Tree age estimates in Fagus sylvatica and Quercus robur: testing previous and improved methods.

Abstract

The accuracy of direct (based on increment cores) and indirect (based on age-size relationships) methods of tree age estimation in Fagus sylvatica and Quercus robur was tested. This was done through increment cores and stem discs taken in an old-growth forest of Northern Spain. It was found that cross-dating was more precise than ring counting by up to 7 years per tree. Furthermore, cross-dating permitted the estimation of the age of trees with floating ring-width series, which were 7% of cored F. sylvatica and 40% of Q. robur ones. In partial cores with the arcs of the inner rings, the length of the missing radius was estimated with both a geometric method, based on the curvature of the arcs, and a new graphical method, based on the convergence of xylem rays at the pith. The graphical method was more accurate when the radial growth was eccentric, as happens in Q. robur, while both methods showed a similar accuracy for F. sylvatica, whose growth is relatively concentric. Empirical models of initial radial growth (IRG), built to estimate the number of missing rings, reduced the errors associated with other methods that assume constant growth rates. Age estimates obtained from the graphical method combined with the IRG models were within 4% of the actual age. This combination ensured age estimates with a mean accuracy of 8 years for 98% of the F. sylvatica trees, and 4 years for 89% of the Q. robur. In partial cores without the arcs of the inner rings, the length of the missing radius was estimated as the distance to the geometric centre of the tree. In that case, age estimates obtained by extrapolating the mean growth rate of the 20 innermost rings in the cores were from 10 to 20% of actual age, which coincided with results obtained in other tree species with this method. Finally, the age-diameter equations of the different cohorts produced better age estimates (from 8 to 14% of actual age) than equations of the population as a whole (from 20 to 40% of actual age). These results proved that the errors derived from doubtful assumptions, such as concentric radial growth, constant growth and recruitment rates, or the absence of anomalous rings, could be reduced by applying more realistic methods of tree age estimation.