Wood decay resistance moderates the effects of tree mortality on carbon storage in the indigenous forests of New Zealand

•Canopy tree mortality was the major driver of contemporary changes in C storage.•Net C loss was lower when the dominant species had highly decay resistant wood.•C loss from legacy CWD lower when dominant species had highly decay resistant wood.•Need more data on wood decay in areas where tree mortality expected to increase.

The maintenance of carbon (C) storage in indigenous forest is a key component of efforts to manage atmospheric carbon dioxide concentrations. Increased pressures from extreme climatic events and invasive pests and pathogens pose major threats to the future stability of C storage in indigenous forests through elevated canopy-tree mortality. We assessed the potential for interspecific differences in wood decay resistance to moderate decadal-scale net C losses following canopy tree mortality. We recorded tree mortality, growth and recruitment over a period spanning almost 40 years in repeatedly surveyed plots spanning a wide range of mortality rates. We combined these survey data with national data on species-specific wood decay resistance (i.e. retention of wood density) to estimate contemporary C lost through decay of trees that died during our study. We also included C losses from CWD contributed by a major synchronous mortality event before the study period (legacy CWD C loss) for a subset of the plots where CWD C storage measurements were available. C flux from live to dead biomass (1.36 Mg ha−1 year−1, s.e. 0.16) was the main factor influencing estimated net contemporary changes in C storage, with the largest net contemporary C losses (−1.5 Mg ha−1 year−1) observed in plots experiencing high mortality. Estimated net contemporary C loss from tree mortality was reduced when the dominant species had highly decay-resistant wood. The ability to predict contemporary changes in C was significantly improved when a plot-level indicator of CWD decay resistance was included in multiple regressions. Mean legacy CWD C loss was 0.39 Mg ha−1 year−1, s.e. 0.16. When legacy losses were incorporated in net C change estimates, both the size of the legacy CWD pool and its interaction with legacy CWD decay resistance explained a significant amount of variation in net C change in multiple regressions. In plots losing large (around 3 Mg ha−1 year−1) amounts of C from live biomass (or with more than 300 Mg ha−1 C stored in legacy CWD at the start of the study) wood decay resistance altered the net C balance by as much as 1.11 Mg ha−1 year−1, which is a considerable effect given that the mean annual C assimilation rate across plots was 1.38 Mg ha−1 year−1. Thus, our study reveals strong potential for interspecific variation in decay resistance to moderate the impact of canopy tree mortality on C storage in forests. We suggest that research effort on wood decay rates should be prioritised toward areas, such as drought-prone regions of Amazonia, where forests are likely to experience synchronous mortality events more frequently in future.