Estimation errors in fine-root prod

Estimation errors in fine-root production
Previous studies indicated that fine-root production estimation is an even more difficult problem than fine-root biomass estimation (e.g., Vogt et al., 1996; Nadelhoffer and Raich, 1992). For example, Vogt et al. (1996) obtained r2 < 0.2 for all relationships between fine-root production and several abiotic and biotic factors at the climatic species scale. Nadelhoffer and Raich (1992) found that there was also no sig- nificant correlation between fine-root production and litterfall when all data points in their study were in- cluded (r2 = 0.05, n = 59). In comparison, the multiple linear regression method proposed in this study was able to explain much of the observed vari- ation, using fine-root biomass, age, and annual mean temperature as independent variables, regardless of the measurement methods, with significantly lower SEE (Figure 9, Table 6). The values of SEE varied among forest types, but the magnitudes were similar. For the datasets compiled in this study, there was little dif- ference whether fine-root biomass diameter classes of < 2 mm or < 5 mm were used.
We recognize that errors may be introduced in this new approach due to uncertainties in fine-root biomass, age, annual air temperature, and their re- lationships with fine-root production, but much of the SEE in the fine-root production estimates could still be due to methodological differences between the individual studies. Vogt et al. (1998) compared five different methods to estimate fine-root production at a single forest stand (sequential core maximum- minimum method using significant differences; de- cision matrix method; sequential core