Cross-sectional vs. longitudinal research: a case study of trees with hollows and marsupials in Australian forests

Abstract

How different are insights based on cross-sectional studies from those of longitudinal investigations? We addressed this question using a detailed case study encompassing a rare suite of inter-connected cross-sectional and longitudinal investigations that have spanned the past two decades and included work on: (1) the decay and collapse of large-cavity forest trees (termed "trees with hollows"), (2) populations of a suite of species of arboreal marsupials that are reliant on trees with hollows as nesting and denning sites, and (3) relationships between the abundance, type, and condition of trees with hollows and the presence, abundance, and species richness of these animals. Our case study was from the montane ash eucalypt forests of the Central Highlands of Victoria, southeastern Australia. Our longitudinal studies led to new insights that either would not have been possible from a cross-sectional study, or which were unexpected because they did not conform, or only partially conformed, to postulated responses made at the outset based on the results of earlier research. These new insights included: (1) a substantial slowing in rates of tree fall between 1997 and 2006, which were significantly lower than predicted from earlier data gathered between 1983 and 1993, (2) no evidence for a decline in populations of almost all species of arboreal marsupials between 1997 and 2007, despite the loss of nearly 14% of the measured population of trees with hollows during that time, (3) changes in nest tree selection by some species of arboreal marsupials in response to these changes in hollow availability, (4) concentration effects, in which populations of animals used the declining tree hollow resource more intensively, and (5) evidence for significant rainfall effects on temporal changes in animal abundance. Our case study underscored the additional ecological insights that can be generated from longitudinal studies, including how relationships between biota and their habitat can change over time. Understanding these temporal changes is essential for informed forest management and biodiversity conservation, and points toward the need for greater use of longitudinal data sets in ecology.