Topic: hi/lo strangeness

[danxx]

Not a big deal, but has anyone noticed that in the 10-day plots, the high and low temperatures listed above the curves are ALWAYS 2 degrees warmer than the curves themselves? Kinda odd. So on a day for which high | lo is "100° | 77° F", the max and min of the curves are 98 and 75. See attached. Always always.

 [ You are not allowed to view attachments ]

[awsum140]

I've noticed a difference, too, for quite some time and never mentioned it.  I just checked and it's showing a one degree difference for me.  Given all the other problems with WU, this one doesn't seem all that earth shaking, but maybe Galfert will want to track it since it is a chronic problem.

[danxx]

As I said, it's not a big deal, but somewhat of a curiosity. Programming-wise, you just need to get your data from the same source.

[awsum140]

Or maybe run cross checks if using different sources/formulas.  The other day a user reported having a dew point of -235 or so with air temps of about 104, he is in Vegas.  It looks like there's more wrong with displayed data as time goes on.

[danxx]

Well, getting really bizarre temps is something that's been happening for months, and was reported here in the spring. It was happening to me, and then it stopped. Those values were just wildly wrong. These values are right, but curiously just a little bit different. Not that 100F is profoundly different than 98F, and I'm sure not going to tell the difference, but it does reveal some degree of sloppiness.

[WSWeather]

Mine wander between 1 and 2 degrees different but considering the inaccurate forecasting I rarely look at those graphs anyway. 

[danxx]

Gee, if you're going to do it wrong, at least do it wrong consistently!