Detecting and attributing climate change effects on vegetation: Australia as a test case

Societal Impact Statement: Climate change is contributing to vegetation changes that threaten life support systems. Yet, inherent climatic variability and past and present human actions—such as clearing, burning and grazing regimes—also alter vegetation and complicate understanding of vegetation change. Australian ecosystems exemplify such complexity. To predict future vegetation changes, proactively guide management and ensure persistent drivers do not disrupt intended outcomes, we need to untangle the effects of these various change drivers on vegetation. Such attribution of change, which is rarely done, requires historical context, long-term datasets of vegetation and environmental drivers and integrating data with process-based understanding. Summary: Climate change is expected to affect vegetation: associated rising atmospheric CO₂, higher temperatures and more variable and extreme rainfall regimes can all cause major shifts in vegetation composition, structure and function. Such effects need to be detected to confirm understanding and to inform models that can predict future vegetation change and guide management efforts. However, many change drivers—some related to, and others distinct from, climate change—simultaneously affect vegetation. These drivers include altered land management practices and shifts in fire and grazing regimes. Untangling the signals of climate-change-induced vegetation change from these other drivers of variation poses significant challenges. These challenges are amplified in regions with high interdecadal climate variability and enduring legacies of shifting human activities. Here, we assess attempts to detect and attribute vegetation change across Australia, a continent that exemplifies such complexities. We develop a scheme to classify attribution efforts according to whether they consider (1) qualitative or quantitative evidence, (2) mechanistic explanations and (3) alternative plausible change drivers. While a significant body of evidence demonstrates vegetation change in Australia, we find that it is difficult to confidently attribute changes to recent climate shifts—noting that few studies have attempted to do so. Several recommendations emerge that may improve attribution worldwide, including explicitly considering attribution strength, committing to long-term monitoring of vegetation and change drivers and recognising multiple drivers of change, especially past and present human influences. Finally, achieving the strongest level of attribution requires linking observations and mechanistic models.