Figure 1
Predicting Seeing
A Weather Based System
by
Eric Douglass
Introduction:
One of the things we planetary observers would like to know is the quality of
seeing we can expect on a particular evening. "We would especially like to know
this hours before dragging out all our equipment! Consequently, I initiated a
study looking for correlations between meteorological variables and optical
turbulence.
A. Study Design
The design of this study was relatively simple. At sunset I collected a wide
variety of meteorological data in the form of maps (jet stream, surface weather
maps, infrared photos), measures (pressures, temperatures, humidity, dew points,
etc.), and observations (cloud types, presence of hazes, etc.). Static measures
(pressure, humidity, etc.), distances to all event boundaries (cold fronts, jet
stream core, pressure systems), and area parameters (temperatures and pressures
at 30-50 mile radius in multiple directions) were recorded on a spread sheet.
Later, at four hours after sunset, I examined stars for the presence of optical
turbulence. This was accomplished using a two star method: stars at
approximately 45 and 90 degrees were observed for amount of visual
scintillation, amount of movement of the stars at 190x, and the amount of cell
movement after defocusing the telescope (figure 1). The use of multiple measures
both improved me quantification of the measure and decreased observer error.
Figure 1
This study was conducted over a 9 month period, and included all nights where
adequate data (observing and meteorological) was obtained. This resulted in 80
data points. These were divided into summer and winter data sets. At a later
time, the information was placed on a spreadsheet and correlations were sought.
High quality seeing evenings were defined as the top 25 of nights, intermediate
seeing was defined as the next 25 of evenings, and turbulent seeing was defined
as the bottom 50 of evenings. Note that this study only examined optical
turbulence, and not optical transparency.
B. Results
Examination of the data revealed several variables that were correlated to
optical turbulence. The first of these is the presence of the jet stream core.
The distance to the core was measured in 75 mile increments to the center of the
core. Plotting a graph of distance to the core against quality of seeing
revealed a negative correlation (figure 2). From this it appeared that anytime
the core was within a specified distance at sunset (225 miles in summer data,
300 miles in winter data), the seeing quality was routinely graded as
intermediate to turbulent (32/34 data points). Only 2 of these data points were
in the high quality group (2/34), while 10 were in the intermediate group
(10/34), and 22 were in the turbulent group (22/34).
figure 2
A second variable was the presence of a cold front boundary within a specified
distance of the observer. Distance to a cold front boundary was-measured in 50
mile increments. Plotting this distance against quality of seeing revealed a
negative correlation (figure 3). From this it appeared that anytime a cold front
boundary within a specified distance at sunset (300 miles in summer data; 250
miles in winter data), the seeing was routinely graded as intermediate to
turbulent four hours later (23/25 data points). Only two of these data points
were in the high quality group (2/25), while 8 were in the intermediate group
(8/25) and 15 in the turbulent group (15/25). It is of interest that the two
high quality evenings where the cold front boundary was within the specified
distance were also the same two evenings where the jet stream was within the
specified distance. This suggests that these two evenings were aberrations,
representing either observer error or observation during transiently calm
moments.
figure 3
A third variable was the presence of clouds of intermediate depth at sunset.
Cloud depth was measured using the infrared images (correlating the cloud depth
with cloud temperature as long as the infrared images also revealed decreasing
temperatures toward the center). The distance to these clouds was measured in 25
mile increments, and measured at -30, -40, -50, and -60 degrees C. Plotting
distance to clouds of these different temperatures revealed a negative
correlation starting at a temperature of -40 degree C (figure 4). From this it
appears that anytime a cloud of intermediate depth is within a specified
distance (150 miles) at sunset, the seeing was generally graded as intermediate
to turbulent four hours later (12/14 data points). Only two of these data points
were in the high quality group (2/14), while 3 were in the intermediate group
(3/14) and 9 were in the turbulent group (9/14). It is of note that one of the
evenings in the high quality group is also one of the aberrant evenings noted
above.
figure 4
A fourth variable was the sunset dew point temperature. Graphing the sunset
dew point temperature against quality of seeing revealed no correlations in the
summer data, but did reveal a correlation in the winter data. Here, it was found
that all nights with lower dew points (below 28 degrees F) all were in the
turbulent group (8/8 data points; figure 5 a). A similar correlation was found
in the winter sunset temperature data: nights with sunset temperatures below 50
degrees were all graded as turbulent (9/9 data points, figure 5b).
figure 5
A fifth variable was the distance to a high pressure system.
Distances to high pressure systems were measured at 50 mile increments from the
approximate center of the system. While there were few data points here, it was
noted that all evenings where a high pressure system was within 50 miles of the
observer at sunset were routinely graded as turbulent (4/4 data points). On the
other hand, high pressure systems that were between 299 and 200 miles at sunset
were generally graded in the high quality (5/8 data points) or intermediate
quality (2/8 data points) groups.
A wide variety of other parameters were measured, but were not found to have any
correlation with quality of seeing. These included:
variation in temp. over 50 mile radius at sunset
variation in press. over 50 mile radius at sunset
sunset pressure
difference between sunset temp. and dew point
maximum daytime temp.
time of maximum daytime temp.
time/amount of minimum morning temp.
difference between maximum and sunset temp.
temp. variation from four hours before sunset to sunset
press. variation from four hours before sunset to sunset
presence of a haze at sunset
presence of clouds at sunset
sunset relative humidity
C. Discussion
Optical turbulence is created when light passes through a parcel of air that is
of different temperature, and thus of different density, than the surrounding
air. Such a parcel will assume a refractive index, causing a transient
'shifting' in the image. This represents a final common pathway for all
meteorological processes that result in optical turbulence. The processes, and
their link to the common mechanism, are examined below.
The near presence of the jet stream is clearly correlated with optical
turbulence. The jet stream is a narrow band (often a few hundred miles across)
of fast moving air in the upper troposphere. Where its edges pass through slower
air, it spins of parcels of air that have a different temperature, and thus a
different density. Light rays passing through such parcels will refract.
The near presence of a cold front is clearly correlated with optical turbulence.
Cold fronts occur where heavy cold air masses invade lighter warm air masses.
This produces a situation where warm parcels are rising and cold parcels are
falling. Any light passing through such an area will undergo extensive
refraction.
The near presence of clouds of intermediate depth is correlated with optical
turbulence. Clouds form when rising warm parcels reach their dew point. The
water condenses into liquid water droplets, releasing a large amount of heat
(called the 'heat of condensation'). This warms the parcel, causing it to rise
even further. Any light passing through such a region will undergo refraction.
From this study, it appears that the optical turbulence generated from such
clouds extends out in distance to 100-150 miles and extends out in time for at
least four hours. Further, it appears that optical turbulence does not extend
out to such a distance or time with clouds that lack this 'intermediate' depth,
and clouds with higher infrared temperatures (-30 degrees C) did not manifest
this correlation.
The sunset dew point and sunset temperature in winter was correlated with
optical turbulence. Here dew points below 28 degrees F and sunset temperatures
below 50 degrees F were always associated with turbulent seeing. A mechanism for
this is not apparent, though I suspect that it has to do with the process of
radiation of infrared wavelength radiation from the earth in air of extremely
low absolute humidity. As water vapor is the primary greenhouse gas on earth,
its near absence permits such radiation to distribute itself in a different
pattern than on warmer nights. I suspect that this pattern of reradiation
somehow produces the unstable seeing condition. It is of note that these
defining temperatures were taken in central North Carolina, and are probably
specific for that region. Each region will probably have a different set of
defining temperatures, and the most appropriate way of finding these numbers is
to construct a graph for each region (of sunset temperatures/dew points to
quality of seeing). An approximation, however, might be to apply the calculated
difference between the average sunset temperature and these specific
temperatures. In my study, the severe optical turbulence occurred at 10 degree F
below the average winter sunset temperature and 30 degrees below the average
winter sunset dew point.
The presence of high pressure systems directly above the observer was associated
with turbulent seeing, while one at 200-300 miles distant was associated with
higher quality seeing. The mechanism for high pressure system development is
well known (jet stream convergence with slow descent of air), but the
connections with optical turbulence are not clear.
D. Conclusions
From this material, one can construct a method of determining the quality of
seeing using common meteorological variables. Any single term may have rare
false negatives (prediction of turbulent seeing mirrors reality), but suffers
from a high number of false positives (prediction of high quality seeing does
not mirror reality). So, for example, the near presence of the jet stream is
nearly always predictive of turbulent seeing, but a distant jet stream is often
not predictive of high quality seeing. This is because optical turbulence is a
combination of a variety of factors. Consequently, combining a variety of
factors decreases the number of false positives.
So, for winter months, poor quality seeing is predicted to ensue anytime any one
of the following condition are met at sunset: the jet stream is within 225
miles, a cold frontal boundary is within 250 miles, clouds of intermediate depth
are within 150 miles, or sunset dew point temperature is below 28 degrees F
(greater than 30 degrees F below average winter sunset temp). Conversely, high
quality seeing is predicted when none of these conditions are met. Application
of this system to my data reveals the following: of the 10 high quality seeing
evenings, 9 were correctly predicted; of the 20 turbulent evenings, 18 were
correctly predicted; of the 13 evenings predicted to be of high quality, 9 were
in the high quality group; of the 26 evenings predicted to be of high
turbulence, 17 were in the turbulent group, 8 were in the intermediate group,
and 1 was in the high quality group.
For summer months, poor quality seeing is predicted to ensue anytime one of the
following conditions are met at sunset: the jet stream is within 300 miles, a
cold frontal boundary is within 300 miles, or clouds of intermediate depth are
within 150 miles. Conversely, high quality seeing is predicted when none of
these conditions are met. Application of this system to my data reveals the
following: of the 10 high quality seeing evenings, 8 were correctly predicted;
of the 20 turbulent evenings, 15 were correctly predicted; of the 18 evenings
predicted to be of high quality, 8 were in the high quality group; of the 22
evenings predicted to be turbulent, 15 were in the turbulent group, 5 were in
the intermediate group, and 2 were in the high quality group.
Finally, there are a number of suggestions in the astronomy community about
factors that affect seeing. This study allowed the evaluation of these. They are
presented below:
(1). Some suggest that high quality seeing often occurs when a haze or for is
present. Hazes and fogs are due to an inversion, which is a stable situation
where warm air sits on top of cold air. While inversions produce little vertical
mixing and so should be associated with non-turbulent seeing, the study
indicated that there was no association between the two (they occur with near
equal distribution in high quality and turbulent groups). This is because such
inversions are generally very superficial, and do not extend through the depth
of the atmosphere. Consequently, the presence of an evening haze or fog is not a
predictive factor.
(2) Some suggest that high quality seeing occurs the night after a cold front
passes through. The study found that seeing was routinely turbulent the night
after a cold front passes, though air transparency is higher.
(3) Some suggest that the coldest nights are associated with the best seeing.
While it is clear that the coldest nights have the highest transparence due to
the lowest absolute humidity, they are also associated with the turbulent seeing
(see text above).
(4) Some suggest that the hottest days are associated with the most turbulent
seeing. The study showed no relationship between highest daytime temperature and
quality of seeing, or between the highest sunset temperatures and quality of
seeing.
(5) Some suggest that the presence of a high pressure system is associated with
best seeing. The numbers for evaluating this parameter are small, but it appears
that the presence of a high pressure system within 50 miles of the observer is
associated with turbulent seeing, but such a system between 200 and 300 miles
distant is associated with high quality seeing (see text above).
(6) Some suggest that the presence of high humidity is associated with high
quality seeing. The data from this study showed no relationship between relative
humidity and quality of seeing.
(7) Some suggest that the presence of cumulus clouds at sunset
is associated with turbulent seeing. While cumulus clouds are especially
associate with convection and optical turbulence, many such clouds are small
(cumulus humilis), and lack the deep convection needed to produce optical
turbulence at four hours after sunset (i.e. after their energy source is gone).
Deeper cumulus clouds, however, fall under the 'clouds of intermediate depth'
(see text above), and do produce optical turbulence deep into the evening.
(8) Some suggest that the presence of any clouds at viewing is associated with
turbulent seeing. While this was not examined in the study, it was clear to this
observer that anytime clouds occurred during observation, turbulence was
high--and that especially for cumulus type clouds (including altocumulus and
cirrocumulus).