Loess deposits contain important records of variations in atmospheric dust, providing evidence which may be used to assess the role of dust in climate change. To examine these records, it is necessary to establish a reliable, high-resolution chronology. Using the radiocarbon ages from palaeosols developed in the loess gives only an average mass accumulation rate for the package of loess that they bracket, and any dust flux calculation based on these ages implicitly assumes that the accumulation rate has remained essentially constant over the period of deposition. However, luminescence dating is applied directly to the mineral grains that make up the loess deposit, and is therefore ideally suited to the investigation of the records of dust accumulation contained in loess.

My studies have applied optically stimulated luminescence (OSL) dating to Peoria Loess deposits from midcontinental North America. These are the thickest deposits of last-glacial loess in the world and were deposited during the time from the Last Glacial Maximum to the Holocene. The mass accumulation rates (MARs) for these last-glacial loess deposits were believed to be high on the basis of thickness, however, little was actually known of the fluctuation in dust accumulation rates both during the last glacial period, and spatially across North America. Luminescence dating of Peoria Loess at several sites in midcontinental North America has revealed that MARs were extremely high between 18,000 and 14,000 years ago - much higher than those calculated for any other pre-Holocene location worldwide.

These unprecedented MARs coincide with the timing of a mismatch between palaeoenvironmental evidence from central North America, and the palaeoclimate simulations from atmospheric global circulation models (AGCMs). The high atmospheric dust loading implied by these MARs may have played an important role, through radiative forcing, in maintaining a colder-than-present climate over central North America for several thousand years after summer insolation exceeded present-day values. These findings highlight the need both for the collection of further ‘ground-truth’ data to assess the role of dust in climate change, and for dust to be incorporated into climate models.