An approach for modelling the mean fine-root biomass of Norway spruce stands

Abstract

The applicability of a heuristic model for estimating mean fine-root biomass of Norway spruce stands based on the coordinates and the diameters at breast height (diameter at a height of 1.3 m, dbh) of their trees was tested. The model was developed based on the following assumptions which were derived from the literature: (1) the maximum distance the roots of a tree can be found depends on the dimension of the tree and exceeds the edges of the crown; (2) fine-root biomass decreases with increasing distance from the tree trunk; (3) fine-root biomass increases with the dbh; (4) maximum fine-root biomass of a tree is not allocated directly around the tree’s trunk but at some distance from the stem. On the basis of these assumptions the model calculates a relative fine-root biomass at a given point within a stand. Four different versions of the model were compared, with each version differing with respect to the assumed decrease in fine roots with decreasing dbh and the approaches used to calculate the contribution of a subject tree to the fine-root biomass at a given point within a stand (additive versus consumptive). Using regression analysis we parameterised each model type with the data of 70 soil cores from a 75-year-old Norway spruce stand in southern Germany (Bavaria). The relative fine-root biomass calculated by the four different model types accounted for 62–72% of the variation of the measured fine-root biomass. The parameterised models were used to predict the fine-root biomass of 60 given points of a second Norway spruce stand based on its dbhs and stem coordinates. The comparison of measured and predicted mean fine-root biomasses of the second stand revealed no significant differences between the measured mean and the means estimated by three of the four model types. Whereas with two of the model types we achieved means and medians, respectively, nearly identical to the measured average, none of the model types was able to predict values as high as the measured maximum. Constraints of the models and points that need to be considered regarding the minimum number of soil cores needed for a reliable parameterisation of the model are discussed.