Expectations that a warming world will be associated with more hydro-climatological extremes has motivated research exploring if an associated signal is evident in paleoclimate archives. Tree-ring chronologies are central to this work because of their high temporal resolution, but they are also potentially compromised by variance artefacts associated with the evolving composition of the chronology and with data processing. Here we present two empirical methods to identify and quantify potential artefacts related specifically to temporally varying growth rate (local level, LL): LL-based partitioning analysis and LL-based chronology stripping. The two methods were developed and tested using a multi-site New Zealand kauri (Agathis australis) living-tree data set. Our results show that the methods are complementary in terms of artefact identification and quantification, and that they can provide useful insight into causal processes when used conjointly. Our results also indicate that data pre-processing to remove LL-related artefacts may be sub-optimal, that there may be an optimal standardization that minimizes bias, and that the evolving variance of kauri master chronologies over the last 500 years is not significantly affected by LL-related artefacts.