Metabolic Scaling Theory (MST) provides a theoretical framework for understanding and predicting how and why the morphology and function of biological systems vary with scale. It remains unclear, however, if the predictions of MST hold in variable environments. For example, in savannas, resource variability, competition, and disturbance may drive allometric relationships away from ideal predictions based upon biophysical constraints. Here we test MST botanical models, and three other prominent scaling models (Geometric Similarity, Elastic Similarity, and Stress Similarity), in non-ideal savanna systems at two hierarchical levels: whole-tree and branch-level. First, we collected branch and leaf data from 30 trees (four species) comprised of 279 branches that we harvested at three savanna sites across Mali, West Africa. Next, we use a hierarchical Bayesian approach to calculate interspecific and intraspecific scaling exponents for several allometric relationships. Specifically, we estimated tree and branch-level scaling relationships between diameter and 1) length/height, 2) aboveground biomass, 3) leaf biomass, and 4) stem biomass. Lastly, to determine if theoretical model predictions can successfully be used to scale allometries in savannas, we compare our empirical estimates to those proposed by scaling models.
Results/Conclusions
Results for each scaling relationship show all species converging toward universal allometries. Further, all scaling models tested capture these ‘global’ allometries at the interspecific and intraspecific tree-level. At the branch-level, only the scaling relationship between diameter and length/height shows any interspecific variability. Likewise, at the whole-tree level, both inter- and intraspecific scaling exponents for diameter vs. length/height trend toward a value greater than the MST prediction of 2/3. We hypothesize there are two orders of processes influencing tree allometries in savannas: 1) fire drives interspecific diameter vs. length allometries toward values greater than 2/3 (a universal effect), and 2) competition for light drives the intraspecific variability in diameter vs. length allometries (a site-specific effect). Still, MST predictions conform to 87.5% of the empirical scaling exponents we calculated and outperformed other scaling models, perhaps due to the extended biological and evolutionary assumptions upon which MST is built. We conclude that if MST is an accurate representation of biomechanical tendencies, then deviations from MST can provide insight regarding specific plant trade-offs in response to abiotic and biotic interactions.