Global Warming is Real, and Proceeding at Around 0.18°C per Decade.
Short term trends cannot tell us anything about long-term changes in global climate. For example,
let's examine some 5-year trendlines in the RSS satellite measurements of global mean temperature:
What do you notice? The trends are all over the place – from -0.4°C to nearly +1°C per decade.
It's simply not possible to determine a long-term trend from this kind of analysis. Solar activity
rises and falls on a roughly 11-year cycle, and ocean currents such as the El Niño Southern
Oscillation (ENSO) also contribute significant interannual variability to global temperatures, so
5-year trends only reflect these natural fluctuations. It's even easier to see this if we plot the
temperature data as 5-year running means:
5-year trends only reflect the rise and fall of natural variability and cannot tell us anything about
long-term global climate change.
How about 10-year trends? -
Less variability than 5-year trends, but still a lot – from about 0°C to +0.4°C per decade. How well
does this give us a handle on the longer term trend? Still not very well, but it’s clearly greater than
zero. Again, it's easy to see these 10-year trends become gradually steeper and shallower
over time as they follow the interannual variability.
How about 15-year trends? -
Now we’re beginning to see more consistency. Clearly the trendlines are all positive, ie. they show
warming. The lowest of these trendlines gives warming of about 0.1°C per decade and the highest
0.3°C per decade, but of course it would be misleading to just choose the smallest (or largest)
trend out of this bunch and say that that is representative of long-term change. Also, 15-year
trendlines are out of sync with the solar cycle. Any particular 15-year trend could start at a solar
minimum and end at a solar maximum, or vice versa, thus exaggerating or underplaying any
real underlying trend - in the case of this RSS data, producing trends which are different by a
factor of three.
Using 22-year trendlines can help to smooth out this natural variability from the 11-year solar cycle,
ENSO etc., and thus show us the underlying trend:
Now we can finally see that the warming trend is around 0.18°C per decade, just as the climate
scientists have been saying for many years.
In general, when looking for trends in data, if a particular sized subset of the data shows wildly varying
trends with different start and end points, then it cannot be claimed that any one of those trends is
representative of the long-term trend - only by selecting a subset for which the trends are in close
agreement can we find a genuine long-term trend.
RSS data from here:
We get the same kind of picture if we use the GISTEMP series instead:
Again, a warming of about 0.18°C per decade.
It's also interesting to note how well the different global temperature series match each other:
There is a very close agreement between the different series, using both satellite and land/ocean-based
measurements, which gives us confidence that the warming trend is real and not an artefact of any
particular measurement system. Here I have offset the individual curves from each other so that they
can be more easily compared, hence the temperature axis can only be taken as a guide to the magnitude
of the fluctuations, not their actual values.