Towards a theory of carbon allocation based on eco-evolutionary optimality principles

• DOI: 10.5194/egusphere-egu23-6459
• Publication Date: 2023-02-25T20:12:41Z
• Authors: Ruijie Ding; Rodolfo Nóbrega; Iain Colin Prentice
• Keywords: Herbaceous plant

Understanding the distribution of assimilated carbon (C) among different plant parts is essential in explaining responses multiple functional traits to climate change. C allocation not adequately represented by current ecosystem models, and general explanatory framework with environmental conditions fragmented. Machine learning approaches applied large data sets have failed reveal principles underlying allocation. Here, we analyse a global set derived from several previous compilations test eco-evolutionary optimality hypotheses that potentially account for controls on root:shoot biomass ratios (R:S) both woody herbaceous plants. These are expressed terms statistical predictors describing aspects environment relevant stimuli. Thus, example, consider growing-season temperatures rather than annual means; include modelled gross primary production (GPP) root-zone water capacity (RZWC) candidate predictors. We hypothesize increasing productivity permits increased stems, automatically reducing R:S. Demand roots less warmer climates because faster nutrient turnover soils. On acid soils, need take up nutrients reduced due more open stomata thus lower optimal photosynthetic capacity. More allocated seasonal mismatches between supply demand, where RZWC required maintain availability during dry season. sandy low water-holding implies further investment uptake. Our analysis broadly supports these hypotheses, an ordinary least-squares linear regression model explains nearly three-quarters observed variation However, strategies plants differ. The expected negative relationship R:S growth temperature, positive sand content, shown only plants; while soil pH findings constitute first step towards theory response resource availability, parsimonious inclusion next-generation cycle models.