The high image quality at the ORM is now verified throughout an extensive statistical database of reliable seeing values.
In Fig. 3 (click here), results corresponding to 9 months of measurements at site A are presented. The lack of data for December 1995 is due to instrumental problems (coma distortion in the telescope). For each month, the statistical parameters obtained from the seeing value distribution are calculated. The monthly trend of mean, median, minimum and the standard deviation is shown. In order to give an idea of the statistical significance of the values the total number of observations taken every month is presented as well as the percentage of observed hours with the seeing monitor. The corresponding numerical values are given in Tables 1 (click here) and 2 (click here).
As is evident in Fig. 3 (click here) and Tables 1 (click here)
and 2 (click here)
the monthly median seeing never reaches values worse than 1''.
The minimum atmospheric degradation
to image quality can be measured throughout the lowest
seeing values (also
presented in Fig. 3 (click here)). In this
sense it may be pointed out that seeing below 0.3'' is achieved in all
sampled months except March.
From our measurements with the DIMM we estimate 72% of observed hours from the available observing time, 11% being lost due clouds, 6% due to winds higher than 15 m s-1, which is the limit of the instrument, and 8% due to high humidity (misting up of the optical surfaces). This percentage could be considered as a pessimistic estimate of useful observing time at the ORM. Given the fact that our instrument operates in the open air, it is more affected by air humidity changes than other astronomical telescopes. Therefore in this sense we certainly are more limited in the useful hours. Also, the instrument is very sensitive to the presence of even tenuous clouds. Cirrus-like clouds are detected by the DIMM, with the consequent loss of tracking of the source. So with the above-mentioned provisos, the results we obtain are in very good agreement with those reported from the Carlsberg Automatic Meridian Circle (CAMC), taken over several years and compiled by Sarazin (1995), which indicate 79% of useful observing time at the ORM.
A possible seasonal dependence of seeing at the Canarian observatories has been the subject of discussion for a number of years. Sanchez (1970) found a slight tendency for nights of very good seeing at Izaña (Tenerife) to occur in summer between April and October (1962-1966). At La Palma, Pike (1984), from data taken at the telescope focus, found the summer months to show an average seeing better than during the winter. The accuracy of the seeing monitor together with the temporal sampling of data presented here make them extremely useful in deciding this question.
From Fig. 3 (click here) it is clear that the seeing abruptly improves between May and June. In Fig. 4 (click here) it is shown results from a comprehensive study of the climatology of the Canary Islands (Font-Tullot 1956). In this plot, using data covering 20 years compiled by the Instituto Nacional de Meteorologıa at Tenerife, the annual variation of the scale height of the sea-cloud (inversion layer) is shown.
Figure 4: Annual variation of the scale height and
frequency
of the sea clouds in Tenerife. Taken from Font-Tullot (1956)
The existence of a temperature inversion layer is a characteristic of subtropical regions, and in the Canaries it is registered about 90% of the time. The efficiency of the inversion layer in separating the boundary layer from the troposphere is clear on the humidity profile, where typically 55% of the humidity is situated below the inversion and 20% above it (Cuevas 1995). The inversion suppresses local convection, which is clearly visible in the presence of a characteristic stratocumulus layer whose top is just below the temperature inversion (Font-Tullot 1956). The frequency of occurrence of the cloud layer is maximum in summer and is anti-correlated with its height (Fig. 4 (click here)). The width of the inversion layer is greatest during the summer period.
The strength of the inversion layer (IL) is defined by the difference of temperature between the top and the base of the temperature inversion, and is maximum during the summer, as is the intensity of the trade-wind regime. The strong temperature inversion layer, and therefore the presence of the sea clouds, is caused by two complementary processes, the trade winds blowing from the NE carrying wet and cool air masses and the subsidence from the NW which occurs above the IL transporting dry air (Cuevas et al. 1996).
Figure 5: Comparison of the statistics corresponding to
winter and summer at site A
Despite the fact that the inversion layer tends to be well below the observatory location, it seems that a dependence of image quality exists, which is correlated with the height of the inversion layer (the sea-cloud location). The best seeing values seem to correspond to months where the altitude of inversion layer is lower than normal (in the 1500-1700 m range), or when the strength of the IL is maximum, which in turn is related to the strength of the trade wind regime.
In Fig. 5 (click here) seeing data are grouped following the above discussion in two blocks, one including June, July and August (summer) and the other containing the rest.The distribution function corresponding to summer gives better mean and median values and smaller stardard deviation. In Fig. 6 (click here) (top), the seeing distribution and the cumulative distribution function corresponding to June 1995 (best seeing) at site A are shown. The attained values are remarkable. The seeing is better than 0.5'' for 60% of the time and better than 0.3'' in 7% of cases. The mean and median values are very similar (0.49'' and 0.46'', respectively) with 0.17'' stardard deviation. These values are very good even when compared with other astronomical sites known for their excellent seeing (see Sarazin 1995 for references). In Fig. 6 (click here) (bottom) the statistics corresponding to all data taken at site A are shown. Half of the time seeing values are better than 0.64'' with minimum values down to 0.17''.
Figure 6: (Top) seeing distribution and cumulative
distribution
function corresponding to June 1995 at site A. (Bottom) seeing distribution
and
cumulative seeing distribution from all data taken at site A