Overview of the JCMT

H.E. Matthews, J. Leech

Date: 10 November 2004

1 Introduction

The JCMT is a 15-m diameter telescope optimised for observations at sub-millimeter wavelengths, situated close to the summit of Mauna Kea, Hawaii. Although the JCMT is a partnership between the United Kingdom, Canada and the Netherlands in accordance with the site agreement with the University of Hawaii, proposals can be accepted from individuals and groups of any nationality.

2 Telescope and Site Properties

2.1 Basics

The James Clerk Maxwell Telescope has an alt-azimuth mounting, and is engineered according to the principle of homologous deformation¹ to maintain a paraboloid at all elevations. It is housed in a dome, or ``carousel'', which is physically unattached to the antenna, but co-rotates with the latter. To protect the antenna and associated equipment from the elements, the carousel has both large doors and a sliding roof, which are stowed during operation. Under normal observing conditions, a Gortex membrane transparent to mm- and submm-waves is in place, protecting the telescope from exposure to the sun, wind, and dust. On account of the presence of the membrane the JCMT is the only mm/submm telescope capable of routinely observing the Sun directly, although this results in distortion of the telescope surface, very significantly affecting the beam shape and efficiency at frequencies above about 350 GHz.

The JCMT antenna is located at longitude(west) 155$^\circ$ 28$'$ 47$''$, latitude 19$^\circ$ 49$'$ 33$''$, at an altitude of 4092m. The local time is Hawaiian Standard (HST; = UTC $-$ 10 hours), and is in effect throughout the year. The antenna has a lower elevation limit of 5$^\circ$, and sources may not be tracked accurately above elevations of about 87$^\circ$; the upper limit for position- or beam-switched observations is likely to be 85$^\circ$ or even lower, depending on the distance between the reference and source positions. The normal operating azimuth maximum slew rate was increased in October 2001 to 1.0$^\circ$/sec following a major upgrade carousel drive system; following many years at the relatively sluggish speed of 0.6$^\circ$/sec this improvement is very welcome.

2.2 Antenna Surface

The continuing improvement of the antenna primary reflector surface is a major goal of the JCMT group. The adjustments necessary to obtain a perfect paraboloid are determined by what we loosely refer to as `holography', in this case by recording the beam pattern at two focus settings using a 94 GHz source located at UKIRT (i.e., within the near field). Such observations are mostly sensitive to the small-scale structure of the telescope surface, so they may be supplemented with in- and out-of-focus beam maps of bright planets to obtain the large-scale surface surface errors, and their variation with elevation. Measurements show that the total rms fluctuations of the antenna surface under stable night-time conditions remain about 30$\mu$m; this is the result of small-scale (fractions of a panel in size) errors having an rms of 16$\mu$m, and large-scale (typically about 5m) structure also with deviations of about 20$\mu$m. The panels have `aged' somewhat over the years, and options for improving this situation are being considered. Temperature changes introduce large effects which can be studied in real time; particularly during the day the overall rms error can increase to 50$\mu$m or more, largely due to temperature changes. This results in a very noticeable daytime decrease in efficiency at the higher frequencies. Observers who are particularly concerned with accurate extended source calibration should allow time in their proposal to supplement their datasets with beam efficiency measurements (for line observations) and maps of the telescope beam (for both line and continuum observing). Additional information, especially relating to the potential for real-time control via FEA modelling of the thermal behavior of the JCMT structure, can be found on the JCMT Web pages. In this context, the author wishes to note here the enormous contribution over many years to the telescope surface project made by Fred Baas, who passed away 2001 April 4 following a short illness. Jan Wouterloot is now leading this project.

2.3 Pointing and Tracking

The pointing model of the telescope is derived from continuing extensive measurements and incorporates azimuth track irregularities. Recent pointing models (see the JCMT pointing Web pages) routinely give pointing accuracies of about $1.5$ arcsec (rms) in both azimuth and elevation. The alignment with respect to the optical axis of each receiver is carefully checked on installation, and on other occasions as deemed necessary. Nevertheless, during an observing run it is advisable to check the local pointing offsets fairly frequently, and some allowance for such measurements needs to be made when calculating the total time required for a program.

The pointing and tracking accuracy of the antenna is carefully monitored. It has been especially rigorously investigated in the past year or two, with particular reference to irregularities in the azimuth track, central bearing, and behavior around transit, and compensation for these aspects incorporated into the on-line control system. The central bearing was successfully replaced in May 1999. In April 2001 the elevation encoder was replaced in response to the observation of large and variable changes around transit. Except for the latter and occasional more extreme azimuth track irregularities tracking appears to be better than 1 arcsec over periods of an hour or so, and may well be dominated by refraction noise in many cases. At higher elevations (above about 80$^\circ$), tracking may become less reliable due to the rapid movement of the source in azimuth, particularly during position-switched observations.

2.4 Beam Path

Incoming radiation is directed by the secondary mirror into the receiver cabin, located below the primary mirror surface between the elevation bearings at the $f/12$ Cassegrain focus. From here the optical path goes by means of a flat tertiary mirror to one or other of the possible receivers, which are mounted on racks within the cabin or on one of the Nasmyth foci platforms. The tertiary mirror is mounted on a turret which is under computer control to permit automatic redirection of the beam path to one or other of the receiver positions. Although the Nasmyth foci are nominally $f/35$² SCUBA uses an $f/16$ focus, obtained by the use of foreoptics on the SCUBA platform. One should be note that the image of the sky at the Nasmyth foci rotates as a function of both azimuth and hour angle. The SCUBA software takes this into account, while a special optical system ( the ``K-mirror'') is being installed to maintain the beam orientation with respect to the forthcoming heterodyne array receiver HARP for telescope elevation. Although a subject of occasional speculation, no plans have been made to provide a beamsplitter to allow more than one receiver to be used simultaneously.

2.5 Chopping Secondary Mirror

For many observations with the JCMT, in particular pointing and focus determinations, photometry, all SCUBA observations, and some spectral line observations, the secondary mirror is `chopped', or nutated, at a frequency and amplitude (`throw') which can be chosen by the user. This provides a detected signal which is the difference between the signals from the sky at the signal and reference positions, and from which most of the atmospheric background variations have then been removed. It is common practice also to `beam-switch', or `nod', the telescope while chopping, so that the source appears alternately in the signal and reference beams.

Most photometry and many spectral line observations are carried out using chopping in azimuth. However, in some cases it will be desirable to chop, say, in Right Ascension and thus ensure that the reference point is always in the same position relative to the source of interest. The secondary mirror control hardware and software permit observations of this type, and allow the user to choose any position angle for the secondary mirror chop direction. It is also possible to drive the mirror axes in essentially any pattern, such as the `jiggle' patterns required by SCUBA.

In principle one should choose the greatest chop frequency, and smallest chop throw, practical to obtain good results. The default chopping frequency for continuum measurements is 7.8125 Hz for both SCUBA and the heterodyne receivers. This gives reasonable atmospheric cancellation and good mechanical performance for at least the shorter chop throws (1-2 arcmin). `Beamswitched' spectral line observations use a lower chop frequency, usually 1 Hz. Chop throws larger than about 3 arcmin are not recommended.

2.6 Frequency-Dependent Telescope and Atmospheric Parameters

As telescope availability permits there is a continuing effort by local staff to obtain performance data for the telescope. An overview is given in Table 1, which includes the aperture efficiency, beamwidth, atmospheric transmission for 0.5 mm pwv³, and percentage of `good' nights at a number of representative frequencies (chiefly those of CO transitions). Since each receiver system illuminates the antenna surface differently, the efficiencies and beamwidths given should be used as a guide only. Consult the sections in this Guide on the spectral line systems and SCUBA for more complete information. Also see the JCMT Web pages.

Table 1: Overview of Telescope and Atmospheric Parameters

Frequency	Wavelength	Aperture	Beamwidth	Atmos.	Nights
(GHz)	($\mu$m)	Efficiency$^{\dag }$	(arcsec)	trans.	(%)
150	2000	0.66	28	0.97	90
230	1300	0.63	21	0.96	90
345	870	0.56	14	0.88	70
492	610	0.46	12	0.43	20
690	435	0.32	8	0.44	25
870	345	0.21	6	0.53	30
$^{\dag }$ Aperture efficiency calculated assuming the rms surface accuracy is 30$\mu$m

Using a figure of 30$\mu$m for the measured surface accuracy as a basis, the the aperture efficiencies given in Table 1 have been calculated using the standard formula⁴, with corrections to the theoretical aperture efficiency for losses and blockage. The fact that the measured values (see the Heterodyne Guide) are somewhat lower than the theoretical calculations indicates additional losses in the system.

At the lower frequencies the beam shape is well determined and an approximately circularly symmetric Gaussian. At the shortest wavelengths at which it has been determined, i.e., 350$\mu$m, the maximum amplitude of the error beam (`sidelobes') is about 10% of the main beam, if the telescope is correctly focussed. However, particularly in recent times, the beam pattern has undergone a number of changes, so that if it is important to data quality it is worth restating here that one should allow time as part of one's observing program to determine the beam pattern, telescope efficiencies and other quantities necessary to good calibration.

Figure 1 shows the calculated⁵atmospheric transmission at the zenith above Mauna Kea as a function of frequency for three different values of precipitable water vapour pressure. Strong absorption lines of atmospheric oxygen and water vapour divide the millimetre and submillimetre band into sharply defined `windows'. At the higher frequencies, the atmosphere allows at most rather less than half the incident radiation to reach the telescope. Representative values of the atmospheric transmission at the zenith are given in Table 1 also, as a guide to allow the calculation of atmospheric attenuation and contribution to system temperature at the frequencies listed. Bear in mind that conditions are highly variable on many occasions, that the given frequencies mostly are close to the peaks of the atmospheric transmission curve, and that the optical depth along a given line of sight varies approximately as the cosecant of the elevation.

Figure 1: Atmospheric transmission calculated (using the IRAM ATM routine; see text) as a function of frequency in the submillimetre window for three different water vapour pressures (1mm pwv is a `good' night, 0.5mm pwv `exceptional', and 5mm pwv is `rather nasty'). Useful observations are possible only in the 230 GHz region in the latter case.

The data obtained by the Caltech Submillimeter Observatory radiometer, which monitors the atmosphere by performing frequent skydips, is used to determine current trends and obtain information from past dates, and the information is retrieved from CSO and updated on-line during observing. Although the radiometer operates at 225 GHz it provides an effective record of the opacity at all frequencies, since the opacity at 225 GHz, $\tau_{225}$ can be directly scaled to the opacity at other frequencies, such that $\tau_\nu = k_1 + k_2\tau_{225}$. The constant term $k_1$ is due to O$_2$. It is found⁶ that at 345 GHz $\tau_{345} = 0.05 + 2.5\tau_{225}$, while at the higher frequency sub-mm bands $\tau_{490} \approx \tau_{690} \approx \tau_{820} \approx 20(\tau_{225} -0.01)$. The opacity at 225 GHz at the summit of Mauna Kea is related to the precipitable water vapour $P$ (i.e. pwv) by

$$\tau_{225} \approx 0.01 + 0.04 P$$

so that, for the atmospheric transmission curves given in Figure 1, at $P =$0.5, 1.0 and 5.0 mm pwv, the value of $\tau_{225}$ is 0.03, 0.05 and 0.21 respectively.

An example of the variation of the zenith optical depth at 225 GHz over a period of a few days is given in Figure 2. As can be seen, changes of a factor of two or more within an hour do not seem to be uncommon. See the the CSO Web pages for up-to-date atmospheric records.

SCUBA is routinely used to obtain skydips and hence the atmospheric absorption within its wavebands, and the relationships with the ``CSO tau'' are on a secure footing. In addition, a water vapor radiometer is in routine use at the JCMT. This relies on observations of the 183-GHz emission line of H$_2$O from the atmosphere and subsequent modelling of the line shape to derive the water vapour content. Because this instrument is mounted in the cabin and uses a pickoff mirror to look almost in the same direction as the observing it provides a on-the-spot measurement of the water vapour content in the direction of the current target. Radiometers of this type are installed by HIA on a pair of the SMA antennas, where they have been demonstrably valuable in real-time path delay corrections.

Figure 2: The typical variation of zenith optical depth (tau) at 225 GHz, as measured by the radiometer at the Caltech Submillimeter Observatory, located close to the JCMT, during a period of a few summer days in 1992. Night-time periods are indicated by the closed rectangles on the abscissa. Note (a) the pattern of local day-time increases in water vapour content on some days, and (b) the absence of local daytime disturbances for the last two days of the period. Information courtesy of Caltech Submillimeter Observatory.

In the final column of Table 1, a very subjective estimate is given of the percentage of nights on which observations may reasonably be expected to be rewarding at the given frequency; hence this represents in some way the chance that your proposed observations may have of clearing the last hurdle towards success, having surmounted all the others (e.g. the PATT). These numbers are expected to be somewhat season-dependent; i.e. higher-frequency observations have a better chance of success in the winter-spring, although there have been periods when the summit has been inaccessible due to heavy snowfall (not significantly in the past decade, and suggestive of a climatic change in view of the translation of Mauna Kea as ``White Mountain'').

It is clear from examples such as given in Figure 2 that there ought to be a certain predictability about the behaviour of the atmosphere on timescales of a few days, and we are increasingly trying to `flexibly-schedule' higher-frequency observations with lower-frequency programs to make more effective use of good conditions. In late 1997 and early 1998 a strong ``El Niño'' effect confounded normal weather statistics, resulting in an exceptionally long period of dry weather and transparent skies, and perhaps resulted in unrealistically optimistic expectations for some time thereafter. Conversely, in the past year or more the weather has been worse than average, with occasional periods of excellent transmission. A weak El Niño effect has recently been announced once more; at present it's unclear how this will affect observations in Hawaii. Weather forecasts specific to the summit over a time scale of a few days are issued by the the Mauna Kea Weather Center in Honolulu.

Because the atmosphere is often unfavourable to submillimetre observations during the daylight hours due to anomalous refraction⁷, increased water vapor levels and cloud cover, observations are usually formally scheduled for only 16 hours per day, beginning in the late afternoon (at 1730 HST), and ending at 0930 HST each day. Temperature differentials in the antenna structure acquired during daytime observing can significantly affect the telescope performance not only during the day, but also for some of the following evening. For this reason, when solar and other daytime observations are undertaken each daytime shift is normally counted against the following evening shift, in order to allow the telescope time to thermally relax.

Despite these concerns it can be technically feasible to continue astronomically useful observing throughout the day. Under normal conditions continuous telescope operator support at the JCMT is guaranteed between the hours of 1730 and 0930 HST, and contingent on available staff support it may be feasible to extend this time, where consistent with other demands on telescope time, and having regard to the fact that daytime observing may cause significant degradation of the beam for some hours afterwards. Note that recent changes in staffing arrangements have seriously reduced the opportunity for such extended observing hours, however, and one should have prior contact with telescope staff in advance if interested in any such possibilities.

Observers should note however that the potential for extended shift hours is not at all common; engineering and commissioning work must often be done during the day and usually takes precedence. In addition, approval needs to be obtained from the Director JCMT if extended observing would imply going beyond the PATT-approved program for the present semester. If observations are allowed to continue, the second-shift observer has priority until 1:30 in the afternoon, and the first-shift observer priority thereafter.

2.7 Source Availability

The latitude of the JCMT allows one to observe a large fraction of the southern hemisphere, as well as all of the northern sky. In principle it is just possible to access all of the galactic plane; exceptional nights do in fact allow successful observations at very large airmasses. The trade-off, compared with northern telescopes, is that there are almost no circumpolar sources. The availability of sources as a function of declination is given below in Table 2, and graphically in Figure 3.

Figure 3: A representation of the tracks of sources at selected Declinations as a function of azimuth and elevation for the latitude of the JCMT. The outer circle represents the horizon ( i.e. elevation 0$^\circ$) with elevation increments of 10$^\circ$ shown as dotted circles, with the zenith at the origin of the plot. Azimuth tick marks are shown every 15$^\circ$ around the outside of the plot. Hour angle tracks are shown for a maximum declination of 80$^\circ$(the small ellipse)with an interval of 15$^\circ$, with hour angle marks ($+$ signs) shown every hour. On the plot sources rise at the right side and set (except for circumpolar objects) at the left. Sources with a declination of about 20$^\circ$ pass very close to the zenith. Plot courtesy of Firmin Oliveira.

Note that sources between 15$^\circ$ and 25$^\circ$ declination pass very close to the zenith, and it may not be possible to observe such objects within up to 20 minutes either side of transit. In addition, because of the JCMT's location in `Submillimeter Valley', nearby cinder cones obstruct the view of the sky at azimuths between 0$^\circ$ and 120$^\circ$, and from 230$^\circ$ through to 270$^\circ$, in places rising to almost 10$^\circ$ elevation. Recent construction of telescopes on the ridge above the JCMT further restricts the view at certain azimuths for rising targets.

Table 2: Source Availability above $20^{\circ }$ Elevation

Declination	Max. Elevation	Availability	Declination	Max. Elevation	Availability
(degrees)	(degrees, South)	(hours)	(degrees)	(degrees, North)	(hours)
$-$60	10.2	0.0	20	89.8	10.0
$-$50	20.2	0.8	30	79.8	10.4
$-$40	30.2	5.2	40	69.8	10.6
$-$30	40.2	6.8	50	59.8	11.0
$-$20	50.2	7.8	60	49.8	11.2
$-$10	60.2	8.6	70	39.8	11.4
0	70.2	9.2	80	29.8	11.6
10	80.2	9.6	90	19.8	0.0

The use of Table 2, along with the approximate local sidereal time at 00$^{h}$ HST as a function of date taken from the Astronomical Almanac, should allow effective planning of an observing session. Figure 4 combines some of this information in a chart showing the optimum month in which to observe a source, given its right ascension. An important general fact to note in this context is that the fall-winter semester is the better time to observe sources in the Orion region, while the Galactic Centre is better placed in the spring-summer semester. Sources in the other quadrants of the sky can be observed at some time in either semester equally well.

Figure 4: This chart (based on an original by R.E. Hills) shows which month to observe so that a source at a given right ascension transits at given (local Hawaiian) time. Upper case indicates spring/summer semester (February through July), lower case the fall/winter (August through January) semester.

About this document ...

Overview of the JCMT

This document was generated using the LaTeX2HTML translator Version 2002 (1.62)

Copyright © 1993, 1994, 1995, 1996, Nikos Drakos, Computer Based Learning Unit, University of Leeds.
Copyright © 1997, 1998, 1999, Ross Moore, Mathematics Department, Macquarie University, Sydney.

The command line arguments were:
latex2html -split 0 -show_section_numbers tel_overview.tex

The translation was initiated by Jamie Leech on 2004-11-10

Footnotes

... deformation¹

See e.g. the initial discussions in S. van Hoerner, Astron. J., 72, 35 (1967)

...$f/35$²

A separate hyperboloid mirror attached to the back of the main tertiary mirror and used to directly access the $f/35$ focal position was removed in October 2001.

... pwv³

pwv = precipitable water vapour

... formula⁴

Ruze, J.; Proc. IEEE 54, 633

... calculated⁵

Using the ATM routine; see Cernicharo, J.; ATM: A program to compute theoretical atmospheric opacity for frequencies lower than 1000 GHz; IRAM Internal Report 15-April-1992. This routine does not include ozone lines, which can be quite optically deep, and are usually taken into account automatically in the channel-by-channel calibration of line data where necessary.

... found⁶

See Masson, C.; in Astronomy with Millimeter and Submillimeter Wave Interferometry, ASP Conf. Ser. 59, 87; eds. M. Ishiguro and Wm. J. Welch (1994).

... refraction⁷

Often referred to as ``bad seeing''; i.e. the source position appears to fluctuate by many arc seconds on timescales of tens of seconds. This is thought to be due to blobs of wet air being drawn up the mountain by convection currents. It is generally worse on summer days than winter ones. Similar effects have been observed at lower frequencies (see Altenhoff et. al.; 1988, Astron. Astrophys. 184, 381).