A combined OSL and 14C dating study of charcoal production in the sandy environment of Zoersel forest (N Belgium)

Zoersel forest is one of the few ancient woodland areas in northern Belgium. It is also one of the two sites in this region where well-preserved remains of past charcoal production have been found. These relics are increasingly studied to inform on former woodland composition and exploitation, which aids in outlining conservation and preservation strategies. The age of the features is commonly obtained through 14C-dating of charcoal, which is abundant in the kiln remains. However, owing to natural and anthropogenic 14C variations, this approach does not allow establishing the chronological order of burning activities during the last few centuries (i.e. from about 1650 CE onwards). In this study, we investigate the potential of quartz-based optically stimulated luminescence (OSL) dating of the sandy sediments that are associated with the kilns and which are expected to have been exposed to heat during their operation. Using a set of 35 sediment samples collected from both charcoal-rich and underlying layers (i.e. kiln remains and substrate, respectively), we first document the OSL characteristics of sand-sized quartz in terms of signal composition, sensitivity, dose-response, and the diagnostic tests commonly used to assess the appropriateness of the single-aliquot regenerative-dose (SAR) protocol. Small (2 mm diameter) aliquots are then used to examine the distribution of equivalent dose (ED) in the samples. In general, the distributions are asymmetric, starting from a finite, non-zero value; OSL-minimum-ages derived from this population tend to be systematically higher compared to 14C-ages obtained for the same kilns. Complementary analysis of single grains of quartz extracted from the charcoal-rich layer of one of the investigated relict kilns shows that, at least for this particular feature, most grains belong to a single dose population of which the average ED is consistent with that obtained using small aliquots. Possible causes for the apparent discrepancy between OSL and 14C-dating will be discussed, together with results from additional ongoing single-grain analyses. From the available small-aliquot and single-grain ED distributions, however, it is already clear that the sandy subsurface at Zoersel forest was prone to one or more heterogenous processes, making it a challenging environment for chronometric analyses