I’m sorry to be writing this here, but I can’t find any other place to leave this information.

This website is very interesting and a good source of information. Unfortunately it is unreadable to me, since it is rendered with white characters on a white background.

I would request whoever programmed the cascade style sheet to either erase the background colouring or include colour settings for the characters.

Thanks for your time,

Luís

Great work! I saw you on Glenn Beck’s show in New York. Glenn is one of my hero’s, but don’t tell him. He’s far more effective thinking no one’s paying attention - “One Man, Alone, with the Burden of Truth; Saving the World, One Show-at-a-Time!” It’s wouldn’t be so funny if he wasn’t so good at it! Your Conference is a perfect example!

I’m happy he introduced you because your work has the type of facts that I find easier to translate. Reports on weather and temperature are fascinating and invaluable. However, locating an official weather station right on top of an Arizona black-top has bogus test-results significance that even a pea-brain could understand! “…the Truth will set us Free”,

With thanks,

Janie

John V has answered the question: He did NOT use adjusted data (other than for TOBS). As this question has been put forward by others, I want to make it quite clear that he did not use filled in or homogenized data.

The graphs you run with these sites are the official USHCN records, right?

And they are adjusted, aren’t they? ‘Course they are. If they hadn’t been, St. Mac never would have uncovered the Y2K error in the first place, would he? He caught the disconnect on one of them there blue graphs. So that means SHAP and FILENET adjustments will have been applied to them there blue graphs.

And what does THAT mean, pray tell? What it means is that those records
a.) Have had their missing data filled in from nearby stations, and
b.) Have been homogenized with the records from nearby stations.

Therefore, good stations have been contaminated by bad stations. And bad stations have been made more like good stations.

Therefore, the differences between bad stations and good stations HAVE BEEN DUMBED DOWN.

Consider:
a.) What? 25% of data is never collected and must be filled in from other statins.
b.) Six out of Seven stations has an CRN-estimated 1°C or more violation.

So what does that mean for a CRN-1 or 2 station when a quarter of its data is filled in and it’s data is homogenized with stations that have a 6/7 chance each being severely out of line?

I therefore hypothesize that the differences between good stations and bad stations are worse, possibly MUCH worse than those blue graph thingies indicate.

I futher suggest that the site error is not being adjusted away, it is merely (via SHAP and FILEMNET) being spread around in a manner so that they are not corrected, but do not stand out. The bias is not lost, oh no. Some of the CRN-5 data error is transmitted to the better stations and some of the CRN “correctness” is transmitted to the worse station.

Result: the complete bias remains, but it does not stand out as much because the CRN-5s are adjusted down and the CRN1sare adjusted up.

THREE EXCLAMATION POINTS ! ! !

I have another related question. Did the John V study comparing CRN-1 and CRN-5 stations use data that had been adjusted using FILENET and SHAP? (This goes beyond the urban-rural problems.)

REPLY: Hmmmm more later