, and that for large objects such asthose shown in Figure 10(a) the distribution is some-what uniform. This suggests an all-sky survey is anappropriate strategy. Figure 10(b) shows all NEAs,displaying a detection bias toward ecliptic oppositiondue to smaller objects that are more likely to requireoptimal phase angles to be detectable. If future searchsystems can achieve a limiting magnitude perfor-mance substantially better than LINEAR, or if theprimary goal of the survey changes, then the questionof observation strategy should be revisited.

Search Capability Quantification. The third task isto develop a measure of search capacity that enables asystematic approach to evaluating the capability ofthe search. The driving metric for a search effort is thevolume of space searched. This volume may be calcu-lated by calibrating, on a field-by-field basis, thedepth of the search for the detection of an 18th abso-lute magnitude object. Once a reliable calibrationmethod is found, the volume of each field can simplybe accumulated over a period of time to generate aneffective search volume.

FIGURE 10. The location of NEA detections with respect to ecliptic opposition; (a) NEAs with absolute magnitudeless than 18, (b) all NEAs. The plots show only the first detection per lunation for a single NEA, and detections madeonly during search mode. These data show that LINEAR detects large asteroids at all declinations, which suggests anall-sky search strategy is appropriate.

FIGURE 9. The cumulative number of large NEAs is esti-mated to be 1090 ± 180. The NASA goal of discovering 90%of the large objects by 2008 received a significant boostwhen LINEAR joined the worldwide asteroid search in 1998.The discoveries of asteroids in the past three decades byother institutions around the world are shown in blue; the re-cent discoveries by LINEAR are shown in red.

0

200

400

600

800

1000

1973 1978 1983 1988 1993 1998 2003

Rest of the worldLINEAR

LINEAR joins the search

1200

1400

Estimated number of large NEAs

(a)

–60

–30

0

30

60

90

–180

–135

–90–45

045

90

135

180

(b)

–180

–135

–90–45

045

90

135

180

–60

–30

0

30

60

90



Achieving reliable field-by-field calibration is com-plex because of the variable conditions under whichthe observing is accomplished and the fact that con-siderable data-dependent processing occurs to detectmoving asteroids. The most obvious method to cali-brate the magnitude performance of a system is topick stars of known magnitude from the fields anddetermine how bright a star must be to achieve somestandard signal-to-noise ratio in the detection system.Given that the CCD pixels are large and the integra-tion times are short, asteroids do not streak and arethus indistinguishable from stars in any given frame.This process of computing a signal-to-noise ratio todetermine a system’s sensitivity takes into accountmany of the factors affecting the performance of thesearch, such as weather and seeing, but fails to con-sider the aspects of the detection algorithm that lookfor moving objects.

In order to validate the star signal-to-noise ratio asa general indicator of search depth, a large set of datataken near the ecliptic was identified and the detec-

tions extracted. Separately, objects in the MPC cata-logue that were expected to be in the LINEAR searcharea with known absolute magnitudes were identifiedand their magnitudes corrected for distance and illu-mination geometry to generate apparent visual mag-nitudes. With these two inputs—the list of detectionsand the list of known objects—a field-by-field cali-bration of the search system’s ability to detect aster-oids could be determined by plotting the percentageof known objects detected as a function of apparentvisual magnitude.

Figure 11 shows examples of such plots. Figure11(a) shows the measured data and a sigmoidal fit fora single night in 2001, resulting in a single detectionefficiency curve. Figure 11(b) shows the variation ofthe detection efficiency for five different nights in2001, with significant differences between clear fallnights and hazy, cloudy summer nights. Unfortu-nately, this curve-fitting process for determining thedetection efficiency breaks down away from the eclip-tic plane where there are few to zero detections per

FIGURE 11. The probability of detection for an asteroid search system decreases as the apparent magnitude of the ob-jects becomes fainter. The limiting magnitude of the system can be estimated from the curve describing this relation-ship. (a) Detection efficiency curve for a single night of observing in 2001, including the bin-by-bin measured detectionrates and the sigmoidal curve fit that models the detection efficiency. (b) Detection efficiency curves for multiple nightsobserving, showing the variation in limiting magnitude due to varying environmental factors.

Pro

babi

lity

of d

etec

tion

(%)

100

80

60

40

20

0

100

80

60

40

20

0

Apparent visual magnitudeApparent visual magnitude

Sigmoid curve fit Percent detected in bin

50%

10%

90%

8 10 12 14 16 18 20 22 8 10 12 14 16 18 20 22

50% detection efficiency

18 August25 August23 September21 October17 NovemberP

roba

bilit

y of

det

ectio

n (%

)(a) (b)

Date Objects detected Declination searched Weather comments18 August 2634 −4º to +4º Haze early, developing into full clouds25 August 4302 +7º to +14º Occasional scattered clouds all night, haze by morning23 September 5449 +0º to +6º Scattered clouds with occasional haze21 October 9634 +0º to +6º Mostly clear with some high clouds17 November 7771 +11º to +18º Scattered clouds



field. However, there are more than sufficient datacollected near the ecliptic to use this moving-objectcalibration process to validate the star signal-to-noiseratio as a measure of the search depth on a frame-by-frame basis. Thus the accumulated search volume canbe calculated by using the star signal-to-noise ratio.Figure 12 shows this volume in cubic astronomicalunits. Changes in slope in the data identify intervalsof particularly good or bad weather and the multiplesystem enhancements.

As the capabilities of a search system evolve, therate of volume searched should grow. The lower curvein Figure 12 shows a running computation of the vol-ume searched by LINEAR in the preceding year. Thesearch capability of the system evolved rapidly be-tween 1999 and early 2000, leveled off in 2000, andgrew dramatically in 2001 and 2002 because of an al-gorithm enhancement and a hardware fix. In late2002 and early 2003 the system was adversely af-fected by mount instabilities that have recently beenaddressed. As LINEAR evolves and achieves the fun-damental inherent capability of the existing tele-scope/detector/processing system, the variations inthe plot in Figure 12 should reflect observing experi-ence (i.e., lunations, weather, equipment failure, andstaff availability) rather than system limitations. Thusthese variations represent a powerful tool to monitor

the operation of the search and to identify problemsthat may arise.

The Future of Asteroid Search Technology

This article provides an overview of the current LIN-EAR search system and its recent evolution. Whatdoes the future hold for the next generation of aster-oid search systems? Because of the exponential marchof Moore’s law, which has given us highly accurateCCD detectors and a phenomenal increase in pro-cessing capacity, there are probably no more factors oftwo in increased search performance for search sys-tems using existing one-meter telescopes.

In designing the LINEAR system, the LincolnLaboratory team put considerable effort into mini-mizing bottlenecks by matching the capabilities ofeach of the subsystems to work well with the rest ofthe system and to maximize the total system capabil-ity. Given soon-to-be-available detector mosaic sizes,the limitation of asteroid search systems will be aper-ture size. Historically, the astronomy communitybuilt telescopes with increasing aperture but withsmall fields of view, compared to that needed tosearch a reasonable fraction of the sky. Sensitivity tocharacterize extremely faint objects has been the maininterest driving the designs of these telescope systems,rather than wide-area search capability. In addition,telescopes with a wide field of view become progres-sively more difficult and more expensive to build asthe aperture size increases. Therefore, few wide-field-of-view telescopes with apertures exceeding one metercurrently exist.

The astronomical community, however, is nowembracing synoptic astronomy, i.e., viewing large ar-eas of the sky for time-variable phenomena. This in-terest will most likely drive the next generation of as-tronomical telescopes to a wider field of view, asevidenced by the proposed Large Synoptic SurveyTelescope (LSST) and the Panoramic Survey Tele-scope and Rapid Response System (PanSTARRS).

In 2003 members of the LINEAR team partici-pated in a special NASA study to address the feasibil-ity of searching for smaller near-earth asteroids uponthe completion of the current NASA goal. The NASAScience Definition Team was composed of a dozentop scientists from around the nation representing the

FIGURE 12. The volume searched by LINEAR is determinedby multiplying the area of sky searched for each field of viewby the depth of the search for that field, where the depth isdetermined from the limiting magnitude for an object ofabsolute magnitude 18. The upper line shows the cumulativesky searched; the lower line shows the annual volumesearched.

2

4

6

8

10

12

March1998

March 1999

March2000

March2001

March2002

March2003

Vol

ume

(cub

ic a

stro

nom

ical

uni

ts)



various asteroid search and impact hazard specialties.The nine-month study resulted in a report that rec-ommends the next NASA goal should be to eliminate90% of the impact hazard risk by detecting 90% of allobjects larger than 140 meters in diameter [21]. Thereport also offered a list of technologically feasible as-teroid search systems that could accomplish such agoal in a given time period. While the recommendedgoal is beyond the capability of LINEAR and othercurrent asteroid search systems, it is certainly attain-able from space, or from the ground with multiplewide-field-of-view, large-aperture systems. If a newNASA goal is actually stated and later achieved, it iscertain that yet another goal will follow—perhaps ad-dressing comets.

Acknowledgments

Many talented people have worked over the years tomake LINEAR a world-class NEA detection system.The authors would like to thank the many people atLincoln Laboratory who contributed to the software,algorithms, and analysis that have helped the systemto continue to evolve. In particular we thank RonSayer, Scott Stuart, Jeff Kommers, Caroline Klose,Herb Viggh, and Eric Pearce. The authors also thankthe people at the Experimental Test System in NewMexico who develop the software, operate and main-tain the system, and process the data, including MattBlythe, Mike Bezpalko, Jeff English, Bob Huber, JulieJohnson, Ray Kracke, Heidi Love, Lisa Manguso,Matt McCleary, Doug Torres, Peter Trujillo, and TomRuekgauer. This work was sponsored by the Depart-ment of the Air Force and by NASA.

R E F E R E N C E S1. W.F. Bottke, Jr., R. Jedicke, A. Morbidelli, J.-M. Petit, and B.

Gladman, “Understanding the Distribution of Near-Earth As-teroids,” Science 288 (5474), 2000, pp. 2190–2194.

2. A. Morbidelli, W.F. Bottke, Ch. Froeschlé, and P. Michel,“Origin and Evolution of Near-Earth Objects,” in Asteroids III,W.F. Bottke, A. Cellino, P. Paolicchi, and R.P. Binzel, eds.(University of Arizona Press, Tucson, 2002), pp. 409–422.

3. T. Gehrels, “CCD Scanning,” in Asteroids, Comets, and MeteorsII, C.-I. Lagerkvist, B.A. Lindblad, H. Lundstedt, and H.Rickman, eds. (Uppsala University, Uppsala, Sweden, 1986),pp. 19–20.

4. S.H. Pravdo, D.L. Rabinowitz, E.F. Helin, K.J. Lawrence, R.J.Bambery, C.C. Clark, S.L. Groom, S. Levin, J. Lorre, S.B.

Shaklan, P. Kervin, J.A. Africano, P. Sydney, and V. Soohoo,“The Near-Earth Tracking (NEAT) Program: An AutomatedSystem for Telescope Control, Wide-Field Imaging, and Ob-ject Detection,” Astron. J. 117 (3) 1999, pp. 1616–1633.

5. C.F. Chyba, P.J. Thomas, and K.J. Zahnle, “The 1908 Tung-uska Explosion: Atmospheric Disruption of a Stony Asteroid,”Nature 361 (6407), 1993, pp. 40–44.

6. J.G. Hills and M.P. Goda, “The Fragmentation of Small Aster-oids in the Atmosphere,” Astron. J. 105 (3), 1993, pp. 1114–1144.

7. D. Morrison, C.R. Chapman, and P. Slovic, “The Impact Haz-ard,” in Hazards Due to Comets & Asteroids, T. Gehrels, ed.(University of Arizona Press, Tucson, 1994), pp. 59–92.

8. O.B. Toon, K. Zahnle, D. Morrison, R.P. Turco, and C. Covey,“Environmental Perturbations Caused by the Impacts of As-teroids and Comets,” Rev. Geophys. 35 (1), 1997, pp. 41–78.

9. S.R. Chesley and S.N. Ward, “A Quantitative Assessment ofthe Human and Economic Hazard from Impact-GeneratedTsunami,” submitted to Environ. Hazards (2003).

10. D. Morrison, R.P. Binzel, E. Bowell, C. Chapman, L. Fried-man, T. Gehrels, E. Helin, B. Marsden, A. Maury, T. Morgan,K. Muinonen, S. Ostro, J. Pike, J. Rahe, R. Rajamohan, J.Rather, K. Russell, E. Shoemaker, A. Sokolsky, D. Steel, D.Tholen, J. Veverka, F. Vilas, and D. Yeomans, “The SpaceguardSurvey: Report of the NASA International Near-Earth-ObjectDetection Workshop,” NASA, Washington, 25 Jan. 1992.

11. J.S. Stuart, “Observation Constraints on the Number, Albe-dos, Sizes, and Impact Hazards of the Near-Earth Asteroids,”Ph.D. thesis, MIT (Cambridge, Mass., 2003).

12. Stuart, S.J., “A Near-Earth Asteroid Population Estimate fromthe LINEAR Survey,” Science 294, pp. 1691–1693 (2001).

13. W.F. Bottke, A. Morbidelli, R. Jedicke, J.-M. Petit, H.F. Levi-son, P. Michel, and T. S. Metcalfe, “Debiased Orbital andAbsolute Magnitude Distribution of the Near-Earth Objects,”Icarus 156 (2), 2002, pp. 399–433.

14. G.H. Stokes, F. Shelly, H.E.M. Viggh, M.S. Blythe, andJ.S. Stuart, “The Lincoln Near-Earth Asteroid Research(LINEAR) Program,” Linc. Lab. J. 11 (1), 1998, pp. 27–40.

15. G.H. Stokes, J.B. Evans, H.E.M. Viggh, F.C. Shelly, and E.C.Pearce, “Lincoln Near-Earth Asteroid Program (LINEAR),”Icarus 148 (1), 2000, pp. 21–28.

16. G.H. Stokes, R. Weber, F. Shelly, D. Beatty, H. Viggh, E. Rork,and B. Hayes, “Air Force Planetary Defense System: InitialField Test Results,” Proc. Fifth Int. Conf. on Space ’96 1, 1996,pp. 46–53.

17. H.E.M. Viggh, G.H. Stokes, F.C. Shelly, M.S. Blythe, and J.S.Stuart, “Applying Electro-Optical Space Surveillance Technol-ogy to Asteroid Search and Detection: The LINEAR ProgramResults,” Proc. Sixth International Conf. and Exposition on En-gineering, Construction, and Operations in Space, Albuquerque,N.Mex., 26–30 Apr. 1998, pp. 373–381.

18. R. Jedicke, “Detection of Near-Earth Asteroids Based uponTheir Rates of Motion,” Astron. J. 111 (2), 1996, pp. 970–983.

19. Minor Planet Center, Harvard-Smithsonian AstrophysicalObservatory, Cambridge, Mass. (2000), <http://cfa-www.harvard.edu/iau/mpc.html>.

20. D.L. Rabinowitz, E.F. Helin, K. Lawrence, and S. Pravdo, “AReduced Estimate of the Number of Kilometer-Sized Near-Earth-Asteroids,” Nature 403 (13), 2000, pp. 165–166.

21. “Study to Determine the Feasibility of Extending the Searchfor Near-Earth Objects to Smaller Limiting Diameters,” Re-port of the Near-Earth Object Science Definition Team, 22Aug. 2003; available at the website <neo.jpl.nasa.gov/neo/neoreport030825.pdf>.



A P P E N D I X : C O N T R I B U T I O N O F L I N E A RT O C O M E T S C I E N C E

In addition to being the world’s most productive as-teroid search program, LINEAR has profoundly al-tered the field of comet science. LINEAR’s detectionalgorithm, based on algorithms used to detect earth-orbiting satellites, is fundamentally a moving-objectdetector. Any object in motion across the fixed starpattern, within the dynamic range of the algorithm(about 0.1 to 10+ deg/day), is duly recorded. Sincethese rates of motion are characteristic of comets asthey enter the inner solar system, LINEAR has dis-covered more than a hundred comets, making it themost prolific ground-based discoverer of comets inhistory. (The space-based Solar and Heliospheric Ob-servatory [SOHO], the most prolific system, has dis-

covered more than five hundred comets, typicallyshortly before they impact the sun.)

Most of the comets discovered by LINEAR arefound on their inbound trajectory, as they pass the or-bits of Saturn or Jupiter. At this point, the cometstarts to brighten as volatile materials are evolved bysolar heating, and the comet becomes detectable byLINEAR. Typically, the LINEAR system at this timedoes not notice any comet-like trailing feature thatwould clearly identify the object as cometary. Thusthe comet detection observations are routinely passedto the Minor Planet Center (MPC) along with hun-dreds of thousands of asteroid observations generatedeach month. At this point, one of two possible actionsresults in the object being identified as a comet: (1)the orbit of the object is calculated and determined tobe comet-like, as opposed to asteroid-like, and the

FIGURE 1. The cover of the 18 May 2001 issue of Science.The entire issue of this magazine was dedicated to comets,with a special focus on C/1994 S4 LINEAR. (Cover image re-printed with permission of Science magazine, copyright 2001,American Association for the Advancement of Science.)

FIGURE 2. The path of comet C/1999 S4 LINEAR throughthe solar system. It was discovered at point 1 on 27 Septem-ber 1999, and LINEAR S4 disintegrated from 21 to 24 July2000 near its closest approach to the earth. The position ofthe earth is shown as a green circle for the discovery epoch(1) and for the disintegration epoch (3).

C/1999 S4

C/1999 S4 LINEARdisintegrates

21–24 July 2000

Earth

Jupiter

Mars

2

1

3

3

31

1 27 September 199925 June 200026 July 20003

2

Earth



MPC requests an observer with a large telescope tocheck the object for a tail; (2) if the object is postedon the MPC confirmation page because of its inter-esting rate of motion, a follow-up observer may de-tect a tail.

This process of comet discovery is fundamentallydifferent from the process prior to LINEAR opera-tions. In the pre-LINEAR era, amateur observers usu-ally discovered comets by scanning regions close tothe sun. By the time a comet is near the sun, it hasheated up and formed a characteristic tail, whichmakes it detectable. The amateur method has two de-ficiencies: (1) only comets that travel close enough tothe sun and are active enough to develop a large tailare discovered; (2) the comet is discovered only afterit has substantially completed its inbound trajectory.Thus the heating and tail formation process are notobserved or recorded.

By finding comets far from the sun, LINEARhelps to solve both these issues. Many comets thatnever form tails large enough to be visible are discov-ered, and—more importantly—comets are discov-ered early in their trajectory. This early detection en-ables comet scientists to gather observations coveringthe interval in their orbit where they become active,

evolve a tail, and break into pieces. In addition,enough warning is provided to allow time to scheduleadditional observations by other assets such as theHubble space telescope and the Keck Observatory.

These observation opportunities have led to somestriking discoveries, and have resulted in the dedica-tion of an entire 2001 issue of Science magazine tocomets, with a special focus on comet C/1999 S4LINEAR. Figure 1 shows the cover of that issue. Thefollowing sections describe a few of the fascinatingcomets discovered by LINEAR.

Comet C/1999 S4 LINEAR

Figure 2 illustrates the orbit of comet C/1999 S4LINEAR, which was discovered on 27 September1999 just inside the orbit of Jupiter. By June 2000,LINEAR S4 had a well-developed tail, as shown inthe CCD image in Figure 3, and was expected to bevisible to the naked eye at a closer approach to theearth (the dark adapted eye at a dark site is sensitive toobjects of 5th to 6th magnitude). In reality, LINEARS4 peaked with an intensity of about 6.5 in late July(visible through binoculars) and then disintegratedfrom 21 to 24 July 2000. Due to the long time be-tween discovery of LINEAR S4 and its closest ap-

FIGURE 3. An image of comet C/1999 S4 LINEAR taken by one of the LINEAR tele-scopes on 25 June 2000, showing a characteristic well-developed tail.



FIGURE 4. (top) Hubble space telescope observations showing C/1999 S4 LINEAR flaring up and be-ginning to disintegrate in early July 2000. (bottom) Later observations show the cometesimals remain-ing a couple of days after the breakup of LINEAR S4 on 24 July 2000. (Hubble images courtesy of H.A.Weaver [Johns Hopkins University], NASA, and the Space Telescope Science Institute.)

FIGURE 5. Image of the Christmas Comet as it appeared inthe southern hemisphere on 30 January 2001. (Image cour-tesy Terry Lovejoy, Australia.)

proach to earth, the Hubble space telescope wasscheduled for observations of LINEAR S4 in July2000 and recorded the comet’s activity and residualcometesimals. These images of comet LINEAR S4provided a wealth of insight into comet evolution andfunction. Figure 4 displays a sample of Hubble spacetelescope image data before and after disintegration ofthe comet.

Comet C/2000 WM1 LINEAR: The Christmas Comet

C/2000 WM1 LINEAR was discovered by LINEAR



FIGURE 6. Composite picture of comet C/2001 A2 LINEAR, showing asharp nucleus and a tail containing several streamers. (Copyright imagescourtesy Gordon Garradd.)

on 16 November 2000 between the orbits of Saturnand Jupiter. This comet is the first truly naked-eyecomet discovered by LINEAR. Because it appeared atthe end of 2000 it was dubbed the Christmas Cometby the media. It was as bright as 5th magnitude in thenorthern hemisphere and then dropped below thesouthern horizon, where it was as bright as 3rd mag-nitude as seen from Australia. Figure 5 shows an im-age of the Christmas comet.

Comet C/2001 A2 LINEAR

As mentioned earlier, one of the important contribu-tions of the LINEAR program has been the discovery

of comets before they have evolved tails. This earlydetection allows astronomers to observe the heatingphase of the inbound trajectory, which results in dataand insight on the behavior of these unique objects.Comet C/2001 A2 LINEAR shown in Figure 6 typi-fies the erratic behavior of comets. It had many out-bursts in brightness—probably associated with astructure change that increased the outgassing of ma-terial—and the nucleus has split into several pieces.As characterized by one astronomer, “This comet be-came the kind of comet astronomers love—truly un-predictable in its behavior.”



. is the associate head of theAerospace division, where heis responsible for the spacecontrol mission area at theLaboratory. He specializes inanalysis, design, and opera-tions of space surveillancesystems, including the Space-Based Visible (SBV) andLINEAR programs. The SBVsystem provides the first space-based space surveillance capa-bility to Air Force Space Com-mand in Colorado Springs,Colorado. The LINEARprogram utilizes space surveil-lance technology developed forthe U.S. Air Force to searchfor near-earth asteroids. Beforecoming to Lincoln Laboratoryin 1989, he worked as a seniorscientist and operations man-ager at Geo-Centers Inc., acontracting company specializ-ing in fiber-optic sensors.Previously, he performednondestructive testing of laserfusion targets at Los AlamosNational Laboratory in NewMexico, and developed fiber-optic data-acquisition systemsand provided field support forunderground nuclear tests inNevada. He has a B.A. degreein physics from ColoradoCollege, and M.A. and Ph.D.degrees in physics fromPrinceton University.

. is a staff member in the SpaceControl Systems group, whereher primary responsibility isproviding technical leadershipin the LINEAR program.Additionally, she specializes inimage processing and dataanalysis related to variousspace surveillance systems. Shealso provides software manage-ment and occasionally still getsa rare treat of performingactual software development.Prior to joining the SpaceControl Systems group, shespent six years in the AirTraffic Surveillance group,where she specialized in track-ing algorithms and played akey role in the upgrade to theASR-9 radar. She has B.S. andM.S. degrees in electricalengineering from the OhioState University, and shejoined Lincoln Laboratory in1990. While proud of hercontributions to numeroussuccessful projects over theyears, she is most proud of heraccomplishments outside ofthe work arena, namely herchildren David and Daniel,two happy, sports-loving,budding scientists. In herlimited free time, she enjoysreading and attempting vari-ous outdoor sports such asskiing and golf.

. is an associate staff member ofthe Space Control Systemsgroup, based at the LincolnLaboratory Experimental TestSystem near Socorro, NewMexico. He develops imageprocessing software that isused to automatically detectasteroids, comets, and satel-lites, as well as real-time con-trol software for positioningtelescope mounts. He alsogenerates system software forinterfaces to custom devicessuch as weather systems,mount encoders, digital-to-analog converters, CCDcameras, and filter wheels. Hehelped develop the Transport-able Optical System (TOS),which the U.S. Air Forceoperates in Spain. He wrotemost of the LINEAR asteroiddetection software, and hecontinues to be heavily in-volved in the operations andimprovement of the system.He joined Lincoln Laboratoryin 1986 after studying com-puter science at the NewMexico Institute of Miningand Technology in Socorro.

Detection and Discovery of Near-Earth Asteroids by the LINEAR … · 2019-02-15 · vere local...

Documents

Transcript of Detection and Discovery of Near-Earth Asteroids by the LINEAR … · 2019-02-15 · vere local...