A recent study in the New Madrid seismic zone demonstrates that large uncertainties, often involved but rarely expressed, in paleoliquefaction studies can be reduced by conducting detailed investigations at the most promising sites for dating liquefaction features. During the site investigations, care must be taken to collect samples that will provide close maximum and minimum dates for liquefaction features. It is advisable to use two-sigma calibrated dates, rather than one-sigma calibrated dates or radiocarbon ages, when estimating ages of liquefaction features. Well-constrained ages of individual liquefaction features should provide the basis for estimating the timing of paleoearthquakes and correlating features across a region. As uncertainty in ages of liquefaction features decreases, confidence in estimates of timing, source areas and magnitudes of paleoearthquakes increases. The New Madrid study also shows that modern or historic earthquakes that induced liquefaction in the same region and whose locations and magnitudes are fairly well know can serve as calibration events for paleoearthquakes. Future efforts that could further improve the usefulness of liquefaction feature in paleoseismology include (1) the development of new techniques for dating liquefaction features directly, (2) case studies of modern earthquakes that focus on the site and spatial distributions of liquefaction feature as well as geotechnical properties of liquefaction sites and (3) more rigorous quantification of uncertainties associated with estimates of timing, source areas and magnitudes of paleoearthquakes.
