This article first appeared in the Autumnal Equinox, 1998 issue of the HORIZON newsletter of the WCAC.

Hey, at Least itís Better Than Bespin!
By Phillip J. Creed

 One of the biggest complaints about astronomy in Ohio are the clouds.   ďItís just too cloudy around here,Ē most astronomers say, and theyíre right to point out that itís not exactly an optimum spot to put up a professional observatory.  Having already done an astro-weather presentation back in May, finding out how many clear nights one has per year is something that naturally garnered my interest.  I didnít figure it would be very hard to ascertain how many clear nights there are in NE Ohio.

Boy, was I wrong.  Hey, ask the National Weather Service, right?  Surely they must have some statistics on how many clear nights we get around here, right?  WRONG.  They do indeed break down sky cover statistics for the month and for the year, but thereís one little glitch--they are DAYTIME statistics only.  Believe it or not, there are no official estimates or records of clear nights ANYWHERE IN THE COUNTRY.  But I did find a way to get a reliable estimate for the number of clear nights in NE Ohio, and in the process uncovered some fascinating results.  By the end of this article, I will show that:

1. Itís not too good in NE Ohio.
2. Itís not as bad as you think.
3. I have no life.

 First, the source of the study.  The National Weather Service keeps monthly climatic records for all major weather stations (In Ohio, they would be the Akron-Canton airport, or CAK for short, Chesapeake-Huntington, Cincinnati, Cleveland Hopkins International Airport (CLE), Columbus, Dayton, Findlay, Mansfield, Marietta-Parkersburg, Youngstown, and Zanesville).  The data from CAK and CLE was available at the University of Akron's Bierce Library in the Government Documents branch.  These sheets will include the highs and lows for each day, daily precipitation, types of precipitation, and sky cover.  For CAK and CLE, there are two columns for sky-cover.  The first is the daytime sky cover, in tenths (i.e. 4 means 4 tenths sky cover, or 40%).  The second column is the 24-hour sky cover for that date, also in tenths.

 Of course, there might be some debate as far as what constitutes a clear night or a cloudy night.  The National Weather Service defines "clear" as sky cover of 0-3 tenths, partly cloudy as 4-7 tenths, and cloudy as 8-10 tenths.  This definition is applied to ALL weather stations, not just those in Ohio.  Before I spent FOUR HOURS analyzing the data ("Hey, this looks like a job for SUPER GEEK!"), I thought that a good indication would come from the daytime statistics alone.  The days in a typical year for NE Ohio can be broken down as follows:

CLE: Clear   66 
 Partly Cloudy 97
 Cloudy  202

CAK: Clear  68
 Partly Cloudy 99
 Cloudy  198

These are averages courtesy of the National Climatic Data Center (NCDC) in Asheville, NC, and are through 1995; these stats are available on-line.  Cloud-cover statistics had been kept at the Akron-Canton Airport for 46 years (1949-1994) and at Hopkins for 53 years (1942-1994).  I didn't have every year available to me that the NCDC had, and I didn't have all the time in the world, so I decided to look at data from 1975 through 1994, for a 20-year average.  I couldn't use data from 1995 onward because the NWS switched to using automated weather stations, and sky cover data have not been put on NCDC sheets since then.  Before 1995, all weather observations were done manually.  In addition to the daily sky cover, there are also sky cover data for every 3 hours from the period 1975-94, and I feel it is very reliable data.  For instance, if there was a veil of cirrus clouds, it would list sky cover as, say, 8 tenths with a ceiling of 25,000 ft, while an automated station might list sky conditions as "fair."

Anyhow, when we look at the averages for CAK and CLE, we realize that, well, they're not good.  At CAK, this means that only 18.6% of all days are clear, and a majority (54.2%) are cloudy.  The situation is a little worse at CLE, but there isn't that much of a difference.  These aren't the worst, though.  All recording stations in West Virginia, as well as western Pa and NY have less clear days than we do (Buffalo averages just 54 clear days).  And the Pacific Northwest, well, they're even worse.  All of Michigan except for the Detroit area (where the sky glow will utterly ruin your skies) has less than 70 clear days/yr--plus longer twilights because they are further north, and later sunset times because they are further west but still in the Eastern Time Zone.  Some say Ohio is the worst state to do astronomy from, but I think there are several states that best us in that regard.  For instance, from West Virginia through Maine, the western slopes of the Appalachians have remote, dark skies, but horrible weather due to orographic effects from a prevailing westerly wind.  East of the mountains, it is not unusual to get 90-100 clear days/yr due to a dry, downsloping wind--but by the time your weather improves to that point, you will easily pick up the huge skyglow of the Northeast Corridor from Washington through Boston.  Bottom line:  There are many people in the U.S. that live in spots that are less favorable for astronomy than Canton, Ohio.  Be thankful for what you have.

When we look at the rest of the world, it becomes clear that NE Ohio is actually a better spot to do astronomy than what half of the world's population must endure.  Cleveland, Ohio, averages 50% of possible sunshine (no value available for CAK, but it's probably very close; maybe 1-2% more), but this would be the envy of Europe, except along the Mediterranean coast.  Rome has about the same amount of sunshine, and Madrid and Lisbon have more, but all the other European capitols have substantially less, usually around 40%.  London is really bad; it only averages about 35% of possible sunshine.  Its sunniest month, July, has 46% possible sunshine--about the same as Cleveland in March.  And December?  Well, Cleveland's average is a dismal 26% (ours is a little higher for sure in December, but they never kept this statistic at CAK), but London's is just 19% (Even Seattle in December averages 21% possible sunshine!). Warsaw has just 10% possible sunshine in December, and St. Petersburg has 2%! (7% for November!)  It makes matters even worse when one considers that the only in summer does most of Europe have a fighting chance to observer, but that's when twilight really drags, sometimes not even disappearing (by comparison, even on June 21 Ohio will have about 5 hours of total darkness).  Other areas of the world that have less than 50% possible sunshine include all of Russia (yes, even Siberia), almost all of the former Soviet Republics, the densely-populated southern half of China, the densely-populated Congo Region of Africa, plus the Equatorial regions of South America.  Closer to home, Canada endures slightly less sunshine than we do, but also at a more northerly latitude.  In fact, as far as major countries go, the United States is about as sunny as they come!

Okay, so we've seen how cloudy the days are in the world and in NE Ohio.  But the burning question still concerns the nights.  How clear are the nights, and how do you determine the number of clear nights?  Well, I didn't actually determine the number of clear nights; rather, I made a guess.  Remember that there are two columns for sky cover--one from sunrise to sunset, and the other for the 24-hour period from midnight-to-midnight.  Since I know both, the nighttime sky cover can be inferred.  For instance, suppose the columns said that the daytime sky cover was 8/10 (cloudy) and the 24-hour sky cover was 5/10.  The NWS will list it as a cloudy day--but wait.  If the daytime average is 8/10, but the overall average of day AND night is just 5/10, this means the nighttime average must be substantially less than 5/10--probably around 2/10, but this depends on the season and thus the lengths of night and day.  This means that it was indeed a cloudy day--but the nighttime was clear.  Care must be taken in interpreting these results; what I am really determining is the amount of sky cover between midnight and sunrise and between sunset and the midnight for the SAME day; what I really need is a sunset-sunrise sky cover statistic that spans the last part of one day and the first part of the next, but the assumption is that both measures averaged over a period of 20 years will be approximately the same.

 Now with that definition of a "night," the nights in an average year for the Akron-Canton Airport (which is applicable for most club members), from 1975-1994, are listed below.  This is averaged out assuming a 365-day year (meaning a slight data correction for 1976, 1980, 1984, 1988, and 1992):

Clear  80
Partly Cloudy 132
Cloudy  153

 Hardly a jubilant picture, but that's substantially better than the daytime stats!  Notice the increased number of clear nights versus clear days, and the sizable drop in the number of cloudy nights versus cloudy days.  Whereas 54.2% of days were cloudy at Akron-Canton, the amount of nights that are cloudy drops to a more palatable 41.9%.  A month-by-month average (this is what really burned all that time in the library!) is shown in Table 1:
 
Month
Clear
PC
Cloudy
January
3.0
8.9
19.1
February
4.0
8.7
15.3
March
4.5
11.3
15.2
April
6.4
11.4
12.2
May
8.4
11.6
11.0
June
8.8
12.9
8.3
July
9.9
14.0
7.1
August
10.1
12.1
8.8
September
9.5
12.0
8.8
October
8.1
11.7
11.2
November
4.4
8.6
17.0
December
3.1
8.3
19.6
Total
80
132
153

The year-by-year tally is seen in Table 2:

Year  Clear PC Cloudy

1975  81 127 157
1976  96 124 146
1977  79 145 141
1978  92 129 144
1979  73 127 165
1980  74 149 143
1981  63 132 170
1982  74 132 159
1983  83 136 146
1984  77 147 142
1985  77 139 149
1986  79 121 165
1987  96 117 152
1988  101 130 135
1989  73 124 168
1990  82 127 156
1991  78 147 140
1992  81 126 159
1993  61 137 167
1994  78 135 152

AVERAGE 80 132 153

So the average breakdown of nights in a year for the past 20 years has been 80/132/153.  The best year to do astronomy in the last 20 years around here appears to have been 1988 (101/130/135, read 101 clear nights, 130 partly cloudy nights, and 135 cloudy nights) and the worst year was probably 1993 (61/137/167), although it should be pointed out that 1981 (63/132/170) had more cloudy nights.

 These results show that, contrary to conventional wisdom among frustrated astronomers, it is more likely to be clear during the night than during the day at CAK.  This can be partially explained by the fact that some of the cloudiness is diurnal, such as the fair-weather cumulus clouds that pop up during the heat of day, cover up half the sky, and then dissipate by nightfall.  These are caused by convection as the warm, moist air at the lower levels, heated by the hot summer sun, goes upward into cooler regions aloft and condenses.  These results also dispel what I call the October myth--the notion that October is the clearest month.  Actually, these results suggest that July has the least number of cloudy nights and that August has the most clear nights; October in both categories is only the fifth-best month.  However, WCAC members have good reason to like October; haze is almost non-existent on a clear night in October and the sun sets before 7:00 p.m. for almost the entire month.  Still, this means that by the time you pick up this newsletter, the chances that each remaining night of 1998 will be clear is already heading downhill.

 What about the statistics for Cleveland?   Well, considering how long it took to compile monthly averages for CAK, I thought that it would be prudent just to tally up the averages for the entire year.  The average breakdown of nights in a year for Cleveland from 1975-1994, are, well, worse:

Clear  63
Partly Cloudy 135
Cloudy  167

Here is a year-by-year tally for CLE:

Year  Clear PC Cloudy

1975  60 132 173
1976  76 137 153
1977  61 152 152
1978  69 138 158
1979  63 134 168
1980  62 136 168
1981  54 123 188
1982  59 149 157
1983  75 123 167
1984  63 144 159
1985  69 129 167
1986  65 107 193
1987  74 128 163
1988  75 135 155
1989  48 152 165
1990  57 145 163
1991  63 142 160
1992  61 120 185
1993  41 135 189
1994  73 137 155

AVERAGE 63 135 167

 For your information, CAK and CLE are 39 miles apart as a sober crow flies.  This means if we assumed that the trend is linear and we go on a line from CLE to the south-southeast to CAK, we'll gain 1 clear night per year for every 2.3 miles, and would lose 1 cloudy night per year for every 2.8 miles!  What a difference 39 miles makes!  Why does Cleveland have so many less clear nights a year?  Think back to the convective cumulus clouds seen around here.  Cleveland actually doesn't have too many of those in the summertime.  Why?  Because Lake Erie is cooler than the surrounding, and usually warmer terrain, it limits convection.  If you look on a high-resolution visual satellite image during a typical summer day, you will see puffy cumulus clouds forming areas except for those just downwind of Lake Erie--otherwise known as the "Lake Shadow."  Thus, during the day, CAK is cloudier than CLE--but they're both about the same at night.

But when wintertime comes, the lake is much warmer than the air aloft in the wake of a strong cold front, and Lake-Effect snows pile up in Cleveland and off to points east of Cleveland.  Rarely does the snow amount to much down here, but we'll usually pick up clouds; though not as much as Cleveland does.  When the air cools at night, the relatively warmer waters of Lake Erie will remain at roughly the same temperature through the course of the night--thus the temperature difference between the air aloft and the lake INCREASES, and can make nighttime cloudiness WORSE for Cleveland and the lakeshore at night than during the day; whereas at CAK the average sky cover is almost always less at night, even during winter.  Unfortunately for Cleveland, it appears that it just as likely, if not a little more likely in wintertime, for the clouds to roll in at night than during the day.  This might also explain why, even though CLE only averages 4 more cloudy days/yr than CAK, it has 14 more cloudy nights/yr between 1975 and 1994.

As would be expected, winter in Cleveland really drags down their tally of clear nights each year (there are many times in the last 20 years that consecutive months went by in winter with 0 clear nights; that never happened at CAK); while at CAK we still manage to squeak a few in.  Another key difference between CAK and CLE is that in March and April, warmer air will often go over a now-cold lake, resulting in low clouds and advection fog for Cleveland, but not causing anything at CAK except for warmer temperatures; this is a more local event confined to Cleveland, but it is enough to add to the disparity of clear nights between CLE and CAK.

Indeed, between 1975 and 1994, NOT ONCE HAS A YEAR IN CLEVELAND PRODUCED MORE CLEAR NIGHTS THAN AT AKRON-CANTON.  While the best year for CAK was 1988 (101/130/135), the best year for CLE was 1976 (76/137/153).  The worst was 1993, and it was abysmal (41/135/189).  Only twice in that 20-year period has a year at CLE had less cloudy nights than CAK, and both times it was very close.  The year that had the most cloudy nights at CLE was 1986, with 193 cloudy nights (the worst for CAK was 170 cloudy nights in 1981).

But we can go one step further.  Remember, the statistics I have examined only go back to 1975.  What about the daytime averages?  Well, it turns out that if we look at the daytime averages from 1975-1994 for both CLE and CAK, it becomes apparent that the 20-year period I examined was actually a bit cloudier than the average, meaning that the clear night calculations are slightly underestimating the long-term occurrence of a clear night in NE OHIO.  Compare the daytime averages from 1975-94 for the long term averages listed near the beginning of the article:

CAK Daytime Averages, 1975-1994 

Clear  65 
Partly Cloudy 96
Cloudy  204 

CLE Daytime Averages, 1975-1994 

Clear  60
Partly Cloudy 91
Cloudy  214

Now, if we look at the difference between the nighttime and the day averages, we find that the difference between clear nights and clear days for CAK from 1975-1994 was +15 (80 - 65).  These results are summarized below:

(Nights) - (Days), CAK, 1975-1994

Clear  +15
Partly Cloudy +36
Cloudy  -51

(Nights) - (Days), CLE, 1975-1994

Clear  +3
Partly Cloudy +44
Cloudy  -47

If we then take these differences between night and day and apply to the LONG-TERM AVERAGES (again, a risky assumption, but it is still bolstered by the law of averages), we get:

Nights from CAK in an average year:

Clear  83
Partly Cloudy 135
Cloudy  147

Nights from CLE in an average year:

Clear  69
Partly Cloudy 141
Cloudy  155

Thus, as you can see, there are spots that are worse than the Canton area; just look at Cleveland, for instance.  If you're like me and live near CAK, then chances are that seemingly-small 39-mile difference between CAK and CLE gives you an extra 14 clear nights a year, averaged over a period close to 50 years.  If conditions are that bad at Cleveland Hopkins International Airport, which is on the west side of town, ever wonder how bad it would be in, say, Lake, Geauga, or Ashtabula Counties, in the heart of the Snow Belt?  Also, since most of the WCAC members live south or west of the Akron-Canton airport (which, though residing in its extreme southern portion, is still in Summit County), you probably have a little more clear nights than it does, especially in winter.  Also, one can use a partly cloudy night as well as a clear night, especially when special events like eclipses or comets come around.  For instance, I saw Halley's Comet 5 times in December, 1985, and 6 times in January 1986. Each month had only 1 and 4 clear nights, respectively, but they also had 8 and 7 partly cloudy nights, respectively, too.  Another example was Comet Hyakutake.  I logged 22 observations in a 41-day period from 3/1/1996 to 4/10/1996--and I know that not all of those 22 nights were clear.

It should also be pointed out that while there is a sizable gain in clear nights per year by heading from CLE and CAK, this increasing trend will only go so far.  Granted, your chances of getting a clear night in Dover, Millersburg, or Cambridge are better in winter, bear in mind that there is still a lot of regional, or synoptic winter cloudiness in Ohio.  Also keep in mind that as one heads further south, the observing prospects in summer become a little worse because fronts tend to stall in West Virginia in the summertime, though the increase in summer cloudiness near the Ohio River is less than the increase in winter cloudiness as one heads toward Lake Erie.  I'd say that given these figures, it is not unreasonable to expect 90+ clear nights per year somewhere south or southwest of Wilmot, and maybe 100 around Columbus/Cincinnati, but those are mere guesses.   And those are also figures for ANY ONE SPOT; staying mobile and having several established observing sites in your portfolio can increase these numbers a little more.

While some astronomers liken Ohio to Bespin, the cloud-covered planet in The Empire Strikes Back, it does in fact have its share of good weather from time-to-time, and if one is willing to establish a network of observing sites in different directions from one's house, the number of usable nights goes up a little further.  There are certainly spots in the country and the world that are better, but there are certainly places that are worse.  So don't give up hope; we get more clouds than a lot of other spots, but that makes us appreciate a clear night a little more than other

Back to WCAC Home Page
| Current Stories/News | Past Meetings | Past Special Events | Past Horizon Newsletters |