Spatially random mortality in old-growth red pine forests of northern Minnesota
Characterizing the spatial distribution of tree mortality is critical to understanding forest dynamics, but empirical studies on these patterns under old-growth conditions are rare. This rarity is due in part to low mortality rates in old-growth forests, the study of which necessitates long observation periods, and the confounding influence of tree in-growth during such time spans. Here, we studied mortality of red pine (Pinus resinosa Ait.) in five old-growth stands in Minnesota, USA, demonstrating the use of preexisting information of cohort age structures to account for in-growth after the most recent co- hort establishment. Analyses of spatial point patterns, using both Ripley's K-function and the pair correlation function, showed that tree mortality was essentially a random process, without evidence of contagious mortality patterns that are often expected for old-growth forests. Our analyses further demonstrated in practice that the distribution of dead trees may differ from that of the tree mortality events, which are constrained to occur within the initial distribution, and how mortality pat- terns can shape the spatial distribution of mature living trees, often attributed to aggregated regeneration patterns. These findings emphasize the need to disentangle the influence of the initial distribution of trees from that of actual tree mortality events.