Go to this page and scroll about half way down to see seven colorful charts indicating average weather patterns in my home town of Springfield, Illinois.
The “Average Temperature” chart is pretty straighforward. I interpret the fuzzy green bar to represent the average range of daily highs and daily lows in the US, with the light green line representing the daily average. The “Average Precipitation” chart, however, makes less sense. The fuzzy green bar that’s supposed to represent the US average of precipitation has a range of about 3 1/2 inches each month, with the green line in the middle. I don’t get it. The US average for precipitation in January is either 2.88 inches (as suggested by the green line), or it isn’t.
Furthermore, the “Average Snowfall” chart makes even less sense, in relation to the US average. The light green line for January hovers around five inches, but the green background for January goes all the way up to 11 1/2 inches.
Finally, the “Cloudy Days” chart. If I’m interpreting it correctly, in January, 57% of the days are cloudy, 77% are partly cloudy, and some number that I can’t figure out are 100% sunny.
Also, compare Springfield’s “Cloudy Days” chart to that of Barrow, Alaska. There’s something happening there in December well over 100% of the time, but I can’t tell what it’s supposed to be.
I’m thinking that the wide green portion of the graph represents the range of high and lows reported from the 4000 stations sited at the top. The thin green line then represents the average.
For example, for Snowfall in January a station in the Northeast US might be at 10 inches, but Phoenix would be at 0 inches, thus represented that way on the graph. Likewise, July is simply at zero, because noboby recorded snow in that month.
That was my first thought too, but it seems a little strange that every city in the US gets between 2 and 4 inches of precipitation in July.
No, the cloudy + partly cloudy + sunny days add up to 100%. So 57% of the days are cloudy, 77-57=20% of the days are partly cloudy, and 100-77=23% of the days are sunny.
And the green shading represents days with precipitation, so 30% of the days have some precipitation. 57-30=27% of the days are cloudy but not rainy.
The 100% graph was explained quite well so I’ll move on to:
I don’t get it either, but I can think of only one rationale explanation. The solid line probably shows the average of all temperatures recorded in some set of places, over some number of years. The wide band shows the range of average temperatures that have occurred across the years.
The solid line for say, April, shows a little over 3 in of precip. That’s an average for all years for which they have data. The top of the green band for April is at about 4.5 in, which is probably for the year for which there was the greatest average precipitation. The low is just under 2 in, probably for the year for which there was the lowest average precipitation.
If I am correct, the green band is just a red herring anyway because you want to compare Springfield for all years to the whole country for all years. The green band doesn’t tell you a thing unless you also have the min and max ranges for Springfield.
The Barrow chart looks like an error. The shaded portion represents days with precipitation, and is an overlay (not part of the stack that adds to 100%). You would expect this would be a subset of cloudy + partly cloudy days. However, this chart shows that precipitation happens even on clear days in January, and as you pointed out, on more than 100% of the days in December.
Also, the Precipitation graph seems to fail to take into account that snow is precipitation. It may be intended to include only rain.
I don’t think so, because that means that places like Phoenix that get very low precipitation all year would drop the bottom of the green band to be very low for the whole year. But you will see that Phoenix has an average precip line that falls below the band.
I think the Barrow thing is just a way to alert the viewer that “cloudy days” is an unusual term compared to the other months. IIRC, there is no sun light for a period of time during December there. Is that really “cloudy days” or non’cloudy or even days?
Hmm… but the bottom of the green band is at the bottom throughout the year. I’m sure there are some places that zero out on the snowfall all year round, it might not be the same place twice though… I mean, it makes sense to me that in January the range is from 0 to 11 inches, but in July it’s 0 to 0 inches.
Will pull it up and look closer now . . . . I really only looked at the snowfall graph…,
Okay, I see what you’re saying – on the Precipitation Graph, the lower line can’t be the lowest range, as Phoenix alone would drop it to at least where the Phoenix line is… interesting! Somehow it’s an average of averages? Or each graph might be different?
I don’t think that can be the case, as the windspeed for Springfield is in the winter completely above the U.S. light green band.
I think instead the light green band on each chart represents where most of the U.S. falls, with the dark green line showing the average. This is a common way of showing an average and also how far from the average things tend to get (which is also useful). Exactly how they define ‘most’ isn’t clear.
It’s still useful, though. You can say that Springfield’s snowfall and temperature are very close to the U.S. average. Rainfall is a little lower in the winter than average, but not remarkably so. But winter wind is a lot higher in Springfield than most of the rest of the U.S.
The spread around the average values are obviously across the 4,000 weather stations. The grapher calls these the “U.S.” values, but since we don’t know where these weather stations are located, it’s really just anyone’s guess if these locations comprise a valid sample across the country. We also don’t know if the spread represents the full range of values or if they’re some other valid measure of dispersion like the interquartile range. There’s not enough of an explanation. In any case, the graphs are interesting.
