Why is it so difficult to identify a single indicator of water stress in plants? A proposal for a multivariate analysis to assess emergent properties.

Because of the complexity of plant responses to water deficit, researchers have attempted to identify simplified models to understand critical aspects of the problem by searching for single indicators that would enable evaluations of the effects of environmental changes on the entire plant. However, this reductionist approach, which is often used in plant sciences, makes it difficult to distinguish systemic emergent behaviours. Currently, a new class of models and epistemology have called attention to the fundamental properties of complex systems. These properties, termed 'emergent', are observed at a large scale of the system (top hierarchical level) but cannot be observed or inferred from smaller scales of observation in the same system. We propose that multivariate statistical analysis can provide a suitable tool to quantify global responses to water deficit, allowing a specific and partially quantitative assessment of emergent properties. Based on an experimental study, our results showed that the classical approach of the individual analysis of different data sets might provide different interpretations for the observed effects of water deficit. These results support the hypothesis that a cross-scale multivariate analysis is an appropriate method to establish models for systemic understanding of the interactions between plants and their changing environment.