The main figure shows eight records of local temperature variability on multi-centennial scales throughout the course of the Holocene, and an average of these (thick dark line). The data are for the period from 10000 BC to 2000 AD, which is from 12000 BP to the present time. The records are plotted with respect to the mid 20th century average temperature, and the global average temperature in 2004 is indicated. An inset plot compares the most recent two millennia of the average to other recent reconstructions. At the far right of this insert plot, it is possible to observe the emergence of climate from the last glacial period of the current ice age. During the Holocene itself, there is general scientific agreement that temperatures on the average have been quite stable compared to fluctuations during the preceding glacial period. The above average curve supports this belief. However, there is a slightly warmer period in the middle which might be identified with the proposed Holocene climatic optimum. The magnitude and nature of this warm event is disputed, and it may have been largely limited to high northern latitudes.
Because of the limitations of data sampling, each curve in the main plot was smoothed (see methods below) and consequently, this figure can not resolve temperature fluctuations faster than approximately 300 years. Further, while 2004 appears warmer than any other time in the long-term average, and hence might be a sign of global warming, it should also be noted that the 2004 measurement is from a single year (actually the fourth highest on record, see Image:Short Instrumental Temperature Record.png for comparison). It is impossible to know whether similarly large short-term temperature fluctuations may have occurred at other times, but are unresolved by the available resolution. The next 150 years will determine whether the long-term average centered on the present appears anomalous with respect to this plot.
Since there is no scientific consensus on how to reconstruct global temperature variations during the Holocene, the average shown here should be understood as only a rough, quasi-global approximation to the temperature history of the Holocene. In particular, higher resolution data and better spatial coverage could significantly alter the apparent long-term behavior (see below for further caveats). For another estimate of Holocene temperature fluctuations, see: [1]
