next previous
Up: Evaluation of Devasthal site


5 Characterization of Devasthal site

As a follow up on the site survey reported above, precipitable water vapour, meteorological and astronomical seeing measurements were carried at Devasthal sites. Primary emphasis was placed on the evaluation of ground level seeing at the two sites mentioned above. Simultaneous measurements of atmospheric parameters such as air temperature, soil temperature, solar radiation, relative humidity and wind speed and direction as well as microthermal measurements were performed at Site 1. These parameters are of major importance in the interpretation of the occurrence of optical turbulence.

5.1 Estimation of precipitable water vapour

The precipitable water vapour content in the atmosphere was measured with the instrument designed and built by Prof. Westphal and kindly loaned to UPSO. The observations were taken between January 1989 to June 1989 at Devasthal. They indicate 1-2 mm, 2-3 mm and > 3 mm of precipitable water vapour for 21%, 43% and 36% of the time respectively.

5.2 DIMM instrumental setup and observations

This section describes the DIMM instruments which were built in UPSO laboratories and used for seeing measurements along with the procedure used for data processing. Differential image motion technique has been used as early as 1960 to provide quantitative seeing estimates. Such work carried out in recent years include those by the European Southern Observatory (Sarazin & Roddier 1990), the National Optical Astronomical Observatories (Forbes et al. 1988), the ESO site testing measurements for the Very Large Telescope (Pedersen et al. 1988), Observatorio del Roque de los Muchachos (ORM ) at La Palma (Vernin & Muñoz-Tuñón 1995) etc. for the evaluation of potential observing sites. The principles in detail of DIMM instrument can be found in Sarazin & Roddier (1990).

A 52 cm telescope with an equatorial fork mounting having a plate scale of 31 $^{\prime\prime}$/mm at the f/13 folded Cassegrain focus is used at Site 1, whereas at Site 2, a 38 cm telescope with single pier mounting having a plate scale of 36 $^{\prime\prime}$/mm at the f/15 Cassegrain focus is used. The front portion of the 52 cm reflector tube is covered by a mask which has two circular holes, each of 6 cm diameter and separated by 40 cm. Similarly, the 38 cm reflector is covered by a mask having two circular holes each of 5 cm in diameter and separated by 24 cm. One of the holes contains a prism which deviates the incoming parallel light of a star by about 30 $^{\prime\prime}$ in the direction joining the line of the centers of the two holes, so that two images of the same star are formed on the CCD detector. The telescopes are equipped with a PC which controls the Santa Barbara Instrument Group (SBIG) ST4 autoguiding CCD camera and thus accumulates image motion data and analyse them online to provide seeing measurements in two mutually perpendicular directions. One pixel of the CCD corresponds to 0 $.\!\!^{\prime\prime}$42 $\times$ 0 $.\!\!^{\prime\prime}$49 at 52 cm telescope and 0 $.\!\!^{\prime\prime}$50 $\times$ 0 $.\!\!^{\prime\prime}$58 at 38 cm telescope. Our setup closely follows that of Wood et al. (1995). The details of the instruments used in the seeing campaign and their characteristics are given in Table 2.

In DIMM, a series of exposures each spanning a period of 10 ms are taken and hundred such exposures are used to derive one estimate of the standard deviation of the relative separation of the images, $\sigma_{{\rm l}}$ and $\sigma_{{\rm t}}$ parallel and perpendicular to the line joining the centres of the two apertures. Using Eqs. (3) and (4) one can then find independent estimates of r0. These independent values of r0 are then converted into seeing estimates using Eq. (1) and a wavelength of 5000 Å. All the above processes are done online and the data are stored in the PC for further analysis. The data obtained using DIMM were then corrected for airmass (X) using the relation

\begin{displaymath}FWHM_{{\rm corrected}} = FWHM_{{\rm observed}}\left(1/X\right )^{3/5} .
\end{displaymath} (5)

The use of this correction is necessary as it allows comparisons to be made with observations obtained at different time in various directions.

5.3 Estimation of errors

5.3.1 System noise

The uncertainty in the determination of centroid of an image due to detector noise introduces an error in the seeing measurements. Since this instrumental noise is not correlated with the true nature of the atmosphere, this has to be subtracted out before an estimation of the seeing is made (cf. Sarazin & Roddier 1990). The estimation of the instrumental noise was carried out experimentally in the optics lab of UPSO. The intensity levels of the images were kept similar to those that we maintain in performing seeing measurements. The uncertainty measured in the separation of the images of two pin holes on the CCD detector gives a value of 0.09 pixel. This system noise when taken into account improves our seeing reported here by 0 $.\!\!^{\prime\prime}$01 only.

5.3.2 Statistical errors

The statistical errors of the seeing measurements were calculated using the formalism given by Frieden (1983). It has been pointed out by Sarazin & Roddier (1990) that the statistical properties of the atmosphere does not change for about a minute. Under typical seeing measurements at Devasthal we are able to process about 50 images in a minute. The number of images obtained in a minute is limited by the exposure time as well as the readout time of the CCD. For our measurements the statistical error turns out to be 12%, and the error for each individaul measurement is about 9%.
  \begin{figure}\includegraphics[width=8cm]{ds8858f4.eps}\end{figure} Figure 4: Seeing at Devasthal Site 2 plotted against UT. Other details are the same as in Fig. 3

Table 3: Results of seeing measurements at Devasthal
...3 & 1.1 \\
\noalign{\smallskip }

  \begin{figure}\includegraphics[width=8cm]{ds8858f5.eps}\end{figure} Figure 5: Cumulative distribution and histogram of seeing measured during the indicated month at Devasthal Site 1. The solid and dotted lines are for the year 1997 and 1998 respectively

  \begin{figure}\includegraphics[width=8cm]{ds8858f6.eps}\end{figure} Figure 6: Cumulative distribution and histogram of seeing measured during the indicated month at Devasthal Site 2

  \begin{figure}\includegraphics[width=8cm]{ds8858f2.eps}\end{figure} Figure 7: Cumulative distribution and histogram of entire seeing measurements obtained at UPSO and Devasthal sites

  \begin{figure}\includegraphics[width=8.8cm]{ds8858f8.eps}\end{figure} Figure 8: Histogram and cumulative distribution of the wind speed at Devasthal Site 1

  \begin{figure}\includegraphics[width=5cm]{ds8858f9.eps}\end{figure} Figure 9: Distribution of wind speed and direction at Devasthal Site 1

  \begin{figure}\includegraphics[width=8.8cm]{ds8858f10.eps}\end{figure} Figure 10: Monthly variation of average relative humidity at Devasthal Site 1. The solid and dotted lines represent the relative humidity during day and night time respectively

  \begin{figure}\includegraphics[width=8.8cm]{ds8858f11.eps}\end{figure} Figure 11: Histogram and cumulative distribution of relative humidity at Devasthal Site 1. The solid line is for the whole data set, whereas the dotted line is for the data set excluding the monsoon months

5.4 Comparison of longitudinal and transverse seeing

For ESO DIMM the seeing values obtained from $\sigma_{{\rm l}}$ and $\sigma_{{\rm t}}$agree with each other within 12% (see Sarazin & Roddier 1990). An attempt has been made by us to see the behaviour of longitudinal and transverse seeing for the observations at Devasthal sites. The seeing measurements were carried out for 88 nights at Site 1 during February 1997 and November 1998, and for 37 nights at Site 2 between October and December 1998. In Fig. 2 we have plotted the longitudinal seeing against the transverse seeing for both Devasthal sites. For comparison, we have drawn a line of unit slope and zero intercept. This indicates a fairly good agreement between the two measurements considering the errors discussed above. Amongst the two independent measurements, we have therefore considered only one, namely the longitudinal component as the seeing of the site. There is however some difference between the values obtained from the two components of image motion; wind speed and the finite outer scale length will affect the correlation between the two apertures and could lead to this difference (see Das et al. 1999).

5.5 Seeing measurements

Seeing measurements were carried out for 88 and 37 nights at Site 1 (during February 1997 to November 1998) and Site 2 (between October and December 1998) respectively. We have plotted in Fig. 3 the nightly variation of seeing against UT for Site 1 while the same for Site 2 has been plotted in Fig. 4. The statistical parameters obtained from seeing measurements for each month such as mean, median etc. are given in Table 3. They indicate that the seeing is relatively better during April, May and June as compared to other months. This tells us about a possible seasonal dependence of seeing, with summer months showing an average seeing better than other months. The monthwise histogram and cumulative distribution of seeing observations at Devasthal Site 1 and Site 2 are shown in Figs. 5 and 6 respectively.


Table 4: Seeing statistics at UPSO and Devasthal sites

UPSO Site 1 Site 2

Total no. of nights observed (datapoints) 18 (698) 88 (3698) 37 (6798)
Minimum seeing ( $^{\prime\prime}$) 0.5 0.5 0.5
Average seeing ( $^{\prime\prime}$) 1.6 $\pm$ 0.4 1.5 $\pm$ 0.4 1.2 $\pm$ 0.3
Median seeing ( $^{\prime\prime}$) 1.6 1.4 1.1
Percentage of data with seeing $\le$ 1 $.\!\!^{\prime\prime}$0 8 7 40
Percentage of data with seeing 1 $.\!\!^{\prime\prime}$0 - 1 $.\!\!^{\prime\prime}$2 10 16 26
Percentage of data with seeing 1.2'' - 1 $.\!\!^{\prime\prime}$4 16 22 17
Percentage of data with seeing 1.4'' - 1 $.\!\!^{\prime\prime}$6 18 21 09
Percentage of data with seeing 1 $.\!\!^{\prime\prime}$6 - 1 $.\!\!^{\prime\prime}$8 17 12 04
Percentage of data with seeing 1 $.\!\!^{\prime\prime}$8 - 2 $.\!\!^{\prime\prime}$0 12 10 2
Percentage of data with seeing > 2 $.\!\!^{\prime\prime}$0 18 12 2


5.5.1 Comparison with test observations at UPSO

A series of test observations with the 52 cm telescope, which was later shifted to Devasthal Site 1 was carried out for 18 nights at UPSO, during November and December, 1996. Figure 7 shows the histogram and cumulative distribution of seeing data obtained at UPSO which indicates a median value of 1 $.\!\!^{\prime\prime}$6. For comparison, corresponding plots for Devasthal sites are also shown in Fig. 7. Table 4 gives the overall statistics of the seeing values for the entire observing run at UPSO and Devasthal sites.

5.5.2 Comparison of seeing at both Devasthal sites

A Comparison of seeing measured quasi-simultaneously at both Devasthal sites on 8 nights during October and November 1998, is presented in Table 5. The differences in median seeing values range from 0 $.\!\!^{\prime\prime}$16 to 1 $.\!\!^{\prime\prime}$24, with better seeing values always at Site 2. Figure 7, Tables 3 and 4 also indicate that Site 2 has better seeing values than Site 1. It is worth mentioning here that at Site 2 in comparison to Site 1, for significantly large fraction ($\sim$ 40%) of the observing time seeing is < 1 $^{\prime\prime}$ (see Table 4). Another factor which favours Site 2 is that the seeing measurements were done at about 2 m above the ground level whereas at Site 1 it was carried at about 4 m above the ground. Studies by Pant et al. (1999), Vernin & Muñoz-Tuñón (1994) and Avila et al. (1998) indicate appreciable degradation in seeing due to turbulence introduced by the surrounding trees, local topography and ground radiation. The seeing improves considerably at height $\ge$ 10 m above the ground as discussed below.

5.6 Microthermal measurements

The detection of the local source of seeing degradation which occur in levels of the atmosphere very near the ground, within a few tenths of metres above ground is of great importance for evaluating the seeing conditions of an astronomical site. Microthermal fluctuations at Devasthal Site 1 were therefore recorded using pairs of microthermal sensors made from Nickel wire of 25$\mu$ in diameter separated by a distance of 1 m and mounted at three different levels on a mast situated respectively at 6, 12 and 18 m above the ground. A description of the experimental setup and the methods employed in deducing the seeing estimates from the microthermal measurements can be found in Pant & Sagar (1998) and Pant et al. (1999). Observations of the surface layer contribution to seeing were carried out during March to June 1998. The results of these observations have been reported by Pant et al. (1999) where they found that the major contribution to seeing comes from the 6 - 12 m slab of the atmosphere and sub-arcsec seeing of 0 $.\!\!^{\prime\prime}$65 and 0 $.\!\!^{\prime\prime}$5 can be achieved by locating the telescope at a height of $\sim$ 13 m and 18 m above the ground respectively. Similar conclusion about the degradation of image due to surface layer turbulence is also given by Vernin & Muñoz-Tuñón (1994). There is a plan to carry out similar measurements at Devasthal Site 2 also.

  \begin{figure}\includegraphics[width=8.8cm]{ds8858f12.eps}\end{figure} Figure 12: Monthly variation of average air temperature at Devasthal Site 1. The solid and dotted lines represent the diurnal and nocturnal air temperature respectively

5.7 Meteorological all weather station

Measurements of atmospheric parameters such as temperature, humidity, wind etc. simultaneous with DIMM and microthermal observations are of major importance in the interpretation of the atmospheric turbulence. To accomplish this a meteorological Automatic Weather Station (AWS) is also established close to the 52 cm telescope and the microthermal tower. The instrument is from M/s Champbell Scientific Inc. from U.S.A. It contains wind speed sensors, wind direction sensors, pyranometer for solar radiation measurement, temperature sensors for sensing the air and soil temperature, electronic tripping bucket rain gauge, relative humidity meter, solar panel for charging a 12 volt battery and data logger.

All the above instruments are mounted at the appropriate places on a mast fitted in a tripod. The tripod is kept upright and oriented, so one leg points due south, and then plumb the mast by adjusting south and northeast facing legs. The tripod including the mast is grounded with the help of a ground rod.

The instruments and the data logger are connected to a 12 volt battery to record the above mentioned meterological data automatically in a module. The data are stored in the module at one hour interval through programming. The capacity of the module is to store data for about a month. After a month the module is taken out of the data logger for processing and another module is connected for further recording. The meteorological data thus obtained are useful to see if there is any effect of these parameters on seeing. The histogram and cumulative distribution of day and night wind speed at Site 1 are shown in Fig. 8. It is found that both diurnal and nocturnal wind speed are less than 3 m/s for 85% and 87% of the time respectively. The prevailing wind direction is mostly NW and it seldom exceeds 10 m/s. This is in agreement with our earlier observations(see Sect. 2.1.4). The distribution of wind speed and direction is shown in Fig. 9. The monthly statistics of average relative humidity are shown in Fig. 10 for both day and night. Figure 11 shows the histogram and cumulative distribution of relative humidity for the whole observing period. Both day as well as night time measurements indicate that 55% of the time, relative humidity lies below 85%. However, this fraction increases to 70% if we exclude monsoon months from July to September. The monthly variation of nocturnal and diurnal average air temperature is given in Fig. 12. The difference between day and night average temperature is less than 3$^\circ$C for all the months except March, April, May and June, where it is somewhat higher but still less than 6$^\circ$C.

  \begin{figure}\includegraphics[width=8.8cm]{ds8858f13.eps}\end{figure} Figure 13: Average seeing of 30 min duration versus UT for the entire observing period at Devasthal

Table 5: A comparison of median seeing values measured at both Devasthal sites. Number of measured data points are given inside bracket
...3(175) \\
\noalign{\smallskip }
\end{array} \end{displaymath}\end{table}

  \begin{figure}\includegraphics[width=8cm]{ds8858f14.eps}\end{figure} Figure 14: Dependence of seeing on total solar radiation, wind direction, wind speed, air temperature and relative humidity at Devasthal Site 1

5.8 Temporal evolution of seeing

From the point of view of astronomers, it is useful to have a characterization of the temporal evolution of seeing, i.e., the variation of seeing quality with time. The results of all observing nights are averaged in 30 minutes bin and plotted against UT in Fig. 13 for Site 1 and Site 2. No significant trend is seen. This is in contrast with the generally prevailing notion among astronomers that seeing is poorer in the beginning of night and improves later in the night. Our results for Devasthal is in agreement with what has been found for the ORM site at La Palma by Muñoz-Tuñón et al. (1997), where no general trend in the seeing evolution is noticed.

5.9 Dependence of seeing on meteorological parameters

We have checked for possible relationships between the meteorology and the image quality for Devasthal Site 1. In Fig. 14, we have plotted the variation of seeing with meteorological parameters namely total solar radiation, wind direction, wind speed, air temperature and relative humidity. We observe that relative humidity does not seem to affect the seeing significantly. Weak correlation is noticed between air temperature and seeing, in the sense seeing improves at higher air temperature. A marginal dependence of seeing on total solar radiation is noticed with sunny days having better seeing. These are consistent with results given in Table 3, where seeing is better in summer months when air temperature is higher than that for rest of the year. The interesting plot concerns the wind speed and direction following the belief that these have an influence on the seeing. Correlation of wind speed is expected but not for low wind speed as the turbulence for wind speed less than 15 m/s is relatively unimportant (Woolf 1982). Contrary to this, a weak correlation is noticed between seeing and wind speed and direction.

5.10 Stability of seeing at Devasthal Site 2

The stability of seeing with time tells us about the quality of the site To get a knowledge of the same we carefully looked into individual nights having more than 6 hours of continuous seeing data. This has been done only for Site 2 as it is better than Site 1. Among the 37 nights of observations at Devasthal Site 2 we have 25 nights which satisfied this criterion. The time during which seeing is less than 1 $.\!\!^{\prime\prime}$0 in a stretch for >4 hour, 2 to 4 hour and < 2 hour are found to be 10%, 47% and 43% respectively.


Table 6: Comparison of present seeing results with those of other sites
Site/Observatories Seeing ( $^{\prime\prime}$) Instrument Source
International sites      
Mt. Graham, Arizona, U.S.A. 0.60 STTa Cromwell et al. (1998)
Mt. Hopkins, Arizona, U.S.A. 0.59 STT Cromwell et al. (1990)
Las Campanas, Chile 0.60 CMb Persson et al. (1990)
La Silla, Chile 0.87 DIMM Murtagh & Sarazin (1993)
Paranal, Chile 0.64 DIMM Murtagh & Sarazin (1993)
Mauna Kea, Hawai, U.S.A. 0.57 SCIDARc Roddier et al. (1990)
La Palma, Spain 0.64 DIMM Muñoz-Tuñón et al. (1997)
SPM, B. C., Mexico 0.61 STT, CM Echevarria et al. (1998)
Freeling Heights, Australia 1.20 DIMM Wood et al. (1995)
Siding Spring, Australia 1.20 DIMM Wood et al. (1995)
Indian sites      
UPSO 1.57 DIMM This work
Devasthal Site 1 1.44 DIMM This work
Devasthal Site 2 1.07 DIMM This work
Leh 1.07 DIMM Bhatt et al. (2000)
IUCAA, Pune 1.50 DIMM Das et al. (1999)

a Site Testing Telescope.
b Carneige Monitor.
c SCIntillation Detection And Ranging.

5.11 Extinction measurements at Devasthal

Routine extinction measurements in standard passbands were carried out using the 52 cm reflector equipped with a solid state SSP3 photometer. The lowest value of extinction was observed on 20 and 21 February, 1998. They are 0.40 $\pm$ 0.01, 0.22 $\pm$ 0.01, 0.12 $\pm$ 0.01 and 0.06 $\pm$ 0.01 in Johnson U, B, V and R bands respectively. A detailed paper on the atmospheric extinction measurements at Devasthal is published elsewhere (cf. Mohan et al. 1999).

next previous
Up: Evaluation of Devasthal site

Copyright The European Southern Observatory (ESO)