As explained previously, the mesospheric sodium has long term (seasonal) but also short term (daily or hourly) variations. The last ones cannot be studied statistically, since they are random. Once the laser power of your LGS AO system has been determined, the interest of sodium monitoring on a daily basis is:
In the following, the different requirements for such monitoring will be presented first, followed by a concise mention of on-going monitoring experiments.
The first constraint is directly linked to the sodium itself. The D1 and D2line are respectively at 5895.94 Å and 5889.773 Å. The last one having an hyperfine structure with a separation of 19 mÅ.
Numerous atmospheric water vapour lines occur in the region of the spectrum around 6000 Å, and two lines in particular (5889.637 and 5890.09 Å) are very close to the rest wavelength of the sodium D2 line but also much stronger. The problems associated with telluric line contamination of sodium absorption spectra have well been documented by [Hobbs] (1978). This implies that any instrument of interest must have a minimum resolving power of 50000 to resolve the D1 line from the closest water line. For the D2 line, the minimum resolving power requested increases to 130000. With a resolution of one million, one is then able to resolve the hyperfine structure of the D2line. The other point to consider is the equivalent width of 1 mÅ of the atmospheric Na D1 line. This means that a signal to noise of 1000 is necessary to detect it with a 15% accuracy.
Using e.g. the CES at the La Silla ESO 3.6 m telescope, with a resolution of 220000,
one finds that a signal to noise of 1000 can be reached in 9 min for 
car (V=1.7)
and in 2.5 hours for 51 Oph (V=4.8). At best, with the brightest star, only
"long term sporadics'' can be studied. The short-life sporadics (few minutes at most)
cannot be detected by this technique!
However, [Patriarchi & Cacciani (1999)] are offering a solution to make short time integrations to monitor the mesospheric sodium layer column density with small aperture
(
50 cm) telescopes.
Their system is based on a magneto-optical filter already widely used in solar observations
([Cacciani et al. 1994]). This filter is compact, stable in wavelengths and profile shape. It displays a very narrow spectral transmission band (down to
70 mÅ) and also has a high peak transmission (up to 40%). Presently no results of this technique have been presented.
The group of R. Angel (Tucson, Az) has been doing atmospheric Na
measurements for several years in relation with their LGS-AO
experiments at the MMT ([Lloyd-Hart et al. 1995]; [Lloyd-Hart et al. 1998]).
Recent observations have been aimed at measuring the sodium column density simultaneously with the LGS visible magnitude,
which has a direct bearing on the expected efficiency of the
LGS adaptive optics correction. Ge et al. (1997) found an
equivalent mV =10.3 star for the LGS, using a 1W projected circularly
polarised sodium laser beam for an atmospheric sodium column density of
cm-2.
They made several successful measurements during March and
May 1997, monitoring similar sodium layer patches, during the simultaneous return experiments.
The MMT experiments established the important result that linearly and circularly
polarized laser returns are proportional to the simultaneous sodium column
density (see Fig. 4 of [Ge et al. 1998]). Moreover, circularly polarized laser
provides 
30% increase in fluorescent return over linearly
polarized laser ([Ge et al. 1997]).
Together with the distributed column density measurements (e.g. seasonal and diurnal variations), it is now possible to estimate laser power requirements for any specified guide star brightness, which is a crucial result, in order to determine the choice of laser type and power for the laser guide star AO systems.
Other groups related to existing or planned LGS-AO systems recognise the importance of studies as described above. Most of them are thus thinking of monitoring the mesospheric sodium but mostly using existing fully equipped nearby telescope, to realise spectro-photometric observations.
| LGS parameter | Atmospheric Na layer characteristics | Means of Observations |
| LGS power | Seasonal variations | Statistical study |
| (spectroscopy) | ||
| LGS focus | Long Term variations | Statistical studies |
| Short term variations (< 1 min) | Monitoring (lidar) | |
| Scheduling of observations | Daily & hourly variations | On-site statistics |
| Monitoring (lidar) |
In order to monitor the atmospheric sodium as efficiently as possible, one would like to have short exposure time. For over 30 years (Ageorges et al. 1999 and reference therein), lidar technology has been used to study the atmosphere and proved its superiority over other techniques. It gives the most information in the least time.
Presently the best solution for short measurements of the Na column density is via lidar technology. These can be quite portable and with a 20 Hz 100 mJ pulse at 590 nm, one can record the sodium layer in about 10 s, in clear nights, with a 2.7 m collector ([Wuerker & Wong 1997]). A 1 m collector, would mean 10 times longer integration time, which is still largely shorter than what is reachable presently with high resolution spectroscopic studies. Another advantage of lidar techniques is the determination of the profile of the sodium layer. The main drawback is the cost of at least few $10000 for the laser. However "low cost'' lidar experiments have started recently ([Michaille et al. 2000]).
The normal lidar involves transmitting short monochromatic light pulses and measuring the time-of-flight of the photons (total of emitted and backscattered) in order to provide information about the scattering medium. It has been proposed to modify the Max Planck for Extraterrestrial physics ALFA (Adaptive optics with Laser For Astronomy) laser ([Quirrenbach et al. 1998]), at work in Calar Alto (Spain), in order to use it as a lidar. ALFA is a CW laser. It is proposed to transform it by introducing an extra-cavity acousto-optic modulator in the outgoing collimated beam. In these experiments, the 3.5 m telescope itself will be used as a collector. The detection will be made on a high quantum efficiency avalanche photo-diode. Each detected photon returned will be time tagged in a multi-channel scaler memory syncronised to the laser modulation.
It is important that such an instrument be designed so that the Na profile
can be measured as quickly as possible, with minimal time lost for
astrophysical observations.
The final purpose is to get a world unique operating system that allows to do both LGS AO observations and simultaneously to monitor the atmospheric sodium and thus correct on-line for its variations (e.g. focus). This experiment actually took successfully place in October
1999 ([Butler et al. 2000]). The next step is now to use the laser launch telescope as a
"lidar collector''.
To summarise, presently, the Na monitoring experiments performed in conjunction with LGS-AO observations have been aimed at determining simultaneously the LGS intensity and the atmospheric Na column density in order to derive information on the LGS efficiency. This is possible because monitoring the laser power and estimating the amount of Na present in the atmosphere at the moment of the observations (monitoring), enable to predict the return from the laser (magnitude of the LGS), only under the assumption that no saturation occurs in the sodium layer. A by-product of this study is the monitoring of the LGS-AO itself, in terms of quality of correction under known observing conditions in comparison to a theoretically perfect system.
What is currectly missing are statistical values for the variations of the centroid of the atmospheric Na height. Only these statistics would determine the real necessity of an atmospheric sodium monitor. Indeed, if the average altitude's height variations are small compared to the WFS error budget, they will then not affect the focus of the LGS. If however they are stronger, two things can happen: their frequency is higher than 0.016 Hz, then they can be corrected by defocusing the LGS WFS; or the variations happen on timescale much smaller than 1 min, then a separate sodium monitor is really necessary to identify these periods for which the AO correction will be poorer.
Copyright The European Southern Observatory (ESO)