Reconstructing fire regimes with charcoal from small-hollow sediments: a calibration with tree-ring records of fire

Interpretations of charcoal records from small hollows lack a strong theoretical and empirical foundation, and thus their potential for providing useful fire-history records is unclear. To evaluate this potential, we examined charcoal records in 210Pb-dated cores from 12 small hollows and looked for evidence of 20 local fires reconstructed with tree-ring records from the surrounding forest. Using all charcoal > 0.15 mm wide we established an optimum threshold that identified charcoal peaks corresponding to known fires while minimizing charcoal peaks identified that were not associated with known fires (i.e., false positives). This threshold detected four of four high-severity fires, five of 10 moderate-severity fires, and three of six low-severity fires. Analysis of larger charcoal alone (> 0.50 mm wide) yielded nearly identical temporal patterns and detection rates, but four false positives were identified, twice as many as identified using all charcoal >0.15 mm wide. Charcoal peak magnitude was highly variable within severity classes: although half of the low-and moderate-severity fires left no detectable peaks, others left peaks larger than some high-severity fires. Our results suggest that fire detection depends strongly on fire severity and that fine-scale spatial patterns of lower-severity burns play an important role in determining the charcoal signature of these events. High detection rates for high-severity fires and low false-positive rates indicate that charcoal records from small hollows will be most useful in systems where fires are large, severe and infrequent.