How much does global climate normally vary from year to year?

Question

How much does global climate normally vary from year to year?

Answer 1

Almost negligible.

Answer 2

9. Climate change

As a bit of a digression, suppose the frequency of "10 year" events is observed somehow to be changing from one century to the next. Any such change might be viewed as a change of the climate, although that might be an arguable conclusion. On the other hand, it would be very hard to infer much about changes in the frequency of "500 year" events in terms of climate change, because over the time from one century to the next, it is quite possible that all that has been observed is a natural fluctuation in the frequency of "500 year" events. The ground here is getting pretty shaky.

Part of the problem is to decide what is really meant by the word "climate"? The word climate generally is taken to mean some sort of average of the weather. That much is fine, but what is the averaging period? And how much data are there to be certain (in a hard, statistical sense) that the average is changing? As already noted, solid meteorological observations are about two centuries old in most of the U.S., and some important observations (notably, those above the surface) have shorter periods of record than that. Of course, evidence can be found for what the climate might have been like a long time ago (ice ages, etc.), so the climate of the distant past was almost certainly very different from the climate of today, but it is really difficult to be certain of the details of the changing climate. And if the climate is changing all the time (as is probably the case), then whatever is called the "climate" is basically only a particular (and basically arbitrary) way of manipulating the data statistically. When the data are viewed with other choices having been made, perhaps the climate will appear more stable than it does when the choices are made another way. There are lots of good folks (as well as ignorant folks in the media and elsewhere) talking about how we humans might (or might not) be "changing the climate" ... and I have no information that says we are (or are not) changing the climate from what it would have been without human activities. However, if climate is changing all the time, how can the changes introduced by humans be distinguished from the changes that would have occurred without humans? It is very difficult to make this distinction. Even experts disagree about such things.[7] How can we have confidence that the media reports have done their job in educating us to be aware of the true situation? My belief is that we cannot rely on the media to keep us informed about such things!

--------------------------------------...

Note (added in August 1997): Recently, Reid Bryson wrote a very interesting essay [in the March 1997 issue of the Bulletin of the American Meteorological Society , Vol. 78, pp. 449-455] in which he proposes the following definition:

Climate (Climatic status) is the thermodynamic/hydrodynamic status of the global boundary conditions that determine the current array of weather patterns.

His idea is that the global boundary conditions determine the sorts of permissible weather patterns. Hence, if the boundary conditions are changing, so is the climate. He notes that this status changes with time and the season, and that the climate includes the weather patterns associated with that status. As interesting as this idea is, it's not clear to me that it really has changed anything. We do not know the precise "boundary conditions" ... the use of this term is associated with treating the problem as a boundary value problem, a mathematical term ... and so we would still have to do some sort of averaging in order to treat the problem in any practical terms. It has not been demonstrated, moreover, that the problem is a pure "boundary value" problem ... it might be that with a given set of boundary conditions, the set of permissible weather patterns could also depend on the initial conditions (an "initial value" problem, in mathematical terms). However, I like the notion that the climate necessarily includes the fluctuations associated with the weather patterns permitted by a particular set of "boundary conditions." Reid's essay is certainly an interesting proposal that at least makes an effort to avoid the logical conundrums of climate as the average of the weather. I'm inclined to be supportive of its direction, if not the practicality and appropriateness of all its abstractions.

--------------------------------------...

One particularly severe winter, or one notably long drought, or some spectacular series of devastating storms does not signal that the climate has changed. Not even all of those taken together are sufficient to foretell the Apocalypse. Nor should it be considered automatically as something unusual or particularly "abnormal," as I have been trying to show.

As I look at the data showing the variation of rainfall in Oklahoma City over a period of 91 years, I have a lot of trouble being confident in saying that climate change is or is not occurring. Yes, there recently does appear to be more precipitation than in the past. However, if I dug back into the record another 91 years, I might find find a period that looked very much like, say, the last 10 years. Given all the year-to-year variability, a true climate change (however we might want to define such a thing!) is pretty hard to detect. Even using sophisticated statistical techniques, it is hard to be completely sure! There are a lot of things that can affect the record, including the exposure of the raingauge, how the readings are taken, the design of the raingauge, the character of the surrounding region, ... many of these things have changed over the years, even at a single site. Assessing climate change is pretty doggoned difficult. Weather (and its average, the climate) changes on many time and space scales, and we can be certain that during the 200 years (or less!) of observations here in the U.S., we have sampled only a tiny fraction of those time and space variations in the weather (and climate).

Answer 3

If by climate you mean average global temperatures, the standard deviation of the GISS Land-Ocean Index over the period between 1952 and 1978 (The last period during which there was little change in average temperature) was 0.095 Degrees Celsius.  This means that the 99% confidence intervals for global temperature during this period were plus or minus 0.285 degrees Celsius.