Identifying southern yellow pine cross sections from the southeastern United States using quadratic discriminant analysis on pith and second annual ring diameters
15. Life on land
01 natural sciences
0105 earth and related environmental sciences
DOI:
10.1016/j.dendro.2021.125904
Publication Date:
2021-11-09T02:24:48Z
AUTHORS (4)
ABSTRACT
Abstract Southern yellow pine specimens collected from historical structures, stumps, and coarse woody debris in forests have been difficult to identify at the species level due to similar wood anatomy. This can be problematic for dendrochronologists when identifying the correct species used in the construction of historical structures, or reconstructing forest history on the landscape and using those specimens in the context of that history. We applied a quadratic discriminant analysis (QDA) to update a century-old method plotting pith diameters against second annual ring diameters to discern one species of southern yellow pine from others. Our analysis estimates error rates for false positive and false negative determinations when comparing longleaf pine (Pinus palustris Mill.) to shortleaf (Pinus echinata Mill.) and loblolly pine (Pinus taeda L.). The cross-validated false positive error rates for the smallest dataset (n = 46), was nearly twice (9.52%) that determined as a simple proportion by counting errant observations (4.76%). QDA of the largest dataset (n = 206) gave a flatter zero contour and false positive rate (3.13%) like the proportionally determined value (1.56%), despite one additional observation being falsely assigned to longleaf pine by QDA. An unknown, unearthed southern pine specimen from southeastern Virginia was radiocarbon dated up to 500 years prior and assigned as longleaf by our method (probability ≥ 0.9998). Thus, through a QDA, it is possible to greatly improve confidence in identifications of key unknown specimens that can provide evidence of discerning one species, longleaf pine, from other southern yellow pines.
SUPPLEMENTAL MATERIAL
Coming soon ....
REFERENCES (54)
CITATIONS (2)
EXTERNAL LINKS
PlumX Metrics
RECOMMENDATIONS
FAIR ASSESSMENT
Coming soon ....
JUPYTER LAB
Coming soon ....