Why is it that people like Zeke can’t understand what they are doing wrong?
Using anomalies, infilling and gridding might be OK if you were happy with your network of stations and were just trying to generate a spatially averaged temperature anomaly across the US.
But that isn’t what I am doing. I am trying to see how much adjusting is being done, and the only sensible way to do that is to average the final station temperature and subtract the average of the raw station temperatures. That gives you a precise measure of the average adjustment for a particular year.
In a network with a systematically changing composition (rural station loss etc.) the very last thing you want to do is smear the data together and generate infilled data for the missing stations. That is 100% guaranteed to give you incorrect results. Besides a multitude of other sins, they are turning rural stations into urban stations and making UHI worse.
Zeke says that they are done adjusting data, but nothing could be further from the truth – the adjustments are increasing exponentially.
The temperature record needs to be turned over to responsible adults in the private sector.
Yep, but I would like to add that what Watts has discovered by investigation is that in groups of perhaps a half dozen,perhaps one outlier had a different temperature and they adjusted the temp of the outlier to that of the group, when in fact the outlier was the only one that met all qualifications and the others should have been adjusted to the outlier.
That is only one factor, but it serves to increase the difference between raw and adjusted as your graph shows.
I do believe that when new sites are put in which eliminate several errors, there will be readings which are cooler because of poor homogenization in the past.
Where I come from, you don’t “adjust” raw data, period. Raw data are raw data. You justify including or excluding raw data in your analysis, based on a defined rationale; but you never “adjust” the raw data.
When the TR7 sonar band was constructed it had a fault wherein the distance reported was off by between 1x and 1.1x the actual distance, depending of the angle of the obstruction.
What they did was adjust the raw data received to get an accurate map. You are saying they were wrong to do that?
Precisely. The keepers of the global thermometer claim UHI does not affect the trend. However, if there is a systematic bias across time associated with the replacement of rural stations with urban stations, then UHI will gradually increase the trend in direct proportion to the rate at which rural stations are dropped. This artifact, if present, should be identifiable if you regress the change in USHCN temp on the change in the “rural ratio” : rural/(urban+rural). The parameter estimate should be negative and significant. If it looks promising, then you would want to add other variables that might affect the change in USHCN, like CO2, ENSO, TOBS, etc.
If the parameter estimates on rural ratio, TOBS, etc are significant after controlling for other variables, then you will have quantified the degree to which siting, TOBS and adjustments explain the change in recorded temperature. You could try using your differences between raw and adjusted data as the y variable in your y = f(b*x) regressions, where the b’s are your parameter estimates and the x’s are your explanatory variables.
What would cause TOBS to increase exponentially? I’m thinking TOBS should be a flat constant number to represent the adjustment until the MMTS units were introduced. Then TOBS should decrease as more MMTS units were brought into the HCN network.
I mentioned it on a previous post, but if TOBS existed in abundance, there would be a greater number of concurrent identical maximum temperature readings in pairs than would exist beyond something found naturally. You could compare the number of maximum temperature pairs from MMTS against what is found with min/max thermometers in long temperature records to see if there is a significant TOBS.
TOBS should be a one-time step change for every individual station affected by it. Not some weird, fungible number put to 130 years for careful observations.
gradual change is constant—UHI due to construction, degrading of stations as in white paint gradually eroding causing some absorption of heat,
More and more sites being compromised could have a compounding effect on the homogenization process.
TOBS = Time of Observation.
If the Tobs is wrong (reseting the min/max Six thermometer in the morning not the evening) The thermometer still gets reset every 24 hours and you are still getting min and max readings for every 24 hours. Since high temperatures normally occur after lunch but before dinner (between 1 pm to 5 pm) only thermometers reset in the middle of the day are going to have a “Double high reading” and are going to make Steve’s observations of the 1930s high readings suspect.
When is the most likely time an observer would read and reset a Six thermometer?
In the morning before work or in the evening after work, not in the middle of the day.
Farmers are the most likely keepers of the Six thermometers before and just after WWII.
1900 – Farmers made up 38% of labor force
1920 – Farmers made up 27% of labor force
1930 – Farmers made up 21% of labor force
1940 – Farmers made up 18% of labor force
1950 – Farmers made up 12.2% of labor force
1970 – Farmers made up 4.6% of labor force
1990 – Farmers made up 2.6% of labor force
Until the seventies, when Earl Butz said “Get Big or Get OUT” farms were multi-crop and many were dairy farms since it was a good year round cash crop. For example Chatham county NC still had 60 dairy farms in the late 1980s early 1990s when we moved into the area.
So when do you milk the cows? – 4 AM and 4 PM making the Tobs, either when you get up and out the door (3:30 am) or possibly when you come back to the house for your second breakfast around 6:00 to 7:00 am. A less likely time is just before bed. During the middle of the day for either a business person or a farmer is NOT very likely because you just aren’t going to remember to do it.
2000 year old problem (cf Mark 6:4).
This issue reminds me of the discredited paper by Steig et al 2009 ( the one that featured on the front cover of Nature) that smeared the Antartic temperatures together to show warming across the continent when in reality only the Antarctic Penninsula shows significant warming.
There are many posts over at WUWT about this – it seems that Anthony “got it” back then even if he doesn’t now !
When are the highs and lows are most likely to occur? About an hour after sunrise and in the after noon.
Actual data for my area just after the winter solstice:
Wednesday, February 5, 2014
Time (EST): Temp.: Dew Point: Humidity: Sea Level Pressure:
12:15 AM 39.2 °F 38.8 °F 99% 30.09 in
12:35 AM 39.2 °F 38.8 °F 99% 30.08 in
12:55 AM 39.4 °F 39.0 °F 99% 30.07 in
1:15 AM 39.2 °F 39.0 °F 99% 30.06 in
1:35 AM 39.2 °F 38.8 °F 99% 30.06 in
1:55 AM 39.4 °F 39.2 °F 99% 30.07 in
2:15 AM 39.4 °F 39.2 °F 99% 30.05 in
2:35 AM 39.4 °F 39.2 °F 99% 30.05 in
2:55 AM 39.4 °F 39.2 °F 99% 30.03 in
3:15 AM 39.4 °F 39.2 °F 99% 30.01 in
3:35 AM 39.4 °F 39.0 °F 99% 29.99 in
3:55 AM 39.2 °F 38.8 °F 99% 29.99 in
4:15 AM 39.2 °F 39.0 °F 99% 29.97 in
4:35 AM 39.2 °F 39.0 °F 99% 29.96 in
4:55 AM 39.2 °F 38.8 °F 99% 29.95 in
5:15 AM 39.2 °F 38.8 °F 99% 29.92 in
5:35 AM 39.2 °F 38.7 °F 98% 29.92 in
5:55 AM 39.4 °F 39.0 °F 99% 29.93 in
6:15 AM 39.4 °F 39.2 °F 99% 29.95 in
7:15 AM 39.4 °F 39.2 °F 99% 29.93 in
8:15 AM 39.6 °F 39.4 °F 99% 29.91 in
9:15 AM 40.1 °F 39.7 °F 99% 29.89 in
10:15 AM 42.6 °F 42.3 °F 99% 29.86 in
11:35 AM 43.7 °F 43.2 °F 98% 29.86 in
12:15 PM 44.6 °F 43.9 °F 97% 29.84 in
1:15 PM 46.8 °F 46.0 °F 97% 29.81 in
1:35 PM 47.5 °F 46.9 °F 98% 29.81 in
1:55 PM 48.4 °F 47.7 °F 97% 29.81 in
2:15 PM 49.3 °F 48.2 °F 96% 29.80 in
2:35 PM 50.7 °F 48.6 °F 92% 29.81 in
2:55 PM 50.9 °F 48.4 °F 91% 29.82 in
3:15 PM 51.4 °F 48.6 °F 90% 29.83 in
3:35 PM 51.8 °F 48.6 °F 89% 29.83 in
3:55 PM 52.2 °F 48.9 °F 89% 29.84 in
4:15 PM 52.7 °F 48.2 °F 85% 29.84 in
4:35 PM 52.7 °F 48.0 °F 84% 29.85 in
4:55 PM 52.3 °F 48.2 °F 86% 29.85 in
5:15 PM 51.8 °F 48.2 °F 88% 29.87 in
5:35 PM 50.9 °F 48.2 °F 90% 29.88 in
5:55 PM 50.4 °F 48.2 °F 92% 29.88 in
6:15 PM 48.0 °F 46.6 °F 95% 29.89 in
7:15 PM 45.0 °F 44.4 °F 98% 29.92 in
8:15 PM 43.7 °F 43.2 °F 98% 29.95 in
9:15 PM 48.9 °F 41.5 °F 76% 29.98 in
10:15 PM 47.8 °F 35.1 °F 61% 30.02 in
11:15 PM 45.0 °F 30.7 °F 57% 30.06 in
11:55 PM 43.7 °F 29.5 °F 57% 30.08 in
An hour after sunrise? You seem to have different dynamics at play in your universe than mine. Temps begin to rise about the time the rays of the sun appear here …
Perhaps you should put up a good thermometer, and check the results. It’s well known that the earliest morning sun is at too low an angle to overcome the thermal momentum of the previous night, by itself. The amount of energy recieved from the sun is dominated by the angle involved. 15 minutes after sunrise, the sun is roughly 1 degree above the horizon. That’s about the same as Barrow, Ak, at noon, in mid-April. Unless your area has already cooled to the temperature of Barrow in mid-April, the sun isn’t providing enough energy 15 minutes after sunrise to reverse the trend.
It’s the same reason the daily high temperature rarely comes at local noon, and summer solstice is almost never the hottest day of the year.
By observation I always had about one half hour before sunrise being the coldest.-But then not much warming for the next 2 hours.
It could be that the surface of the land in a particular area is a factor. I live within five miles of nine different lakes and within ten miles of about 30 lakes. –Just sayin’
—Have to think about it. Rayleigh scattering might effect the H2O absorption, now I have to really think about it.
Also obviously the latitude and longitude would make a difference because of the angles of irradiance.
Actual data for my area yesterday, after the summer solstice:
Friday, June 27, 2014
Time (EDT): Temp.: Dew Point: Humidity: Sea Level Pressure:
12:15 AM 77.9 °F 68.4 °F 72% 29.97 in
1:15 AM 75.2 °F 68.5 °F 80% 29.99 in
2:15 AM 75.2 °F 70.5 °F 85% 29.99 in
3:15 AM 73.9 °F 71.4 °F 92% 29.99 in
4:35 AM 72.5 °F 71.1 °F 95% 29.99 in
4:55 AM 73.2 °F 71.1 °F 93% 29.99 in
5:15 AM 72.9 °F 70.7 °F 93% 29.99 in
5:35 AM 72.7 °F 70.7 °F 93% 30.00 in
5:55 AM 72.7 °F 70.7 °F 93% 29.99 in
6:15 AM 72.5 °F 70.7 °F 94% 29.99 in
6:35 AM 72.5 °F 70.9 °F 95% 30.00 in
6:55 AM 72.5 °F 70.7 °F 94% 30.01 in
7:15 AM 72.5 °F 70.9 °F 95% 30.01 in
7:35 AM 73.2 °F 71.6 °F 95% 30.01 in
8:15 AM 74.3 °F 71.6 °F 91% 30.02 in
9:15 AM 77.0 °F 72.5 °F 86% 30.03 in
10:15 AM 77.9 °F 72.7 °F 84% 30.05 in
11:15 AM 79.9 °F 72.7 °F 79% 30.06 in
12:15 PM 82.4 °F 72.3 °F 71% 30.06 in
1:15 PM 82.4 °F 70.5 °F 67% 30.06 in
2:15 PM 83.7 °F 69.8 °F 63% 30.05 in
2:35 PM 85.3 °F 70.7 °F 62% 30.04 in
2:55 PM 86.0 °F 71.6 °F 62% 30.03 in
3:15 PM 84.4 °F 71.6 °F 65% 30.03 in
3:35 PM 86.5 °F 72.1 °F 62% 30.02 in
3:55 PM 86.4 °F 72.7 °F 64% 30.02 in
4:15 PM 87.3 °F 71.4 °F 59% 30.02 in
4:35 PM 87.3 °F 70.0 °F 56% 30.01 in
4:55 PM 87.1 °F 70.0 °F 57% 30.01 in
5:15 PM 88.3 °F 69.4 °F 53% 30.00 in
5:35 PM 88.3 °F 70.9 °F 56% 30.01 in
5:55 PM 87.8 °F 69.4 °F 54% 30.02 in
6:15 PM 82.4 °F 69.4 °F 65% 30.04 in
7:15 PM 79.0 °F 67.1 °F 67% 30.06 in
8:15 PM 77.4 °F 68.9 °F 75% 30.06 in
9:15 PM 74.1 °F 70.0 °F 87% 30.07 in
10:15 PM 72.5 °F 69.8 °F 91% 30.09 in
11:15 PM 72.5 °F 70.2 °F 92% 30.12 in
11:55 PM 71.8 °F 70.7 °F 96% 30.11 in
Mucking up the low readings by doing the observation in the mornings before work, is much more likely that mucking up the high readings by doing the readings in the middle of the afternoon.
evanmjones, who did a lot of the work on sifting through the data for Anthony Watts said there were Tobs issues. But they did the adjustments correctly, by hand one observation at a time instead of the lazy way of using some computer algorerhythm that is continually readjusted.
There are also the issue of fronts coming through. Are the Tobs issues being spotted in retrospect due to a front coming through and the observer noticing and changing how he re-sets the Six thermometer accordingly?