AmeriDendro 2013 registration is open: follow the link on the Overview page.
Controlling for modern sampling bias through tree-level effects
Presented by Mr. Jacob CECILE on 15 May 2013 from 18:00 to 21:00
Type: Poster Presentation
Session: Poster Session + Reception
Board #: 24
Non-random sampling protocols and differences in mortality between trees of different growth rates has been suggested to cause widespread, systematic bias in dendrochronological reconstructions of tree growth, a phenomenon known as modern sample bias. By making explicit the model of tree growth used in regional curve standardization, it becomes evident that tree ring data can be stored and manipulated in arrays, while the regional curve and standardized chronology (forcing) can be thought of as vectors. In this framework, we describe the origin of trend-in-signal bias in regional curve standardization and the improvements shown by using signal-free standardization. We then extend signal-free standardization and show its equivalence to a new standardization technique, factor regression standardization. A tree-specific effect can be added to the model to account for differing-contemporaneous-growth rate bias, eliminating modern sample bias and providing a novel tree-level index of productivity.