I’m traveling today, so an open thread seems useful. Some topics might be:
1. Schollenberger’s coup with SkS ratings data
2. Will Steve Goddard issue a correction, or just ignore it?
3. Why is Dr. Mann playing with fire on Twitter? Is he just being an emotional child or does he want another lawsuit? We’ve been down this path before. See: Monday Mirthiness – Watch the genesis (and retraction) of a smear.
Yet, he persists, as if he can’t help himself:
Michael Mann-child accuses @WattsUpWithThat of being paid to lie by @HeartlandInst. Example? Or just libel? pic.twitter.com/KkMfDS3rqc
— Steve Milloy (@JunkScience) May 10, 2014
Mann’s bogus claim asked and answered here: http://wattsupwiththat.com/about-wuwt/faqs/
If I’m paid (I’m not) to spout “disinformation” as Mann claims, why do I publish posts like this one and this one correcting other skeptics and true disinformation.
Of course, his hateful claim might just be misdirected rage at having been sliced and diced by McIntyre again.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
JamesS says: May 11, 2014 at 1:49 pm
“Where does any bias come in? Why even bother with the tmax or tmin readings from the thermometer itself, whether mechanical or digital, when you’ve got the hourly readings to look at?”
You don’t have hourly readings to look at. You just have 100+ years of records where people observed and reset min/max thermometers once a day at an agreed time. By agreement that time could and did change. That’s the record we have.
Its really hard for me to see how TOBS could be a problem if a rational decision on when to read/reset the instrument is chosen. I would naively choose 10 PM, 7 hours after today’s maximum (at 3 PM) and 7 hours before tomorrow’s minimum (at 5 AM). 3 PM or 5 AM would be chosen if one wished to deliberately confound the record. I am wondering what assumption about “prehistoric” temperature records is made to justify a TOBS adjustment. Surely not that the observers were deliberately choosing 3 PM to observe the day’s readings. What were they told to do? Or is it assumed that they made the worst possible choice?
@Nick Stokes. Thanks. I did some Googling of “time of observation bias,” and I can see that’s a problem if all you have is tmax and tmin for a day. But Mosh was talking about having 100 stations in Ohio with hourly records, and setting aside 50 of them and building an empirical model out of the other 50, etc.
Where is the need for TOB adjustments there? Surely, in the 21st century we can get automated hourly recordings for every weather station out there so that this statistical nightmare can go away from now on.
JamesS says: May 11, 2014 at 2:31 pm
“But Mosh was talking about having 100 stations in Ohio with hourly records…”
Yes, that’s how they now try to figure it out. TOBS creates a bias because if you read a max on a warm afternoon, it’s likely the second day will record a max which was not that day, but a second counting of the warm afternoon, post 5pm (if that’s the reading time). If you read in the morning, you’d double count cold minima. TOBS doesn’t say the actual reading was wrong, but the statistics are biased. That hurts when you change reading time.
You can quantify how much by looking at modern hourly records (same place or nearby) to get the diurnal cycle. If you imagined a min/max thermometer with setting time but knowing those hourly records, how much would the bias be? That’s what Steven is talking about.
“Where is the need for TOB adjustments there?”
This is all about making sense of old records. Modern data – no problem.
Reading what Nick wrote, it makes me wonder in there is a latitude bias to TOBs.
See how day length varies http://www.nature.com/nclimate/journal/v2/n4/fig_tab/nclimate1430_F1.html
In reply to John Eggert May 10, 2014 at 3:11 pm, in turn quoting ilma630 at May 10, 2014 at 1:18 pm:
Just as there’s no such thing as a “flashdark” (casts a beam of non-light), the cold body does not and cannot impart heat to the warm body. It CAN impart a bit of *radiation* but radiation is not HEAT.
The EFFECT of doing so is the same as insulation. Insulation does not make your house warm. It merely reduces heat loss. If your house is -10 C and you wrap it with a meter of fibreglass insulation it will still be -10 C.
Unless of course you have a heat source inside. THAT is what warms your house.
CO2 does not warm the earth. It does have an ability to help the earth KEEP some of its heat.
Anthony Watts says: May 11, 2014 at 3:54 pm
“Reading what Nick wrote, it makes me wonder in there is a latitude bias to TOBs.”
I’m sure latitude has an effect. Even time zones do – 5pm is different meteorologically depending on where you are. But Vose et al 2003 describe in detail how TOBS is implemented. They do a simulation of the kind Mosh describes at each of 500 US stations (map at Fig 2). Each station then gets the TOBS adjustment calculated on the stats of the nearest of that 500 (and of course, the known change of TOBS). So that should account for the latitude effect.
OK, I have a stupid question.
Why are we worried about adjustments for TOBS at all?
Yes I am serious. If the task is to determine the temperature of the earth. would not the most sensible procedure to be to take all the temperature measurements at the exact same time? Yes that would mean taking the temperature in some places in noon and on the other side of the earth at midnight, but provided that you had enough spatial coverage (I would argue that we don’t) you would get the temperature of the earth at that point in time. Min and max and adjustments for TOBS all become immaterial if you do it that way.
If one was trying to determine the temperature of a large installation like a steel mill for example, you wouldn’t take the min and max temps at various points in the plant at various times of the day and then try and come up with some way to average it. The only sensible thing would be to take the temperature at as many different points as possible all at exactly the same time. You’d then repeat that daily. or weekly or whatever interval made the most sense.
By worrying about mins and maxes and TOBS, we’re really mixing up a lot of weather that just makes the whole think that much more complicated by naturale variability and that much harder to tease the trend from the noise.
Yes, I know, it is unlikely we have the observational data to do that, not even the satellites come close. But if you asked ME to determine the temperature trend of the planet, that’s pretty much what I would do. Get as many measurement points as possible, and pick a time. Noon eastern standard time for example.
Comparing Tmax at 2:00 on a given day on a given spot on earth to Tmax at 3:30 PM in a completely different time zone just seems like a rather awkward and nonsensical way to go about it. If you want to know the temperature of the earth at any given time, then PICK THE TIME and read all the thermometers in the world AT THE SAME TIME.
If course that still leaves the issue of a metric that doesn’t vary linearly with energy flux anyway, but it would be a more reasonable approach in my opinion.
I’m rather gobsmacked by the Vose et al 2003 paper. In it, the author states, “suppose an
observer reads the maximum and minimum thermometers at 1700 LST on April 1, then a cold front passes through the area overnight. If the temperature on April 2 never exceeds
the value at 1700 LST on April 1 (when the thermometers were last reset), then the recorded maximum will actually be the temperature at 1700 LST on April 1.”
Is he really saying that there is no way to reset the thermometers to ensure this doesn’t happen? I understand the point, but what are the thermometers reset to that would allow this to happen. I’d have thought the min would be set to 200F and the max to -100F so that you’d know you’re getting a true min and max for the day.
Obviously, if you go to an hour-by-hour collection, the TOB goes away. But I am absolutely amazed that no one ever came up with a way to ensure that the max/min temps collected really were the values for that day, and not possibly carryovers from the previous day.
JamesS says:
The point is that they do not KNOW if it happened or not and have no way of finding out now, but they apply the “correction” any way.
And they call it Scientific?
The basis for the TOB adjustment seems to be the assumption that, over the period when no metadata giving the actual time of observation is available, the observers’ preferred observing time changed from that giving the least favorable TOB adjustment (evening observations when high temperatures may carry over) to that giving the most favorable one (morning observations when low temperatures may carry over), neither of which is a rational choice for an observing time (see my above comment).
Steven Mosher says:
May 11, 2014 at 12:32 pm
In answer to my question Steven catagorically states “1. TOBS bias is real”
But is it real for every applicable station?
Is it the same amount for every station?
How often do they calculate the possibility of bias occuring?
How big is it, I have seen anything from 0.03 to 0.6 degrees quoted?
How much affect on the overal trend can that bias have?
The usual suspects and the MSM have gone absolutely bonkers. Wessssssst Annnnnnntarrrrrrrrctic Ice Sheeeeeeeeeeet allllllllllllllllllrrrrrrrrready innnnnnnnnnn colllllllllaaaaaaaaapse! Read all about it ….
That’s not an accurate analogy, because CO2 is a one-way insulator. It lets in energy in one form but reflects back (in part) that energy after it has bounced off the earth and been changed into another form. So adding more of it will have a warming effect on the earth.