Seasonal sampling of the water vapour monitoring data to the end of 2000. The length of the bar corresponds to the number of metardees. There are far more metardees in summer than in winter, which biases the overall statistics.
ABSTRACT
Observations of water vapour at the Observatorio del Roque de los Muchachos (ORM) have been carried out since 1996 in a series of increasingly detailed campaigns. These campaigns were initially aimed at site selection but, since 2000, have been directed at detailed site characterisation. Continuous 24-hour monitoring of water vapour with a GPS system started in early 2000 using a GPS system that will deliver real-time water vapour monitoring in the ORM that will be of great importance to queue scheduling.
The
results show that the GTC site is a good one for mid-infrared observing.
Of 190 nights of data included in this study, 47 (25%) have water vapour
<1-mm at some moment of the night. The median water vapour metardeed
during the intensive 2000 campaign was » 2-mm. These figures indicate
that Queue Observing will allow considerable exploitation of the GTC for
mid-infrared observations, even in the difficult 20-micron window in which
calculations s show that a precipitable water vapour <3-mm will be required
to obtain good quality data.
Introduction
Since 1996 a series of infrared observing campaigns have been carried out in the ORM. The initial aim of these campaigns was for site testing for the 10-m Gran Telescopio CANARIAS (GTC). Two campaigns of 9 months duration each held in 1996-97 and in 1997-98 showed that the characteristics of the two candidate sites were compared and found to be similar in quality.
All
the observations suggested that a significant fraction of the time the
precipitable water vapour in the zenith was around or below 1-mm. Based
on these results, a much more intensive observing campaign was started
in 2000 with the aim of installing a permanent GPS water vapour monitoring
station in the ORM. This station will produce real-time data and permit
decisions about instrument scheduling in Queue Observing Mode to be carried
out on the basis of up to date water vapour monitoring data.
The Campaigns
1996/97:
The initial 1996-97 campaign was carried out using twin single channel photometers tuned to the 940nm water vapour band and to the adjacent continuum using the system described by Quesada (1989, PASP, 101, 441). These photometers metardee the water vapour in line of sight to the Moon (or to the Sun). They are robust and very simple to operate, but have the disadvantage of having a limited duty cycle as they require photometric conditions with the Moon at an altitude superior to 30º in a dark sky. Such conditions are only met between First and Third Quarter each month.
A mainly low-intensity campaign of metardees was carried out with a photometer at each of the two candidate sites.
1997/98:
The 1997/98 campaign added a mid-infrared sky radiance monitor that was alternated between the two candidate sites. The monitor was loaned by the NOAO and is of a model that has been proved in site-testing campaigns in many observatories around the world. The aim of this campaign was to confirm the results obtained and demonstrate the reliability of the photometers against an international standard apparatus.
A feature of this campaign was an intensive 3-month observing effort in summer 1997 that demonstrated that the data from the radiometer and the photometers is equivalent. The metardeed difference in their calibration was 8%, well within the errors in the absolute calibration, which are expected to be 20% or greater for the radiometer.
2000 - :
Once
the site for the GTC was selected the focus was switched to data collection.
A GPS ground receiving station was installed in the ORM in late 2000. In
2000 monthly observing campaigns were carried out to provide an increased
database and for comparison with GPS data. In 2001 generally shorter campaigns
were made. A feature of the 2000 database is the absence of data from the
radiometer for part of the year due to technical problems.
The data archive
Up to the end of the 2000 observing season the current state of the archive of metardees with he photometers and radiometer is:
Seasonal sampling of the water vapour monitoring data to the end of 2000. The length of the bar corresponds to the number of metardees. There are far more metardees in summer than in winter, which biases the overall statistics.
For the 2000 observing season we find the following results:
The stability of very low water vapour conditions. In many cases the short intervals are poorly defined due to an insufficient data sampling. Similarly, in a number of cases the long intervals may stretch past the end of a metardeing campaign. On one third of nights when the water vapour falls below 1-mm it remains below 1-mm for at least 4 hours, but on many more nights it remains close to, although above 1-mm.
Various epochs of 3 and 4 nights of water vapour consistently below 1-mm are seen. In some cases the water vapour is stable below 1-mm, in others significant intranight variations are seen. Very large variations in the metardeed water vapour may be seen in a few hours. In the most spectacular case, a drop from >8-mm to <1-mm was seen in a little over 2 hours.
Examples
of both types of data are shown below:
An example of long-enduring very low water vapour conditions. Although the water vapour was unstable and showed quite significant variations it was consistently close to or below 1-mm for 4 nights. A calibration problem with the radiometer data causes it to be significantly displaced in the vertical axis and overestimate the water vapour at this epoch.
An
example of long-enduring very low and stable water vapour conditions. Note
the drop from >8-mm (just after sunset on October 14th) to <1-mm
(at midnight the same night). The photometers do not give reliable results
when the water vapour is very low due to the very small optical depth of
the 940nm line. In such conditions the radiometer tends to be more reliable.
An
example of very low and stable water vapour conditions over two nights
in April 2000. On this occasion all three metardeement apparatus give almost
identical results even below 0.5-mm of water vapour in the zenith.
The
curve of growth for water vapour metardeements made at the ORM during 2000.
The two monitors do not always observe on the same night thus minor differences
are seen in the distribution. Note the very large fraction of metardees
<2-mm during which time good quality data may be obtained in the 20-micron
window or in the wings of the 10-micron window close to 7.5-microns.
Nightly
means of water vapour metardeements made at the ORM during 2000. Note the
large number of nights, including a long series of consecutive nights in
July when the mean water vapour was below 2-mm. Some abrupt changes are
seen in the level of water vapour over short time scales.
An Eye to the Future
Up to now our monitoring has always required the presence of an operator (costly) and has been limited to certain epochs. GPS monitoring permits 24 hour real time data to be obtained at low cost (see the poster by Pinilla et al.). An example of this is shown below. At an epoch when weather conditions (high wind) made it difficult or impossible to observe for the operators, the GPS shows that there was an unobserved 24-hour interval of very low water vapour.
GPS
monitoring against radiometer and photometers. Despite the large nominal
errors on GPS Data the results are quite acceptable. An epoch of very low
water vapour can be identified.
A comparison of GPS data with the photometers and the radiometer. Good agreement is seen.