Climatic Data - Climate Change: Identification and Projections - page 14

2.2.3. Paleotemperatures
Temperature reconstructions from before 1850 are based on archive
material: traces left on the planet by the climate in the past. Measurements of
substitution, or proxies, are extracted from this data. One of the most well-
known examples is tree rings, whose thickness varies depending on climatic
conditions. Others are based on oxygen-18 isotope ( 18 O), whose ratio with
16 O in precipitation varies according to climatic conditions. Water from rain
or snowfall is either stored in ice caps or trapped in the form of calcium
carbonate found in corals, foraminifera, lacustrine or maritime sedimentary
layers, stalagmite deposits, and so on. The value of such a proxy is that the
relation between the ratio of 18 O and 16 O isotopes and temperature is now
well established. The “isotopic thermometer” is now well calibrated, but
dating remains a source for errors. As a result, the isotopic proxies are
renowned for their good resolution at low frequency (high level of accuracy
on the average value over a long period). Although tree rings more faithfully
track annual temperature variations, they are not as good at recording
gradual trends. There are a number of proxies: the profile of temperatures
measured in boreholes can be used to reconstruct past temperatures by
inversing the laws of thermal conduction, archives recording harvest start
dates, function of summertime climatic conditions, and so on. The difficulty
comes from selecting the proxies, merging them, calibrating them and
aligning them with modern measurements.
Figure 2.2 [MOB 08] shows a sample of eleven of the proxies used,
amongst others, by Moberg and Loehle to build their respective
reconstructions. Many of them confirm the existence of a medieval warm
period (MWP, between 800 and 1200), and show that the little ice age (LIA,
between 1600-1850) was not an event limited only to central England. Some
proxies can even be used to visually assess the temperature deviations from
minimum to maximum to within around 1°C . Understandably, final
reconstructions can differ markedly, depending on the selection and
statistical processing of proxies.
Of the climate archives from which we were able to gather data, we kept
four among those which went back further than the year 1000. We took on
no more data so as to prevent the number of combinations with
reconstructions of solar activity from becoming unmanageable.
Nevertheless, these four paleotemperatures (Figure 2.3) constitute a