The objective of this study was to test ways of overcoming extraneous wood color variation during low-magnification reflected-light image analysis of conifer tree rings and thereby improve the applicability of reflected-light image analysis in dendrochronology. Increment cores from ponderosa pines exhibiting strong heartwood discoloration were examined using image analysis, The research design included three sample preparation treatments (CONTROL, ESTRACT, or BLEACH) crossed with two dendrochronology treatments (STANDARD or RESIDUAL) crossed with two data treatments (SPLIT at the heartwood-sapwood boundary or left at FULL length) to remove the effects of the extraneous color variation. Using a combination of two ring-brightness variables and total ring width, the climate-ring growth model of the EXTRACT-RESIDUAL-FULL was strongest and explained 31.2% of variation in July-October precipitation of southeastern Arizona. Organic extraction (EXTRACT) was helpful in this study but did not fully remove heartwood discoloration. Weak bleaching (BLEACH) removed extraneous color, including heartwood discoloration, but it also removed the ring-brightness signal related to climate. Removing autocorrelation from brightness variables (RESIDUAL) overcame the problem of extraneous color but also possibly removed environmentally relevant information. Keeping brightness series at full length (FULL) worked satisfactorily. Hopefully, future research can successfully isolate some other bleaching, extraction, and/or staining treatment that removes only extraneous color variation from the wood while retaining environmentally relevant color variation so that low-magnification reflected-light image analysis can be widely applicable in dendrochronological studies of conifers.
