Evaluation of a semi-empirical model for predicting fine root biomass in compositionally complex woodland vegetation

Abstract

We used measures of plant size, distribution and root core data to evaluate capability of the model of Ammer and Wagner [2] for spatially explicit prediction of fine root biomass (FRB) in Eucalyptus populnea-dominated woodlands from xeric and mesic regions of Australia. Tree diameter and height were tested as proxy variables for plant size. For the xeric site, which had no understorey grass cover, both the height- and diameter-based models gave reasonable estimates of FRB. However, the height-model provided a better match to the measured data than the diameter-model. For the mesic site, which had a substantial ground cover dominated by C4-grasses whose contribution to FRB could not be captured by the model, neither the height- nor the diameter- model was able to predict FRB satisfactorily. This was also the case even when the contribution of the C4-grasses to FRB was estimated and accounted for after δ13C analysis of fine root samples. Overall, while it is evident that the model can be a useful tool for estimating FRB from aboveground stand inventory in both even-aged plantations and compositionally complex natural vegetation, it is also clear that it does not always provide satisfactory prediction, e.g., the mesic site. Thus, to improve the wider applicability of the model further work is needed to identify why it fails and situations it is likely to be useful. © INRA, EDP Sciences, 2007.