Comparing tree-ring chronologies and repeated timber inventories as forestmonitoring tools

Historical information on forest growth is essential to evaluate and unders tand change in managed and unmanaged forests. Two ground-truth nondestructi ve sources of information on interannual to interdecadal changes are (a) re peated timber inventories and (b) tree-ring chronologies. I present here a case study of how those two types of data can complement and benefit each o ther. At the Gus Pearson Research Natural Area, a ponderosa pine stand near Flagstaff (Arizona, USA), timber inventories were repeated by the U.S. For est Service from 1920 to 1990. The analysis of those data has revealed a de cline of individual tree growth over the 20th century, attributed to increa sed stand density. Monthly precipitation and temperature at the study area showed no overall trend from 1910 to 1990. Tree-ring data collected at the area after 1990, and spanning the last few centuries, were compared to the inventory data to represent growth trends. Periodic basal area increment co mputed from forest inventories showed parallel trends but higher absolute v alues (especially for small pines) than periodic basal area increment compu ted from increment cores. Among selected ways of developing a tree-ring chr onology, average ring area closely matched repeated forest inventories for 20th century trends and revealed that decadal-scale growth rates in the 190 0s have been anomalous compared to the previous 300 years. The mensurationa l and dendrochronological approaches to forest monitoring showed advantages and disadvantages. Repeated forest inventories quantified growth of indivi dual trees and of the entire stand, thus providing a complete picture, even in retrospect; but they had longer-than-annual resolution, and covered onl y the last decades. Dendrochronological data quantified annual xylem growth of individual trees over their whole life span, thus placing recent growth trends into a much longer historical perspective; but they had limited spa tial coverage and could lead to different trends depending on the type of s tandardization option. Overall, the combination of both approaches is recom mended for evaluating changes of forest growth at multiannual scales.