## North America – Canonical Graphs

### What Are These?

These are graphs of the monthly change of temperatures, year over year, and the sum of those changes over the years. A very pure form of “anomaly” process. Each month is compared to the next valid data for that month and the difference is noted (so if January is 1 C warmer than the last time January had a valid data item, you would get the JAN line at 1C on the graph). In theory, over time, this would accumulate to the change of temperature to date. In practice, we see evidence in these graphs that there have been changes to the thermometers and the data record that bias the outcomes.

### But we saw North America already?

I’d started doing a “paced” analysis and that prior North America posting was a few specific interesting observations. There were then some requests for specific graphs and a desire for a bit of a ‘jump ahead”, so I responded. I did the Pacific Basin, South America, and Asia sets of graphs as mostly “just the graphs”. Now, looking back at that North America posting, it’s a nice intro / analysis, but not a good vehicle for “A Gaggle of Graphs”…

So here I’ll be putting the graphs, in a more consistent “Hair” format. The “hair” is all that monthly detail. You could suppress it by only showing the “annual average” of them that I’ve labeled “dT/yr”, but it does not give quite as much information (even if it does look cleaner and is less confusing at first sight.)

So this will be a set of graphs just showing all of North America as “monthly change of temperature” lines for each month (the “hair”), the running total of their annual averages (that dT line) and a count of the thermometer records in each year.

North America Hair Graph by Segments

### The Big Parts – USA, Canada, Mexico, Greenland

While this accounts for the bulk of the land area, sometimes it’s the “little bits” that have the more interesting story to tell, so don’t expect these guys to “tell the tale”…

U.S.A.
Well, the USA has a LOT of data. This makes the “All Data” graph very hard to read. So I’ve made one, but you will need to click on it to see anything much. Long, flat, USA is a Carbon-Guilt Free zone, and has been for over 200 years. Particularly interesting to me is the hot 1930’s, just as our written history ( and my Dad’s descriptions to me…) recorded.

USA All Data Hair Graph of Monthly Anomalies and Running Total

Even cutting it down to 1825 start date is a bit large. Notice that the trend of the “blade” on our little hockey stick tries to get a lift in trend, but our recent cool spate and sleepy sun have quashed it. Unless you measure from a “cherry picked baseline”, there is no warming. Everything is substantially like it has been in the past. We did have some cold times, but we’ve had warm times too. 100 year weather changes. 30 year “climate” is an oxymoron.

USA 1825 Start Date Hair Graph of Monthly Anomalies and Running Total

And I’d really like to know how CO2 made it so hot in the early 1829 and the 1930’s…

Canada

Canada Hair Graph by Segments

Mexico

We looked at Mexico before as a straight graph of all data without any segmentation by date or by thermometer count:

Mexico graph of dT and monthly dT/year

And we easily see that the tops are not getting any hotter (though some of the bottoms have stopped going as far down as they used to go.

So what does it look like as segments by thermometer count with trends lines? It looks like several cooling segments spliced together and with a bit of a hockey stick blade glued onto the end, but still not enough to make a warming trend unless you measure against the cold “Dip” of the 1960’s and 1970’s that make the GIStemp “baseline” of 1951-1980 a “cherry pick”.

Mexico Hair Graph Monthly Anomalies by Thermometer Count Segments

Greenland

Do the Dane’s play hockey in Greenland? “You Betcha!” A really nice blade on the end of the stable shaft of the 1930 to 1990 dropping trend. This “pivot” of trend happening right at the same time the thermometers all get swapped around… (notice how many records are dropped right at the pivot point.)

Greenland Hair Graph of Monthly Anomalies and Running Total

### Central America

Belize

Belize Hair Graph by Segments

Guatemala

Guatemala Hair Graph by Segments

El Salvador

El Salvador Hair Graph by Segments

Honduras

Honduras Hair Graph Monthly Anomalies and Runnig Total

Nicaragua

Nicaragua Hair Graph Monthly Anomalies and Running Total

Costa Rica

Costa Rica Hair Graph by Segments

Panama

Panama Hair Graph Monthly Anomalies and Running Total

### Caribbean Islands and Bermuda

The Bermuda Islands are not really part of the Caribbean Islands, but I’m going to put the graph here as it is more like them than like the mainland.

Bermuda

Nice mostly flat dropping trend from a single thermometer. Nice.

Bermuda Hair Graph

The Bahamas

The Bahamas Hair Graph by Segments

Cuba

Jamaica

Jamaica Hair Graph of Monthly Anomalies and Running Total

Cayman

Don’t know quite what to make of this. There is a very short 29 year data record that ends in August of ’87 then jumps a gap to 2008 and 2009.

Cayman Islands Hair Graph of Monthly Anomalies and Running Total

Haiti

Don’ know what to make of this. Data effectively ends about 1965 ( when it iflls up with a load of dropouts) then we get one January datum in 1993 that’s 3/10 C cooler than January of 1967 that was itself 6/10 C cooler than the prior year to it… I’s almost like someone checks in every few decades and finds no warming so drops it… And of course, now with the quake, it’s going to be hard to get valid data for a few years.

Haiti Hair Graph of Monthly Anomalies and Running Total

Dominican Republic

Dominican Republic Hair Graph by Segments

Puerto Rico

Just for fun I’ve done a log fig trend line. Looks like a better fit than a linear line. Notice that the tops tend to bounce off the zero line, but the lower end tends to “clip” a bit over time. The end result of this “volatility compression” is more of an approach to no change at zero than an ongoing linear increase. Perhaps I ought to have tried this fit on some other graphs as well?

Puerto Rico Hair Graph of Monthly Anomalies and Running Total

Virgin Islands (U.S. and U.K.)

The U.K. Virgin Islands have NIL data.

U.S. Virgin Islands Hair Graph of Monthly Anomalies and Running Total

St. Kitts and Nevis

Eight years of data. Hardly worth it.

Saint Kitts and Nevis Hair Graph of Monthly Anomalies

Antigua and Barbuda

NIL data in the file

Guadeloupe

Don’t know exactly what to make of this other than that we had a lousy thermometer early on; and not much has changed since the 1950’s:

Guadeloupe Hair Graph

Dominica

NIL Data in the file

St Lucia

NIL data in the file

Barbados

Barbados Hair Baseline Segegments

Barbados Hair Pivot Segments

St.Vincent and the Grenadines

NIL data in file

St. Pierre and Miquelon Island

Saint Pierre and Miquelon Island Hair Chart by Pivot Segment

Grenada

Don’t know what to make of this one either. It pops up, then cuts off in 1963. Can’t have hot places in the GIStemp baseline?

Grenada Hair Graph

Martinique

Wonder what that thermometer change was in the early 1960’s? Maybe a move to a new Jet Airport? It’s about the right ‘era’ for airport growth… OK, show of hands: Who wants to go to Martinique and research the history?….

Martinique Hair Graph by Pivot Segments

OK, you, yes, you in the back, sitting down. You go. Report back when you run out of vacation time.

Netherlands Antilles

Oh, I’m beginning to really like the Dutch ;-) They seem to really like consistent instrumentation. So, if you have a vacation choice between the Netherlands Antilles vs. Trinidad and Tobago, well, just remember that Netherlands Antilles are a Carbon-Guild Free Zone, while Trinidad and Tobago are not…

Netherlands Antilles Hair Graph Monthly Anomalies and Running Total

Trinidad and Tobago

Wow! That is one heck of a Rocket Ride of hot! Given how surrounding countries are not warming, I’d expect there is a significant problem with the thermometers on Trinidad and Tobago! Yet this rocket ride, being fairly long lived and complete, will be smeared and spread to all the “nearby” places up to 1000 km away to fill in missing data (via “homogenizing”) and then it, and those other places, will be spread up to an added 1200 km as “Grid / Box” anomalies in GIStemp. Go figure…

At any rate, Trinidad and Tobago get a “Double Dip” Carbon-Guilt rating. Anyone contemplating a vacation there must keep in mind that no A/C is to be used and no fuel powered transport. You are required to walk to the pool or beach, park it in a suitable place in the sun or shade, and have your beverages (and preferably meals) brought to you by the staff. Tips to be 20% as penance… Only locally grown food to be consumed. Drinks preferably from local rum. Suffer your Carbon Guilt until you can’t stand it any longer. Then visit the Netherlands Antilles …

Trinidad and Tobago Hair Graph Monthly Anomalies and Running Total

Me? I’d figure it was a broken temperature series and go look at the instrument history.

Turks and Caicos Islands

NIL data

### Conclusion

Well, that’s it for North America. We have a lot of graphs going in different directions. We can see some Volcanic action in the early ’70s. We’ve got LOADS of missing data in Central America and far more missing in The Caribbean Islands. And at the end of it all, we see a great deal of impact from “playing with the instruments” and not very much else. In many cases, the “baseline” period used by GIStemp drops on added thermometers, but then we “have global warming” as they are taken away until we are left where we were before the baseline. Not exactly what you would expect from a regularly accumulating gas, like CO2, with maximum impact early on and exponential decrease of to date.

Basically, the data profiles here are great for seeing folks fooling around with the thermometers and the data record; and pretty lousy at showing any impact from CO2 at all.

## About E.M.Smith

A technical managerial sort interested in things from Stonehenge to computer science. My present "hot buttons' are the mythology of Climate Change and ancient metrology; but things change...
This entry was posted in dT/dt and tagged , , . Bookmark the permalink.

### 10 Responses to North America – Canonical Graphs

1. Don MatÃ­as says:

“…, but Iâ€™m still adding some islands, USA, and Greenlandâ€¦. This notice line evaporates when they are in.” Aha, AGW strikes again!

2. oldtimer says:

Based on all the work you have now done with the downloaded temperature records, would you characterise them as not worth the paper they were not written on?

3. E.M.Smith says:

FWIW, I think the data are quite usable up to 1980. Then there is a small change in the processing and we get a small step function higher in most places. I think that could possibly be corrected with the present data. Then 1990 hits.

I’ve not gone off to chaise down “what it was” but suspect it’s a change of “QA process” that tosses out both upward and downward excursions as “probably wrong” but does not allow for the fact that there are valid excursions to the cold side that exceed those to the warm side.

A thermometer over a hot surface will have convection take that heat up and away. A thermometer in a cold dip will have cold air just sit still radiating IR off into space. IF you “clip peaks equally” you will have a differential tendency to toss out valid cold readings.

That’s my working thesis for now, anyway.

The result is that, IMHO, the data since 1990 are trash.

Somewhere there ought to be the original daily input data to the “monthly mean” creation process and it could probably be ‘re-processed’ back into a useful mean, but as it stands now, we’re just “Upward Sticky” on temperatures. One hopes that eventually the “process” runs out of room to fudge; and the drop at the end of this series about 2000 gives hope.

But something broke in 1990 and it visible in pretty much every part of the world. And it isn’t CO2. It’s the change of processing that got a new “Duplicate Number” assigned to the data series in that year. You can see this as a change of the 12 digit of the “StaionID” in the original data.

4. P.G. Sharrow says:

Chiefio; your above comment caused me to recall a comment I read about a week ago from a commenter from IIRC New Zealand that in 1989 a NGO or UN orginization “helped” them to upgrade their temperature recording and data collection system. This was part of a world wide climate data collection.
I spent several hours last night trying to retrace my tracks to rediscover my source, no luck. Damn.
Maybe some one out there can help.
The person that posted wanted to point out that after the improvements the locally published temperatures no longer tracked with his own.

5. Kari Lantto says:

“But something broke in 1990 and it visible in pretty much every part of the world.”

To show this to a lazy person: Why not do two separate (sets of) diagrams. Split data 1990 and plot only (i.e. at least no monthly data) the trends before and after for as many countries as possible without making a complete mess.

Perhaps also try with orthodox first differences to avoid possibilities for criticism from hostile people. You have no lack of data?!

REPLY; [ One of the first “Ah Ha!” moments I had was a “by decade” chart of average temperatures that showed a big spike of rise just as the “Duplicate nuimbers” (or mod flags) changed. It was that “March of the Thermometers” with that spike in the newer data that was the “tell” sending me down this path of investigation. Some folks got wound around the axle about the fact I was averaging temperatures (instead of anomalies) and completely missed the point that I was looking for a “tell” not for a “Global Average Temperature”. As of now, I’ve seen that the “pivot” in 1990 comes differentially in some countries and not in others. It also looks like it comes with a new “Duplicate Number” flag. So I’ve tried a couple of ways of looking at The Pivot in the data but not settled on one yet. I did a “sort by 0, 1, 2” Duplicate Number flags vs 3 and above, but the result was a bit “dull”. I think it just drops too many whole countries in the 3+ data. So when I desperately need a break from doing graphs, that’s one of the things I’m playing with. How to best visualize The Pivot in the bulk data. Though, frankly, part of the “issue” has the appearance that the data have been “tuned” so that the diddling does not show up in the bulk view, but only when looked at in subsets (and preferably sub-sets of a form the “designers” did not expect to be used… so “by country” rather than “by Grid / Box”… ) At any rate, balancing that rampant speculation against valid analysis and against “presentation values” means some times some things I try or look at don’t get into postings for a long while…

Per “hostile people” and criticisms: It is not possible to avoid, nor even in any practical way to reduce, the criticisms of folks nor the hostility of some of them. There are people who’s pleasure comes from tossing rocks at others. Thus has it always been. Thus shall it always be. The only “sin” it to care about what they think. To accept that folks who toss rocks for sport have “merit”. They don’t. So I will do exactly nothing because of what “hostile” folks might or might not do nor because of what they might or might not criticize. It bothers me far more if I tell a friend they look “beautiful” and realize I ought to have said “stunning!”. Friends matter. It bothers me vastly more if a “newby” asks a question from genuine desire for understanding and I don’t have time to provide a quality and clear answer. Truth seekers, of any sort, matter greatly. That a “hostile” person might have a criticism based on not using classical FD? Don’t care. Not a bit. I can waste my time and my spirit trying to “make them happy” and it will never happen. They will just find some other thing to toss rocks about. So “disappear them” from your concerns. They matter not at all.

OK, with that preamble: After I’ve graphed the world, I intend to compare my results with some of the “other systems” to see how they vary. I’ve done a couple of GISS graphs already. I’ll like do a FD equivalent at some point, but that’s because I’m curious about how both dT/dt and FD perform in different contexts. It’s about something that matters… Also on the “someday” list it so do a “baseline” based system and compare dT/dt to “baseline” to GIStemp (baseline with homogenizing and fill-in and UHI and…). In particular, I suspect that the “unterminated series” you get when a thermometer ‘just ends’ is part of the “problem”; so I want to try finding a way to clean that. Yet for now, that flags the places (and countries) where the data have the most problems. So I’d not want to “fix” dT/dt just to find that it no longer tells me who has clean data and who suffers most from gratuitous thermometer change… Basically, there is a built in conflict between the “Climate Researcher” goal of fixing all the data issues so that they completely disappear; and the forensics analysis that is looking for “where are the problems” so you can find out what to check for in the other programs. So I’m starting with the most “surface the problems” code possible (straight temperature averages) and then working through successive “fixes” and seeing what problems they each surface in the data. And along the way, but more so toward the end of the process, will be looking at GIStemp to see if it handles those issues “well”, “poorly”, or “well, poorly” ;-)

FWIW, I expect FD to fail miserably on the data with lots of dropouts, and expect nearly no difference at all between FD and dT/dt on data with few dropouts. The only real difference is that I “span the gaps” while FD “takes a reset”. If there are no gaps, there is no difference. (As near as I can tell, the FD definition does not require the use of an “annual average” so my “month vs same month” variation could still be done in a FD process.)

Well, I see I’ve done another 100 line response to a 10 line comment. Guess I ought to go make more graphs … 9-)
-E.M.Smith ]

6. Kari Lantto says:

“How to best visualize The Pivot in the bulk data. ”

Why not divide each station’s (or country’s) series with its temperature 1990 and plot that for all (many) countries. As the series are now all one (1) in 1990 this will work rather well visually. If this works you could try to split these plots by 0, 1, 2+ Duplicate Number flags for some interesting sets of series. I know the software R would do that for you!

7. mandolinjon says:

I found an interesting evaluation by William Briggs about the health care bill in which he found a reference to an ideal climate as a health issue. The measure apparently requires states to establish and maintain ideal temperatures within zones. I thought you might be interested in the wording of the law. The ideas seems to me to be impossible. What do you think?

http://wmbriggs.com/blog/?p=2159

REPLY: [ Sounds a bit like Ahab and The Whale to me… I’ll look at it when I get a chance. Right now it’s near 2 am and I need some sleep… I have noticed, though, a remarkable lack of “reality awareness” among the Green Movement in general and Socialists in particular. They, collectively, seem much more interested in what “sounds good” or what “makes them feel good” than in the nature of (often unpleasant) reality and the non-negotiable nature of much of it… Oh Well, they can go ahead and “break their teeth” on it. I’ll just step aside and let them have at it. California is about to “auger in” and that may serve as a “wake up call”. We’ll see. When the State stops paying wages, welfare, and pensions, well, someone might just notice… They are already putting teachers on unpaid “furlough days” and colleges are cutting back schedules. And they will certainly be willing to tell the Fed’s to “go stuff it” on spending money for “climate control”… The Federal Government is hideously bankrupt once you count the unfunded entitlements, and the Health Care Bill was just a way to “kick the can” down the road 5 more years by sucking in loads of funding for 4 years from a fresh crop of younger folks while not expending for them. It all “hits the wall” in 5 to 10 years now, and twice as hard. We had a little hope of saving Medicare / Medicaid; up until the bill. Financially it pushes the crisis onset out just a little, but accelerates the breakage dramatically after onset. So a “starts now slow ramp over 20 years” failure becomes a “don’t start for 4 more then collapse entirely in 10”. Set your alarm clock. 2020 there is no more Medicare / Medicaid unless the Chinese will let us put it on the credit card or the bill gets dumped. Basically, the Socialists solution to socialism writ small failing, is More Socialism. Unfortunately this inevitably leads to financial ruin and collapse. (We had Medicare and it was doomed, so expanded with Medicaid et.al, that is now bankrupt, so expanded to pull everyone in often against their will… and this too shall collapse; for there is no on else left to pull in…)

Basically, IMHO, the whole agenda of this administration is a “Jump The Shark” moment. The piggy bank is empty, the public has no wealth left to tax, and the expansions being spent at the \$Trillions rate must go on the Chinese Credit Card (or as hyperinflation), but the Chinese are moving to reduce the balance and are cutting back on the “available limit”. Yeah, all in very slow motion, but a non-negotiable reality. So lay in a stock of popcorn and wine (for when the shortages hit) and watch the show. After the Great Depression II we can pick up the pieces and start over. For now the message to business is “Beatings will continue until morale improves!” and as we’ve seen in California, that just drives out business, jobs, and tax base. Until that’s learned in DC (and I think it will take “regime change” to do that) we’re headed for the hard crash. So strap in and hold the popcorn tight. It’s gonna be a bumpy ride… vague “feel good” climate mandates will be the least of it. -E.M.Smith ]

8. mandolinjon says:

ChiefIO: Thanks for your response. I agree that the progressive movements have never be able to look beyond their requirement to the unintended consequences of implementing it. I guess they feel good about destroying wealth including their own well being.

Comments are closed.