From erik Tue Mar 13 08:03:57 2001 Subject: CUO-84 NEO detection To: gaia-sag@astro.estec.esa.nl (GAIA SAG) Date: Tue, 13 Mar 2001 08:03:57 +0100 (MET) X-Mailer: ELM [version 2.5 PL2] Content-Length: 10615 Status: OR Detection of NEOs with GAIA and ground-based follow-up ====================================================== GAIA-CUO-84 E. Hoeg, J. Knude 12 Feb. 2001 Full text on http://www.astro.ku.dk/~erik/gaia/84.neo.obs ABSTRACT: We proposed in June 2000 in the report CUO-77 a modification of GAIA suited for efficient detection of NEOs. With this modification GAIA will be able to detect a large number of NEOs closer to the sun than with any other method: 35 degrees or 0.6 AU. This has recently been shown in realistic simulations by F. Mignard. We discuss the possibility for ground-based follow-up quickly after of such detections. Main points are that one GAIA observation is sufficient for starting a ground based follow-up for orbit determination, and that such follow-up is nearly always possible with a moderate size telescope of, e.g., 1.5 m aperture. Observation group (OG) ---------------------- It should be assumed that a second sky mapper (SSM2) is introduced in the Spectro telescope as proposed in CUO-77. SSM2 is situated at the end of the Spectro field-of-view (FOV), 120 seconds of crossing time after SSM1. An observation group (OG) for an object, star or NEO, is defined as the consecutive observations obtained with GAIA within one revolution of 3 hrs (OGrev), or within a whole visibility period (OGvis) during which a telescope may measure a given star at 3hrs intervals until the motion of the spin axis has moved so far that the star is not seen again for about one month. Definitions: Basic observation : an observation on SSM1, SSM2 or ASM1 OG3 : observation group with the basic observations (SSM1+2, ASM1) OGrev : observation group of either (SSM1+2, ASM1) or (SSM1+2, ASM1, ASM1) OGvis : the observations in a visibility period containing several OGrev It appears that one "observation" in FM-007 is defined as one OGrev (see Section 5) with the basic observations from the two Astro telescopes during 3 hours: (ASM1, ASM1). In FM-008 the Spectro telescope is included, and an "observation" with Spectro means the observation group (SSM1+2, SSM1+2) obtained during 3 hours. An "observation" with the Astros has the the same meaning as in FM-007. We shall in this report a similar loose terminology so that a GAIA "observation" usually means an "observation group", without always specifying which kind it is. It is shown in CUO-77 that an observation group, OG3, of three basic observations: one in SSM1, one in SSM2 and one in ASM1 of the Astro telescope will give a precision of 50 mas for the position and 0.035 arcsec/hr (= 0.010 mas/s) for the velocity of an object of V=20 mag. A NEO typically moves 40 mas/s (=1.0 deg/day) according to FM-007-Fig.12, and very seldom more than 100 mas/s. This implies that a detected NEO after 4 days will typically have moved to a position 4 deg away and uncertain by roughly 4 arcsec. Follow-up from the ground ------------------------- We only discuss the follow-up required after a GAIA detection of a NEO in order to determine an orbit. Any other follow-up to obtain spectroscopy and photometry can then be scheduled and is not discussed here. One observation group (OG3) gives a position and a velocity vector of such precision that the NEO can be found by a ground-based observation using a field of a few arcmin size if the observation is obtained within a few days. This requires observations at distances to the sun as small as 35 deg. The minimum practical distance is about 30 deg, e.g. with the Danish 1.5 m in Chile. The sky is dark when the sun is 15 deg below the horizon, and observation is possible to about 15 deg above the horizon. The resulting angle between sun and object depends of course on geographic latitude and declination, but it seems possible to have telescopes available in north and south which can quickly follow-up on almost any NEO detected with GAIA. The SNR has been calculated with the ESO imaging exposure time calculator available on the ESO website. Assuming altitude 15 deg, thus airmass 4.0; seeing 1.2 arcsec at zenith, and sqrt(airmass) x 1.2 = 2.4 arcsec; full moon, thus 20 mag/arcsec^2 in V and 19.2 mag in I; this background happens to be the same as of the zodiacal light 35 deg from the sun; SNR=10, sufficient for a safe detection; then 65 s exposure is required for an object of V=20.0 mag, and 45 s for I=19.0 mag, in the respective bands. I=19.0 corresponds V=20.0 since V-I=1.0 for a typical NEO. The telescope should be set to track with the known velocity of the NEO so that it would appear as a point and the reference stars would be elongated. An exposure of 60 s would typically give an elongation of 60x0.04=2.4 arcsec for the stars which is about the size of the seeing disk at an altitude of 15 deg. Thus, tracking is often not at all required. Elongated stars of V=20 mag would mostly be visible and could in principle be used as reference stars. By the time GAIA flies we can at least assume that all stars with V<18 mag have sufficiently good positions and proper motions to be used as reference stars. This gives a average density of 3 stars/arcmin^2. Thus a fairly small field of view is sufficient for our purpose. The limiting magnitude would be about V=22 mag with the 1.5 m using the integrated light, without filter on the CCD, if the atmospheric dispersion is compensated. This could be done by means of a simple variable prism in front of the CCD. It could consist of two identical prisms each with a dispersion corresponding to half the maximum atmospheric dispersion to be compensated. The prisms would be rotated before an observation so that the required compensation is achieved. The resulting position of the NEO would be used in orbit determination. Thus all NEOs detected by GAIA can be followed with quite short integration times on 1 - 1.5 m class telescopes on the ground. Completeness of GAIA detection ------------------------------ It appears from Table 1 of FM-008 that about 50 per cent of NEOs with absolute magnitude H=18.5 (diameter about 1 km) will be detected in at least one observation group by GAIA. The fraction is only 12 per cent for H=20, diameter 0.5 km. It would be interesting to know the distributions 1) of orbital elements for the detected and undetected fractions; 2) of the minimum distance between the orbits of earth and NEO; NEOs with small minimum distances are more liable to be perturbed to hit the earth in the near future. A small minimum distance will on average imply a brighter magnitude during the GAIA observing period. Perhaps the detected NEOs have generally smaller minimum distances than the undetected ones, meaning that the more dangerous NEOs tend to be sooner detected than the less dangerous. It would also be interesting to know 3) the distribution of the closest-point angle, ie. the angle between the two orbits at the closest point because the smaller this angle the higher the chances that a perturbation may bring the two bodies to a collision one day. Perhaps the product of minimum-distance and closest-point angle is a good measure for the danger: the smaller it is the more dangerous the NEO, and the higher is also the probability of detection by GAIA. The inverse of the product could be called the threat. It would be interesting to see 4) the distribution of the product or the threat. The NEOs which get only one GAIA observation group are most in need of follow-up from ground. Of course we do not know at the time of an observation of a NEO whether this object will get another GAIA observation later on. But it is interesting to know from simulations what fraction of the NEOs which get only one GAIA observation group can be followed-up from the ground. So we would like to know for these 5) the distributions of the angular distances to the sun and the apparent magnitudes at the time of observation, 6) these distributions, e.g. 4 days after the GAIA observation, and 7) how frequently are follow-up observations required for a safe orbit determination. Further simulations with the sample of 20 000 NEOs could answer these questions. This sample of NEOs and their orbits could perhaps be made public as a common basis for comparison of the performance of various observing methods: GAIA, BepiColombo, ground-based. This would, e.g., be interesting for a group in Denmark. Fainter detection limit in the sky mappers ------------------------------------------ It should be considered to extend the detection limit of the Spectro sky mappers to a fainter magnitude, e.g. V=21.0 instead of 20.0. The positions obtained in the sky mappers for all these objects should be transmitted to ground in order to possibly find NEOs among them. The extra load on telemetry would be moderate since no medium-band photometry should be transmitted for the stars fainter than 20.0 mag. The SNR for these two magnitudes is given in Table 1 for two values of integration times, t=2 and 3 s. It seems that an integration time of 3s as already foreseen in the GAIA baseline would also be sufficient for V=21.0. However, the question of the resulting number of false detections should be studied, as they should not be allowed to overload the on-board data handling or the telemetry. The number of false detections could be kept very low if the SSMs are placed in pairs, as in the Astro telescopes where ASM3 is sampled around the position of the star detected in ASM1. Similarly, the first SSM may consist of SSM1+SSM2 and after the 120 s crossing of the field would follow SSM3+SSM4. This means 3+3 seconds more for the sky mappers at the expence of available integration time in the medium-band photometer. All pixels should be read from SSM1 and SSM3 and all stars with V<21.0 mag should be detected. From SSM2 and SSM4 only 5x5 pixels centred on the detection should be read as in ASM3 in order to verify the detection. The readnoise in these would therefore be less, about 3e-, resulting in higher SNR than in Table 1. Table 1. Detection in Spectro, with no filter, ie. in the broad G band. SNR for integration times t=2 and 3 seconds, always assuming a sky background of 21.0 mag/arcsec^2, total readnoise = 5.3 e- per pixel. The SNRs are derived from Columns 2 and 3 in Table 3 of SAG-CUO-67. V SNR SNR t 2 s 3 s mag 20.0 6.2 8.3 21.0 2.6 3.4 REFERENCES: SAG-FM-007 of 12 Feb. 2001 by Francois Mignard SAG-FM-008 of 28 Feb. 2001 by Francois Mignard SAG-CUO-77 of 23 June 2000 by E. Hoeg. --------------- end -----------------------------------------------