During the past few years important progress has been made
toward recovering paleoclimate information from lake sediments.
Earlier work that documented simple, direct relations between
changes in pollen frequencies and magnetic properties covering
the last glacial-Holocene transition [summarized by Thompson
and Oldfield, 1986] has been amplified by records that now
extend back several 10
years.
From the thick lacustrine deposits in Lake Baikal, Siberia, Peck et al. [1994] demonstrated the influence of orbital periodicities on magnetic mineralogy by applying spectral analysis to magnetic-property profiles. This correlation indicates a preserved climate record extending back at least as far as 250 kyr. The climate record in Lake Baikal was also inferred from a pattern of magnetic-property variations that characterized the youngest and well-dated glacial-interglacial sedimentary units and that was repeated in older sediment [ Peck et al., 1994]. The patterns were interpreted to indicate the dilution of magnetite by biogenic production during warm intervals and an enhanced eolian input of both magnetite and hematite during cold periods when the Chinese loess deposits formed.
Recent studies of three smaller lakes in different settings
have established magnetic records of climate change by
comparisons to pollen records [ Snowball, 1993;
Rosenbaum et al., 1994; Thouveny et al., 1994]. In a
study of soil and Holocene lake beds in a glaciated valley in
Lappland, Sweden, Snowball [1993] interpreted variations in
magnetic properties in terms of fluctuations in glaciers and in
soil development. Indirect evidence was found for the
post-depositional dissolution of detrital magnetite during warm
intervals related to increased input of organic carbon and low
sedimentation rates. Sediment from Buck Lake, a small,
unglaciated lake basin in the Cascade Range, southern Oregon,
provides a pollen record of major climate change, with a closely
matching magnetic record, over an interval about 270,000-450,000
years ago [ Rosenbaum et al., 1994]. High magnetic
susceptibility and high magnetite content relative to hematite
characterize two cold-climate zones, whereas low magnetic
susceptibility and high relative amounts of hematite characterize
two warm-climate zones. The magnetic and pollen records,
however, are slightly offset over two cold-to-warm
transitions---the magnetic properties change below the changes
in pollen---suggesting that factors responsible for the
distribution of magnetic iron oxide minerals responded more
quickly to climate change than did plant cover. In addition, the
magnetic properties display high-frequency variations closely
similar to the broad magnetic variations and thus imply a more
detailed climate-change record than provided by pollen. Detailed
records of magnetic susceptibility, pollen, and organic carbon
from Lac du Bouchet, a maar lake in the Massif Central, France,
provides a detailed climate record over the past 120 kyr [
Thouveny et al., 1994]. As at the other lakes, high magnetic
susceptibility, from detrital magnetite, corresponds to cold
climates. A magnetic susceptibility-age profile, constructed
from radiocarbon ages for the last 40,000 years and from
correlations between vegetational markers and 
O
isotopic stages of the marine record for the remainder, shows a
close correspondence to the 
O curve from a Greenland
ice core.
In particular, the magnetic susceptibility curve displays abrupt
shifts during oxygen isotope stage 5e (part of the last
interglacial stage, about 115-125 kyr ago) as does the
O curve from the ice core. From this comparison
Thouveny et al. [1994] concluded that the Lac du Bouchet record
supports the idea of rapid climate changes during the last
interglacial and that the changes affected continental Europe.
In the sediments from the lakes discussed above the magnetic properties may have been controlled mainly by some combination of the following factors: (1) mechanical and chemical weathering processes in the catchment that influence the amounts, types, sizes, and availability of the detrital iron oxides; (2) hydrodynamic response of the heavy minerals, which include the magnetic iron-titanium oxides, to processes (e.g., fluvial or eolian) and energies of sediment transport; (3) depositional conditions and energies that may concentrate or disperse heavy minerals; (4) the frequency and magnitudes of events capable of transporting the detrital heavy minerals; and (5) the amount of vegetation cover as it affects sediment availability. Important goals for future work are to link in detail catchment processes to paleoclimate and to develop new diagnostic tests for detrimental alterations, such as the post-depositional production of greigite, a magnetic iron sulfide mineral [ Snowball, 1991; Verosub and Roberts, in press], and the dissolution of detrital iron oxide [ Snowball, 1993].