Fróðskaparrit - 01.01.1998, Blaðsíða 151

UMSKIFTI í HOLOSENA VEÐURLAGNUM í NORÐURATLANTSØKINUM,
SUM TAÐ SÆST AV TORVMÝRUROYNDUM
157
(Barber, 1981; Barber et al., 1994a, b) and
Ireland (Blackford and Chambers, 1995).
The aim of these studies has been to identi-
fy points in the peat profile where the sur-
face environment became wetter, or
changed to a drier state. The data show
some evidence of palaeoclimatic changes
that occur in a number of sites at approxi-
mately the same time- within the con-
straints of the current dating framework- al-
though other apparent changes are less
widespread. Some changes in peat bog hy-
drology appear to have been local features,
not consistently represented on a local or
regional basis.
Certain dates within the last 5,000 years
appear to occur frequently, implying wiđe-
spread change in peat-forming conditions,
the most likely explanation of which is a re-
gional-scale climatic change towards wet-
ter, and /or cooler conditions. These dates
are centred around 550, 1,000, 1,400,
2,100, 2,700, and 3,500 radiocarbon years
BP (before present, uncalibrated). The evi-
dence for some is greater than others, for
example the change at around 2,700 BP has
been precisely dated from the Netherlands
(Kilian et al., 1995; van Geel, 1996),
whereas most ages are based on conven-
tional radiocarbon dating of bulk peat sam-
ples, often with a standard deviation in the
range 50-80 years (Blackford and Cham-
bers, 1991). The exact, precise, correlation
of proftles from different sites is not yet
possible, although the introduction of wig-
gle-matched dating (van Geel, 1996) and
tephrochronology (Chambers et al., 1997)
to multi-proxy studies makes such correla-
tion likely in the near future. Fig. 1 shows
the humification curve from Rystad, north-
ern Norway (after Nilssen and Vorren,
1991). There are clear points in the profile
where abrupt changes to wetter conditions
are shown. However, not all of the most
common recurrent dates are recorded here.
Evidence for cyclic or periodic patterns
Many peat-derived proxy climate curves
appear to have a periodic component. Fig.
2, for example, shows a proxy for degree of
decomposition from blanket peat at
Migneint, north Wales, UK. Changes in the
profile appear regular and rhythmic when
smoothed using an unweighted, 3-sample
running mean. Quantitative analysis of
peat-derived time-series has shown a vari-
ety of different signals. Wijmstra et al.
(1984) suggested a 210-year signal in a
combination of peat and pollen data from
the Netherlands. Aaby (1976), had found
what appeared to be a 260 year cycle in peat
data from Denmark, and Barber et al.,
(1994a) suggested an 800 year cycle in
Sphagnum macrofossil data from Bolton
Fell Moss, northem England. Chambers et
al. (1997) produced a radiocarbon-dated
blanket peat humification curve from
southern Scotland, and spectral analysis of
the curve showed a 208 year recurrent in-
terval. Although these studies show the
likelihood of a recurrent pattern from a
number of sites, there are some problems
also. Why, for example, do different peri-
odicities appear to occur at different sites?
Part of the answer may lie with the dating
uncertainty and part with the differences
between sites, with some being more ‘sen-
sitive’ (close to detectable environmental