Cloudy Day Sunny Day

February 3, 2012 7:35 AM

What is the sunniest day of the year? and how is that determined?

Of course this varies by location. For me, right now, that's Portland Oregon, but next month these architects I work with might consider a building in a different location.

I am asked, given TMY3 data (here's a list of the relevant data in the TMY3 set), how to determine the sunniest days and most overcast days in the typical meteorological year data set.

Are "sunniest day" (or "least overcast day") and "most overcast day" defined in meteorology? or are they bogus data deployed by evening news weatherpeople to fill air time?

Is there an accepted means to determine the record day for least and most overcast, taking into account direct/diffuse normal/horizontal illuminance and hourly sky cover?

Bonus question: can anyone recommend a good textbook introduction to meteorology that provides an understanding of the definitions and statistics used in the field?

posted by Prince_of_Cups to science & nature (3 answers total)

This is a difficult question. First, the National Weather Service monitors all meteorological parameters at official stations (mostly airports) using ASOS (Automated Surface Observing System). Here is a link to the manual that gives a huge amount of information regarding how the various measurements are made, including sky conditions. A brief summary: ASOS uses lasers to determine cloud heights and coverage. It then uses an algorithm to convert that data into a report. There are 5 possible outcomes as far as the official sky condition:
Clear, Few Clouds, Scattered Clouds, Broken Clouds, and Overcast. Each one of these categories carries with it a specific % of cloud cover (see table 3 on page 25 of the report). ASOS also has a sunshine sensor and for each station it keeps tracks of total sunshine minutes in a given day and compares that to the potential number sunshine minutes in a given day. So, for example, if the time between sunrise and sunset was 8 hours on a particular day, there are 8 hours of possible sunshine. If the sun was out for only 4 of those hours, then that day had 50% of the possible sunshine. Normally this data is tabulated by month as well so you can get a feel for seasonal changes. You can look at this site for cloud cover of different locations all over the USA and this site for % possible sunshine data. You can also check the official NWS site for Portland Oregon for climate data.
posted by Seymour Zamboni at 9:53 AM on February 3, 2012 [1 favorite]

I don't think those are well-defined meteorological terms. (note, I know more about daylighting words than I do about meteorology words)

If someone has asked you for "the sunniest day" then there's a definite possibility that they don't really know what they're asking for, and you'll have to decide what they really want to know about, and give them a number while telling them how it makes sense to use the number. Even if you call up the weatherman and ask what he meant when he just said "this is the sunniest day we've had all month!" you'll have to decide if that definition is one that actually applies to the thing you're doing. (but yes, call the weather man and ask)

The TMY3 data gives a lot of information for some date and time, presumably once an hour since it says each point is averaging over an hour. You can definitely use that to pick some type of data (say, total horizontal) and plot it out for each day. You could take the peak value and that would tell you how bright "sunny" can be. You could take the total of all the hours for each day and that would be the integrated amount of light received (but that's more relevant to photovoltaics than daylighting). You could divide that total by the length of the day and get the average brightness of the daylight hours. You could do that treatment with the cloud cover number instead of the illuminance (but then you'd have to decide whether to weight it for clouds being unimportant after sunset). Any of those numbers mean something that's relevant to "sunniness", so in some sense it doesn't matter which one you use, so long as you say what you mean.
posted by aimedwander at 9:55 AM on February 3, 2012

Geometrically, midsummer is the sunniest day: the sun's at it's highest, and is up the longest. But cloud cover is variable, and that's where site data comes in.

The tinhg about TMY data (and I may misunderstand this) is that it represents the typical year, without outliers. You're looking for the least statistically significant events, so TMY may not help you. Climate normals or long term measured data might be more useful.
posted by scruss at 9:58 AM on February 3, 2012