First results and Further steps
The following section gives an overview about the equation comparison based on various analytical methods. Further steps will probably include a closer look on species specific equations for promising mixtures.
Equations:
Equations:
Visualization
Anyways, the first step was comparing several estimated biomasses with the real biomass, to find the best way to work on with. The displayed plot contains 5 biomasses calculated with general and broad leave specific formulas.
All equation do more or less overestimate the biomass of small trees and underestimate the biomass of bigger trees. Obviously, the fittest formulas are the ones of the general allometry (orange line) and site specific one (blueline).
1. Additive formula, which takes the parts of a tree in account
2. A general formula based on estimating the volume of a tube and adding the density 3. Power equations that rest on RCD and height, height only and RCD only Results: Overestimating over all equations. Predictions based on equations on the substance of volume provide the best result. |
This plot gives a closer look on the power equations used by Peter Annighoefer and the general, single coefficient equation.
AIC's show, that two coefficient equations are better than single coefficient ones. The existing equation seems to be fit enough to picture the reality without bigger errors.
AIC's show, that two coefficient equations are better than single coefficient ones. The existing equation seems to be fit enough to picture the reality without bigger errors.
2 Additional ways
A stargazer table, calculated with the lm function, combines various analytical methods for reliable comparisons.
One of the most powerful gadgets is Adjusted R^2 which shows us the goodness of fit. In this particular case, the equations based in Height tend to be less accurate.
The slope of the predicted biomass compared to the real biomass (see also figure 18) identifies how close the estimated biomass is to the real biomass. Furthermore the slope tells us if an equation tend to over- or underestimate the biomass.
As a last action, the comparison of equations based on height and volume are presented below. The results are similar to the results in figure 20. Two coefficient equations are better than single coefficient equations. Popular existing equations predict biomass with an error within tolerable bounds.