Jon's astropages

Light pollution in Crete is moderate, but increasing. All touristified areas have an abundance of poorly designed lights which shine most of their light sideways and upwards.

Transparency at sea level is almost always poor, especially in the summer. Below about 1000 - 1500 m altitude there is haze all summer and autumn, and often in the winter and spring as well. The haze is composed mainly of aerosols from various pollutants, sea salt grains, and dust (Science 298). Actually, the Mediterranean air is surprisingly polluted, and most of the pollution is transported from western and eastern Europe and Asia. Smoke from local burning waste dumps can also be annoying on nights when an inversion layer develops. A few times a year large amounts of Saharan dust is carried by strong winds from south. Naturally all this airborne stuff is not good for telescope optics, and particular care has to be taken to keep the optics clean.

At an altitude of 1750 m at the Skinakas observatory, you are above most of the light pollution and haze, and the sky can be very dark indeed. The 1.3 m telescope there is currently the largest in Greece. The site is closed for the public except a few sundays a year around full moon when they have open days. Watch out for the crazy Anogians :-). In the winter and spring the area is inaccessible because of snow.

A hazy evening on Crete. Looking north at the Psiloritis mountain. This evening most of the haze lies below about 700 m, but some is visible up to about 2000 m altitude. Click on the image for full size.

Seeing
The astronomical seeing in Crete can be good at times, but during much of my stay it was poor to mediocre, at least from the balcony of my apartment in Gournes (ca. 50 m altitude). Atmospheric seeing is due to mixing of air layers with different temperatures. This mixing principally occurs in three layers:

1. In the jet stream, which is centered at about 12 km height, or a little higher at this southern latitude. The jet stream intensifies and extends southwards in the winter. Actually, during much of the winter the jet stream is less intense further north than Crete, so in this case the seeing is probably better further north in Europe.

Jet stream observations at the 300 milibar level. Click on the images for full size. Modified from CRWS.

2. Mid-altitude turbulence result from friction between large objects like mountains and the moving air. Mountains can cause turbulence further than 100 km downwind. Wind from west forces the air over two mountain ranges, and seem to cause medium levels of turbulence at my downstream location.

3. Close to the ground, mixing occurs as the ground cools during the night. The ground cooling during night seems to be causing most of the bad seing at my location. The ground cools by radiating in the infrared, and the air immediately above the ground is also cooled, creating a radiation inversion. This uncouples the ground layer from air higher up, and in the transition between these layers turbulence occurs. This uncoupling also means that the wind on weather charts does not necessarily correspond with the wind felt on the ground. The cold air close to the ground will flow downwards under its own weight, from mountains and hills towards the coast. This wind is known as the nocturnal drainage flow, and it can be quite strong. One early morning I could see the effect of this flow as ripples in the sea surface extending as far as a little beyond the islet Dia, which is about 10 km offshore. The flow can differ in strength and direction over extremely short distances.

Seeing observations from my apartment. The value along the y-axis is the estimated seeing on the Pickering-scale. This scale was made for a 5" telescope, but I used it unmodified for my 10" telescope. The gaps between data points correspond to overcast conditions or in a few cases a missing observation. Between September 2002 and January 2003 I did not keep a record.

The charts shows how the seeing conditions from my balcony location have mostly been poor to fair during the period. This stands in contrast to this study (pdf-document) of seeing conditions at the Skinakas observatory, about 35 km away, which found very good seeing (covering a different period, however). In all likelihood the seeing tends to be much better at the 1750 m high peak holding the observatory.

Relating the seeing to the weather
The following may get a bit technical and a minimal understanding of statistics is an advantage. I have made an attempt to correlate my own seeing observations (covering the period Aug 2002, Jan-Aug 2003) with meteorological conditions. Fortunately, every day and night a weather balloon is released from Iraklio airport, about 10 km away, providing a wealth of information through a column of air. The balloon is released at 00 UT, which is about 3-6 hours after I make most of my observations. This temporal separation, and the distance of 10 km, means that it is harder to show possible correlations between seeing and meteorological variables - the correlation coefficients will probably be underestimated.

The meteorological variables I used were: wind speed at the 300 hPa (hPa=milibar) level, which is a bit below the level of the jet stream; inversion strength (explained below); inversion height; wind speed at the ground (39 m above sea level); relative humidity at the ground; at 200 m; at 1000 m; at 2000 m; and at 5000 m. I also included humidity at 1000 m at 12 UT, which is about 6-9 hours before I make my seeing estimates. Data on the wind speed at the 300 hPa level (about 9000-9500 m altitude) were found from BOLAM, and are forecast numbers, not actual observations. The numbers do however correspond well to observed wind speeds at 7-10 km height from the weather balloon data.The inversion strength was taken as the temperature increase from the minimum temperature under the inversion (typically at ground level), to the maximum temperature (typically at an altitude of 100 - 500m). The inversion height was the altitude of the maximum temperature.

I calculated Pearson's linear correlation coefficient (R) between all variables, along with P-values showing the significance level. The common practice is to regard P-values lower than 0.05 as significant. A positive R means a positive correlation between the seeing and the meteorological variable in question; a negative vice versa.

Table 1. Correlations between seeing and various meteorological variables. R-correlation coefficient, N-sample size, P-significance level.

From the above table it can be seen that seeing was best correlated to inversion strength. Next was relative humidity at 1000 m height, then humidity at 200 m. Humidity at 1000 m at 12 UT was only weakly correlated to seeing, which means it was not a good predictor of seeing conditions the following evening. The inversion strength was itself negatively correlated to the relative humidity at 1000 m (R=-0.55, P<0.0001), which makes sense; a high humidity blocks the infrared radiation from the ground and reduces the chance of an inversion forming. Somewhat surprising, the relative humidity at ground level (39 m) was not at all related to seeing. The wind speed at 300 hPa was only poorly correlated to the seeing, although I should mention that the core of the jet stream is probably quite a bit higher up than this at my location. Detailed table of results.

My limited experience suggests that (surprisingly) the local ground type (grass/concrete/asphalt etc) was of little importance. I found no trends in seeing vs. time of night, except a tendency for better seeing right after sunset. I did not experience better seeing before dawn. Bear in mind that my results are strongly influenced by the fact that my location was in a shallow valley. Had I lived on a hilltop the importance of inversions would probably have been far less.

I have found that it is often possible to avoid radiation inversions by moving a short distance. On rare occasions I could get above a shallow inversion simply by observing from my rooftop (~7 m above ground). At a nearby 200 m hilltop I got above the inversions about half of the time, with improved seeing. At the 800 m peak of the Juchtas mountain I was always above any radiation inversion, and usually the seeing there was 2-5 classes better than below the inversion. It can be seen from the below graph that inversions became more frequent and intense during the summer. During the winter most inversions were below 200 m, while in the summer they often extended to 500 m or more.

The strength and height of radiation inversions over Iraklio airport.The gaps correspond to nights without temperature inversions. Click on the image for the full-size graph. Data from University of Wyoming.

Reduces turbulence close to the ground:
- partial cloud cover blocks infrared radiation and reduces ground cooling
- sufficiently strong wind will prevent formation of an isolated cold ground layer. If this wind is from the sea, the seeing can be very good. Too strong wind is not good, however.

Increases turbulence close to the ground:
- rapid cooling during cloudless nights with clean, dry air

How to avoid near-surface turbulence:
- Observe just after sunset when a layer of cold air has not yet had time to form.
- observe on nights with partial cloud cover and/or wind from the sea
- observe from a tall building, a hilltop, or better a mountain top above the cold ground layer
- avoid observing in valleys or downwind of a hill or mountain.

From the above graphs you might conclude that by getting above 600-700 m altitude you will always avoid the inversions. Not so fast! These temperature measurements were made from weather balloons, which of course are floating free in the air. Amateur astronomers, on the other hand, are usually confined to the ground. Even if you are on a mountain at 1500 m, you may still have an inversion above yourself, because of ground cooling. The local topography plays a very important role here. My favourite location is the 800 m Juchtas mountain, which has very steep sides and is surrounded mostly by flat terrain. An inversion is unlikely to develop over this isolated peak, and usually on the top the wind is coming in from the sea. I sometimes observe from close to the Skinakas observatory, at about 1500 m altitude. Usually the seeing there is not nearly as good as at Juchtas, even though the altitude is much higher. Unlike Juchtas, my Skinakas location is not on the peak itself, but on a plateau surrounded by a rugged terrain and several higher peaks. The surrounding peaks can cause turbulence in the synoptic airflow that degrades the seeing. Local inversions and drainage flows can and do form. I have experienced VERY large temperature gradients over very short distances there, caused by radiation inversions. I speculate that very strong local inversions can develop at high altitude locations because the typically clean air allows infrared radiation from the ground to escape easily.

Links
The atmosphere and observing
AstronomyWeather Yahoo group
Clear dark sky
Shallow drainage flows
Beating the seeing
Predicting the transparency and seeing conditions
Meteorological conditions that determineobserving quality at telescope sites
Relating seeing quality to meteorological conditions
Astronomical seeing