Fire dating from tree rings in western cascades Douglas-fir forests: An error analysis

Cross dating, the matching of tree-ring patterns to determine absolute date s or tree-ring series, is a valuable technique for dating wildfires. Howeve r, most recent fire history studies conducted in Pacific Northwest Douglas- fir forests have not employed cross dating. The error associated with non-c ross-dated, field-counted, fire history data was assessed at four sites in Douglas-fir forests of the western Cascades, Oregon. Fire scar and tree ori gin years were dated in the field by counting tree rings on minimally prepa red stump surfaces. Wood samples from these same stumps were then prepared in the laboratory, where tree rings were recounted and cross dated. Fire hi stories from field-counted, laboratory-counted, and cross-dated efforts wer e compared. Cross dating required 22 times the effort of the field-counted fire history reconstruction, and 87% of fire-scarred samples could be cross dated. The field-counted data generally underestimated ages of fin scar and tree origi n years, and Ares reconstructed from field-counted data were estimated as h aving occurred from 1 to 16 years more recently than they actually did. Fie ld-counted scar years were within 10 years of their true values for about 7 5%, and within 20 years for about 87% of observed cases. Errors in fire fre quency estimates were small unless an incorrect number of fires was reconst ructed. Also, the error associated with careful laboratory counts on well-p repared surfaces was minimal (mean error of 1.5 years) even when cross dati ng was not conducted. We recommend that future fire history studies in the Pacific Northwest employ cross dating.