I think the easiest way to think of a temperature anomaly is to think of the comparison between absolute temperatures in Kelvin and the temperature Joe Public might more commonly use: degrees Celsius.
The temperature near me now is 291K. However, if you regarded the melting point of ice to be a reference point then the anomaly from that point would be +18°C , in round numbers. If you plotted the temperature history both in K and °C then the curves would have the same shape but be separated by approximately 273 Kelvin-sized degrees.
One difference, however, is that instead of taking a single reference value, it is usual to take a different one for each month. So you would have an average value for January calculated over 1951 to 1980, say. Similarly, for all the other months. The effect of this is to smooth out the seasonal variations. You would be able to see immediately whether either January of June was warmer or colder than the average of the reference period.
Also, if two places were near together but their temperatures differed by 2 degrees you would be able to see quite easily if that was an unusual difference or whether it was normal for those two places.
You could get a possibly unexpected situation where the absolute temperature in January was higher than in February but the anomaly showed that February was warmer. All it means is that Februarys are usually much colder.
More details are included in the link
Earth’s temperature can be expressed an anomaly and / or an actual temperature, it is both, there is no contradiction.
I fear you may not have understood what an anomaly is, nor what a statistic is.
Anomalies
ˉˉˉˉˉˉˉˉ
Imagine the average American has an income of $35,000 a year, your income is $45,000. If asked how much you earn you could answer “$45,000” or “$10,000 more than the average American”. Both are correct, both are expressing the same thing but in a different way.
It’s the same with global temperatures.
They can be expressed as an ‘actual’ or absolute value such as 14.7°C, or as an anomalous value such as 0.7°C. The anomalous values are expressed against a fixed baseline figure. The baseline is the average over a long period of time. NASA’s baseline is 1951 to 1980, HadCRU is 1961 to 1990, UAH uses 1979 to 2008.
Statistics
ˉˉˉˉˉˉˉ
A statistic is a numerical representation. Actual temperature values are statistics, anomalous values are also statistics. It is not, as you imply, a case that actual values are not statistics.
Calculating Actual and Anomalous Values
ˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉˉ
To calculate the actual temp from the anomaly all you do is add the baseline figure and the anomaly together. Example: baseline = 13.6°C, anomaly = 0.5°C therefore actual temp = 14.1°C.
To calculate the anomaly simply deduct the baseline from the actual value. Example: baseline = 14.3°C, actual temp = 14.7°C, therefore anomaly = 0.4°C.
More Examples:
ˉˉˉˉˉˉˉˉˉˉˉˉ
? NASA announces that the global temp anomaly for May 2014 was 0.99°C (this is the actual dTs NASA value). We know that NASA use 1951 to 1980 as the baseline during which time the average global temperature was 14.00°C. from that we can calculate that the actual global temp in May 2014 (based on dTs) was 14.99°C.
? JAXA (Japanese Space Agency) publish on their website that yesterday the Arctic sea-ice extent was 9.835,948km2. We can look up the historical data which tell us that the average for June 23 over the last 10 years is 10,149,470km2. If we’re using this as the baseline then the anomaly for yesterday is -213,522km2.
Comment
ˉˉˉˉˉˉˉ
You mock GARY F yet his answer is absolutely correct. Other people have also answered the question for you, it’s clear you’re not prepared to take any notice of them, you are purposely choosing not to learn. Is it any wonder you make so many very basic errors.
Graphical Format
ˉˉˉˉˉˉˉˉˉˉˉˉˉˉ
Your contention is that the temperature data expressed as anomalies differ from ‘actual’ temperatures.
Here’s a graph, I randomly selected the NASA data for the last five years, it could have been any temp record for any period of time. The actual temps are shown in red and plotted on the y2 (right) axis, the anomalies are in blue and on the y1 (left) axis.
The y2 values are offset by 0.05°C, but only so the two lines don’t overlap each other. Apart from that the graphs are absolutely identical, only the values of the datapoints differ:
I doubt that any explanation is Idiot-Denier Proof, but…
In 2003, the following cities had these numbers of murders:
Dallas 226
Detroit 366
Houston 278
Memphis 126
New Orleans 274
New York 597
Philadelphia 348
In which city are you most likely to be murdered?
Now, let’s create an “anomaly” score (aka: a linear transformation based on a standardized reference value)
In 2003, the murder rate (anomaly) for the cities was:
Dallas 18.4
Detroit 39.4
Houston 13.6
Memphis 19.3
New Orleans 57.7
New York 7.4
Philadelphia 23.3
Which set of numbers gives the best estimate of the city where you are most likely to be murdered?
Since we used the number of murders to calculate the murder rate, we can just as easily reverse the arithmetic and use the murder rate to calculate the number of murders.
Similarly, since we use temperature degrees to calculate temperature anomalies, we can just as easily reverse the arithmetic and use the temperature anomalies to calculate the temperature degrees.
Get it?
Probably not – but everyone else does.
=====
edit --
Yeah, I should have said something like actual temperature degrees since anomalies are also in degrees; but, it's hard to find a simple analogy for people who do not understand the relationship between raw counts and percentages.
The point is that comparing raw scores can introduce either real or perceived bias - and in the case of temperature it creates real bias when they are combined. Scaling them as departures from the mean makes them easier to compare and provides more accurate information on each record's trend.
At one point last night, the temperature at the top of Tucson's Catalina Mountains was in the 60s (F), while the temperature in the foothills was in the 90s (F) [the temperature at the airport was over 100 F].
Knowing that, however, does not tell you whether the temperatures at each station were warmer or cooler than normal. This makes comparisons difficult and can bias interpolations across space.
If you understood linear transformations (the general linear equation is: y=ax + b), you would not have to ask this question because you would know that temperature anomalies retain all of the original information while improving the reliability of comparisons between individual records and in averaging records.
So, how about this: x – y = z (anomaly); y + z = x (original temperature). It’s so simple, even a Caveman can understand it. Now, if only Deniers were as smart as Cavemen.
=======
Zippi62 --
>>How do statistics work when temperatures can change 27C in 1 place in 2 minutes?
... and then change back within a couple of hours?<<
The same way they work if it does not change at all. Those are just data points and - according the Law of Large Numbers and the Central Limit Theorem - if we have a statistically significant sample size, we can determine the statistical parameters of any population.
You really are clueless about how things work. You should be ashamed because, as Galileo said, "mathematics is the language with which God wrote the universe" - which means you cannot understand what God is saying, but millions of atheists can - and that is Goddamn funny.
============
edit --
>>science has no clue about identifying the internal variability of the climate and its temperature.<<
Since you lack any scientific knowledge, that is a lie - and you know that you are lying
Anomalies are better for understanding and expressing what is going on. They are used to show trends across time and across geography. To say that the Northern Hemisphere warmed from January to May tells you nothing. But plotting the anomalies add usable points of data.
For example, in the news today is that May 2014 was the hottest May ever recorded. But if you read that the average global temperature was 59.9 you learn very little. You want to know if that is different. Expressing the anomaly expresses meaning: The entire globe was 1.3 (f) degrees warmer than the 20th century average. That tells you something. You can then better break it down. In the northern hemisphere it was 2.7 degrees warmer than the 20th century average over land and 1.1 degree warmer over the oceans. In the southern hemisphere it was 2.2 degrees over the average over land and 1.1 degree warmer over the oceans.
You could express the actual degrees for each of those, but you have to add the baseline for each for anybody to understand if the temperature is warmer or cooler than average.
If you really don't understand anything about the use of averages and other statistical quantities in science, you really need to take some classes. The utility of these concepts is universal in science, it's not particularly revolutionary that they appear in atmospheric science.
Perhaps you should also study Platonic ideals so you can understand that the things you think of as "actual" temperatures are not the actual things either--nothing measured by a thermometer is an actual temperature, it's a useful attempt to quantify the actual temperature.
EDIT: Apparently Zippi62 is carrying the delusions of young Earth creationism one step further: in addition to rejecting biology, geology, physics and astronomy, he has rejected arithmetic to boot. No commutative law of addition for him--that's just the evil serpent swallowing his own tail! (And we know about the serpent)
Earth's temperature is an average
Earth's temperature is an average, and yes, an average is a statistic. But even an average temperature can't increase without adding energy to a system.
GaryF is mostly right, except anomaly baselines are different for different areas and times.
Can't wait to see the rest of the 'skeptics' fall in line with this argument.
Temperature anomaly records! You guys are simply climate science dolts when you confuse "actual" temperatures with temperature "anomalies". You only get it right when you state the idea that they are simply "statistics". Is Earth's temperature a "statistic" or an "actual" temperature?
http://www.blackhillsweather.com/chinook...
How do statistics work when temperatures can change 27C in 1 place in 2 minutes?
... and then change back within a couple of hours?
